PhosCo

SUSTAINABLE | ASX PHOSPHATE | PHO

ASX Announcement

5 May 2026

Gasaat Phosphate Project

Maiden resources on new deposits boost scale and economic outlook

Additional low-strip resources stand to increase early production and cashflow, and will feed into the updated Scoping Study

92% of global Mineral Resource inventory is Measured & Indicated

HIGHLIGHTS

Maiden Mineral Resource estimates (MRE) for the KM and SAB phosphate deposits within PhosCo's wholly owned Gasaat Phosphate Project total 20.2 Mt at 20.5% P₂O₅, including 14.2 Mt classified as Indicated.
The KM maiden Resource is 12.0 Mt at 20.5% P₂O₅ (Indicated + Inferred) with an average strip ratio of 0.4:1
The SAB maiden Resource is 8.2 Mt at 20.6% P₂O₅ (Indicated + Inferred) with an average strip ratio of 2.6:1

Table 1a – Total MRE for KM & SAB, May 2026 (JORC 2012) at 8% P₂O₅ cut-off

Classification	Tonnage (Mt)	P₂O₅ (%)
Indicated	14.2	20.7
Inferred	6.0	20.1
Total (KM & SAB)	20.2	20.5

Table 1b - KM Deposit MRE, May 2026 (JORC 2012) at 8% P₂O₅ cut-off

Classification	Tonnage (Mt)	P₂O₅ (%)
Indicated	8.6	20.6
Inferred	3.4	20.3
Total (KM MRE)	12.0	20.5

Table 1c - SAB Deposit MRE, May 2026 (JORC 2012) at 8% P₂O₅ cut-off

Classification	Tonnage (Mt)	P₂O₅ (%)
Indicated	5.6	20.9
Inferred	2.6	19.9
Total (SAB MRE)	8.2	20.6

PhosCo Ltd | Suite 2, Level 11, 385 Bourke Street, Melbourne, VIC, 3000 | ABN 82 139 255 771

T: +61 3 9692 7222 | E: [email protected]

The Gasaat global Mineral Resource inventory now stands at 166.6Mt at 20.6% P₂O₅. These represent the only phosphate resources controlled by an ASX listed company in the North African phosphate mega-province.
Measured (29%) and Indicated (63%) Resources represent 92% of total Resources at Gasaat, providing a strong foundation for a Bankable Feasibility Study.
Gasaat has potential for the development of a large scale, long life world-class phosphate mining operation and is strategically located in Tunisia close to key export markets and end users.

Table 2. Gasaat Phosphate Project Global Mineral Resources

Prospect	JORC 2012	Mt	% P₂O₅
KM (May 2026)	Indicated	8.6	20.6
	Inferred	3.4	20.3
	Total	12.0	20.5
SAB (May 2026)	Indicated	5.6	20.9
	Inferred	2.6	19.9
	Total	8.2	20.6
KEL (March 2022)	Measured	49.1	21.3
	Indicated	6.4	20.3
	Total	55.5	21.2
GK (November 2022)	Indicated	83.7	20.2
	Inferred	7.2	20.1
	Total	90.9	20.2
Global Resources	Measured	49.1	21.3
	Indicated	104.3	20.3
	Inferred	13.2	20.1
	Total	166.6	20.6

PhosCo Managing Director, Taz Aldaoud said:

"The new KM and SAB deposits are expected to have a significant impact on the early production and cashflow at Gasaat due to their low strip ratios and close proximity to the proposed processing plant.

"The already substantial global resource has grown further, extending Gasaat's scale and potential mine life, with five prospects still to be fully evaluated.

"Mine planning is underway and metallurgical test work is well-advanced, both of which will be incorporated into the updated scoping study due in the third quarter of this year.

"This study and mine planning will be done in parallel with ongoing exploration, ensuring we have several avenues to continue creating value as we seek to capitalise on growing global demand for much-needed new sources of fertiliser".

ASX Release 5 May 2026

Maiden Mineral Resource Estimates KM & SAB

PhosCo Ltd ('PhosCo or the 'Company') (ASX: PHO) is pleased to announce maiden MRE's for the KM and SAB prospects within Company's wholly owned Gasaat Phosphate Project in Northern Tunisia.

KM Mineral Resource Statement

SRK Consulting has completed a maiden JORC (2012) MRE for the KM Prospect comprising 12.0 Mt at 20.5% P₂O₅ (Indicated + Inferred).

Mineralisation is stratabound within a four-unit Eocene marine phosphorite hosted by the Chouabine Formation and preserved within a fault-bounded northeast-trending graben. The mineralised package is substantially thicker than at the neighbouring Gasaat prospects, reaching a maximum of 56.4 m. The CAP unit – the principal apatitic phospharenite mineralisation – accounts for 77% of declared tonnage and 85% of contained P₂O₅.

The maiden KM Mineral Resource, reported above an 8% P₂O₅ per-cell cut-off, is presented in Table 3. KM has not previously been the subject of a MRE.

Table 3. KM Prospect — Mineral Resource Estimate (JORC 2012) at 8% P₂O₅ cut-off

Classification	Mt	P₂O₅ (%)	Contained P₂O₅ (Mt)
Indicated	8.6	20.6	1.8
Inferred	3.4	20.3	0.7
Total	12.0	20.5	2.5

Notes:

The Mineral Resource is reported in accordance with the JORC Code (2012 Edition).
The Competent Person for the MRE is Oliver Willetts, MAusIMM(CP – Geology) member number 312940, a full-time employee of SRK Consulting (Australasia) Pty Ltd.
The Competent Person for the reporting of Exploration Data, Sampling Procedures and Results is Aymen Arfaoui, MAusIMM, Exploration Manager at PhosCo Ltd/Himilco Pty Ltd.
Mineral Resources are reported at an 8% P₂O₅ per-cell cut-off applied to the four-unit mineralised package (CLP + CDP + CAP + CMP) within the classified KM Mineral Resource volume.
Tonnages are reported on an air-dried in situ basis using a single per-domain bulk density (CLP 2.59 t/m³, CDP 2.58 t/m³, CAP 2.67 t/m³, CMP 2.72 t/m³) derived from sealed Archimedes water-displacement measurements at the PhosCo/Himilco core-processing facility.
The Mineral Resource is reported on the basis of extraction by open pit mining and reverse flotation processing.
Mineral Resources are not Ore Reserves and do not have demonstrated economic viability. There is no certainty that further exploration, feasibility studies, or development will result in the conversion of Mineral Resources to Ore Reserves.
Rounding may cause apparent differences in summation.
The effective date of the Mineral Resource estimate is 5 May 2026.

ASX Release 5 May 2026
Page 3

SAB Mineral Resource Statement

SRK Consulting has completed a maiden JORC (2012) MRE for the SAB Prospect comprising 8.2 Mt at 20.6% P₂O₅ (Indicated + Inferred).

Mineralisation is stratabound within a five-unit Eocene marine phosphorite hosted by the Chouabine Formation (Metlaoui Group) and preserved within five fault-bounded sub-basin blocks of the Rouhia Graben. The mineralised package is laterally continuous across each block, averaging 5–15 m thickness and reaching a maximum of 18 m in B6 — the largest of the five blocks. The CAP unit — the principal apatitic phospharenite mineralisation — accounts for 65% of declared tonnage and 75% of contained P₂O₅.

The maiden SAB Mineral Resource, reported above an 8% P₂O₅ per-cell cut-off, is presented in Table 4. SAB has not previously been the subject of a MRE.

Table 4. SAB Prospect — Mineral Resource Estimate (JORC 2012) at 8% P₂O₅ cut-off

Classification	Mt	P₂O₅ (%)	Contained P₂O₅ (Mt)
Indicated	5.6	20.9	1.2
Inferred	2.6	19.9	0.5
Total	8.2	20.6	1.7

Notes:
1. The Mineral Resource is reported in accordance with the JORC Code (2012 Edition).
2. The Competent Person for the MRE is Oliver Willetts (MAusIMM CP(Geology) 312940), a full-time employee of SRK Consulting (Australasia) Pty Ltd.
The Competent Person for the reporting of Exploration Data, Sampling Procedures and Results is Aymen Arfaoui (MAusIMM), Exploration Manager at PhosCo Ltd / Himilco Pty Ltd.
3. The 8% P₂O₅ cut-off is applied per cell to the three-unit mineralised package (CLP + CAP + CMP) within the classified SAB Mineral Resource volume.
4. Tonnages are reported on an air-dried in-situ basis. Bulk density is assigned per lithology from 94 audited Archimedes water-displacement measurements: CLP 2.61 t/m³, CAP 2.69 t/m³, CMP 2.70 t/m³.
5. The Mineral Resource is reported on the basis of extraction by open-pit mining and reverse-flotation processing.
6. Mineral Resources are not Ore Reserves and do not have demonstrated economic viability. There is no certainty that further exploration, feasibility studies, or development will result in the conversion of Mineral Resources to Ore Reserves.
7. Rounding may cause apparent differences in summation.
8. The effective date of the MRE is 5 May 2026.

ASX Release 5 May 2026

An additional 1.0–1.5 Mt at 19–22% P₂O₅ Exploration Target has been identified and reported separately for fault-bounded blocks within Blocks 6 and 7 not supported by sufficient internal data to qualify as Mineral Resource.

Table 5: SAB Prospect — Exploration Target (ranges of tonnage and grade)

Domain	Tonnage Range (Mt)	P₂O₅ Range (%)
B6	0.7 – 1.0	19 – 22
B7	0.3 – 0.5	19 – 22
Total	1.0 – 1.5	19 – 22

Cautionary statement (JORC Code, 2012, Clause 17): The potential quantity and grade of the Exploration Target is conceptual in nature. There has been insufficient exploration to estimate a Mineral Resource for this material and it is uncertain if further exploration will result in the estimation of a Mineral Resource. This material is not included in the Mineral Resource statement, and the term Mineral Resource must not be used in connection with the Exploration Target. Ranges are approximations.

The Exploration Target is based on actual exploration results from the SAB prospect. The volumes sit beyond the fault-bounded classification limits (B6_EW1 fault in B6; western and southern bounding structures in B7) where the structural model is not directly informed by internal drill or trench data and classification as Inferred Mineral Resource is not supported. The material is derived from the same block model as the classified Mineral Resource.

Gasaat Project Geology

The phosphate mineralisation at Gasaat is contained within the Eocene Chouabine Formation. This unit can be further divided into five sub-units (CLN–CLP–CAP–CMP–CMR) with the marine phosphorite occurring in the three central layers.

Unit	Lithology	Role in Estimate
CLN	Dolomitic nummulitic limestone	Hangingwall — waste
CLP	Phosphatic limestone	Upper transitional zone
CDP	Rudaceous phosphorite	Upper-main transitional
CAP	Apatitic phospharenite (primary ore)	Main economic unit
CMP	Marly phospharenite	Lower transitional zone
CMR	Mudstone/marl	Footwall — waste

ASX Release 5 May 2026
Page 5

ASX Release 5 May 2026
Page 6

Next Steps

Mine planning has commenced on the SAB and KM prospects in support of the updated scoping study, due Q3 CY2026. Given their low-strip nature, and proximity to the proposed processing site, SAB and KM will be prioritised in the mine plan.

Multiple work streams for the updated scoping study are well advanced, including preliminary metallurgical test work which is focused on identifying the optimal processing route. The results of this work are imminent. Work has also commenced on a mining concession application, with a Bankable Feasibility Study scheduled to commence later this year.

Assay results from Phase 1 drilling at the DOH prospect, where PhosCo announced a new phosphate discovery of scale, are expected shortly. The Company will assess the potential to define a maiden MRE at DOH to complement the existing 167Mt Gasaat resource.

Follow-up exploration and drilling is also being planned at PhosCo's wholly owned Simitu copper and base metals project, following the return of high-grade copper assays from a recent rock chip sampling program.

This announcement is authorised for release to the market by the Board of Directors of PhosCo Ltd.

For further information, please contact:

Taz Aldaoud
Managing Director
T: +61 473 230 558

In Follow PhosCo on LinkedIn
X Follow @PhoscoLtd on X

Previously released ASX Material References that relate to Gasaat Phosphate Project

2012 - Chaketma Scoping Study Delivers Positive Results, 14 August 2012
2012 - Maiden JORC Resource: 37 Mt at Chaketma Project, 9 November 2012
2013 - Chaketma Phosphate Project Gassaa Kebira Initial Resource Triples Resource inventory 18 June 2013.
2022 - Chaketma Resource Update Delivers 50% Increase at KEL Deposit to 55.5mt 15 March 2022
2022 - 90% Conversion of Inferred to Indicated Resources at GK, 17 November 2022
2022 - Scoping Study Confirms Outstanding Economics for Chaketma, 9 December 2022

Figure 1. KM Section A-A' showing generalised geology as determined by drilling, trenching, and outcrop mapping, with drill holes projected onto section

Figure 2. KM Prospect showing location of drill holes and trenches, and Section Line A-A'

Figure 3. SAB - Cross section B-B' (Block 6) showing stratigraphy, $\mathsf{P}_2\mathsf{O}_5$ , block grades and modelled structures

Figure 4. SAB Prospect showing locations of resource blocks, drill holes and trenches and Section Line B-B'

Figure 5. Location of the KM, SAB and other prospects within the Gasaat permit

COMPETENT PERSON'S STATEMENT

KM

The KM Mineral Resource Estimate is supported by two Competent Persons under JORC (2012):

The information in this report that relates to Mineral Resources is based on information compiled by Oliver Willetts, MAusIMM (CP – Geology) 312940, a Competent Person who is a Member of The Australasian Institute of Mining and Metallurgy. Mr Willetts is a full-time employee of SRK Consulting (Australasia) Pty Ltd. Mr Willetts has sufficient experience that is relevant to the style of mineralisation and type of deposit under consideration and to the activity being undertaken to qualify as a Competent Person as defined in the 2012 Edition of the 'Australasian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves'. Mr Willetts consents to the inclusion in the report of the matters based on his information in the form and context in which it appears.

The information in this report that relates to Exploration Results is based on information compiled by Aymen Arfaoui, MAusIMM, a Competent Person who is a Member of The Australasian Institute of Mining and Metallurgy. Mr Arfaoui is a full-time employee of PhosCo Ltd / Himilco Pty Ltd. Mr Arfaoui has sufficient experience that is relevant to the style of mineralisation and type of deposit under consideration and to the activity being undertaken to qualify as a Competent Person as defined in the 2012 Edition of the 'Australasian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves'. Mr Arfaoui consents to the inclusion in the report of the matters based on his information in the form and context in which it appears.

SAB

The SAB Mineral Resource Estimate is supported by two Competent Persons under JORC (2012):

The information in this report that relates to Estimation and Reporting of Mineral Resources for the SAB prospect is based on, and fairly represents, information and supporting documentation prepared by Mr Oliver Willetts (MAusIMM, CP(Geology) 312940), an employee of SRK Consulting (Australasia) Pty Ltd. Mr Willetts has sufficient experience that is relevant to the style of mineralisation and type of deposit under consideration and to the activity being undertaken to qualify as a Competent Person as defined in the JORC Code (2012). Mr Willetts has undertaken multiple site visits between 2011 and 2013 and has visited each of the SAB Mineral Resource blocks. Mr Willetts consents to the inclusion in this report of the matters based on his information in the form and context in which they appear, and in accordance with JORC Code (2012) Clause 17 takes responsibility for the form and context in which the Exploration Target (Section 10) is reported.

The information in this report that relates to Exploration Data, Sampling Procedures and Results is based on, and fairly represents, information and supporting documentation prepared by Mr Aymen Arfaoui (MAusIMM), an employee of PhosCo Ltd in the role of Exploration Manager for the Gasaat project. Mr Arfaoui has sufficient experience that is relevant to the style of mineralisation and type of deposit under consideration and to the activity being undertaken to qualify as a Competent Person as defined in the JORC Code (2012). Mr Arfaoui designed the SAB exploration programme, has overseen field data acquisition and sampling, visits site regularly, and has witnessed both the CPSA and GADD-2025 exploration programmes. Mr Arfaoui consents to the inclusion in this report of the matters based on his information in the form and context in which they appear.

ASX Release 5 May 2026
Page 10

Previously Reported Results

There is information in this announcement relating to historic data and Exploration Targets, Exploration Results or Mineral Resources which were previously announced on 15 March 2022, 17 November 2022, 9 December 2022, 3 October 2024, 26 November 2024, 13 January 2025, 11 March 2025, 19 March 2025, 28 July 2025, 10 September 2025, 18 November 2025, 18 December 2025, 28 January 2026, 12 February 2026, 26 March 2026 and 7 April 2026. Other than as disclosed in those announcements, the Company confirms that it is not aware of any new information or data that materially affects the information included in the original market announcements. The information in this announcement relating to the Company's Scoping Study are extracted from the Company's announcement on 9 December 2022 titled 'Scoping Study Confirms Outstanding Economics for Chaketma'. All material assumptions and technical parameters underpinning the Company's Scoping Study results referred to in this announcement continue to apply and have not materially changed. The Company confirms that the form and context in which the Competent Person's findings are presented have not been materially modified from the original market announcement.

Gasaat Exploration History

The first hole was drilled at Gasaat in 2011, 189 holes have been drilled across all prospects (Table 6). Most of the early drilling focused on KEL (117 holes) and GK (31 holes). Since being granted the Gasaat permit on 6 March 2025 PhosCo has drilled 31 new holes totalling 2,503m, with the focus being on KM (13 holes) and SAB (6 holes) with the aim of identifying easily accessible low strip phosphate.

Table 6. Gasaat Distribution of Drilling by Prospect

Prospect	Area Km²	Number Holes	Total Metres
KM (Kodiet Mssafin)	0.18	13	867
SAB (Sidi Ali Ben Oum Ezzine)	0.40	15	922
KEL (Kef El Louz)	3.56	117	9,128
GK (Gassaa Kebira)	2.23	31	4,355
KEA (Kel El Agued)	0.53	2	100
DOH (Douar Ouled Hamouda)	1.36	10	774
GE (Gasaat Ezerbat)	0.82	1	102
GS (Gassaa Shegira)	0.75	4	585
Total	8.90	189	16,248

Page 11

Table 1 – JORC Code 2012

KM (Kodiet Mssafin) Prospect – Table 1a

JORC Code, 2012 Edition – Table 1

Section 1 Sampling Techniques and Data

(Criteria in this section apply to all succeeding sections.)

Criteria	JORC Code explanation	Commentary
Sampling techniques	■ Nature and quality of sampling (e.g. cut channels, random chips, or specific specialised industry standard measurement tools appropriate to the minerals under investigation, such as downhole gamma sondes, or handheld XRF instruments, etc.). These examples should not be taken as limiting the broad meaning of sampling.
■ Include reference to measures taken to ensure sample representivity and the appropriate calibration of any measurement tools or systems used.
■ Aspects of the determination of mineralisation that are Material to the Public Report.
■ In cases where ‘industry standard’ work has been done, this would be relatively simple (e.g. ‘reverse circulation drilling was used to obtain 1 m samples from which 3 kg was pulverised to produce a 30 g charge for fire assay’). In other cases, more explanation may be required, such as where there is coarse gold that has inherent sampling problems. Unusual commodities or mineralisation types (e.g. submarine nodules) may warrant disclosure of detailed information.	■ Two modes of sampling contribute to the Kodiet Mssafin (KM) Mineral Resource, both completed in the 2025 campaign by Himilco Pty Ltd (a wholly owned subsidiary of PhosCo Ltd):
– 13 HQ-diameter (63.5 mm) diamond drill holes (GADD-2025 series, 818.0 m total)
– 5 surface channel trenches (GAT-01 to GAT-05).
■ The dataset comprises 483 laboratory major-oxide records covering nine drill holes and all five trenches; an additional 118 portable x-ray fluorescence (pXRF) records cover three drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) where laboratory assays are deferred to the standard core-sampling step.
■ Sample types comprise 319 half-core and 164 full-core records, all collected at nominal 1 m length (mean 0.98 m, range 0.4–1.5 m), to lithological contacts.
■ HQ core was half-cored using a diamond saw; samples were bagged and dispatched to ALS Spain where the entire submitted sample was crushed to –2 mm.
■ Trench sampling used continuous channels (~20 kg/m, average 0.98 m bin length); trenches were cut perpendicular to stratigraphy, corrected for slope and surveyed by differential GPS (DGPS).
■ All material mineralisation is hosted within the Chouabine Formation, which forms a single, laterally continuous four-unit mineralised package: CLP (upper transitional phosphate), CDP (coarse coprolitic phospharudite), CAP (apatitic phospharenite – primary ore) and CMP (lower phosphatic marl).
Drilling techniques	■ Drill type (e.g. core, reverse circulation, open-hole hammer, rotary air blast, auger, Bangka, sonic, etc.) and details (e.g. core diameter, triple	■ All KM drilling is HQ-diameter diamond core, conducted by Himilco Pty Ltd in the 2025 campaign:

Table 1 - JORC Code 2012

Criteria	JORC Code explanation	Commentary
	or standard tube, depth of diamond tails, face-sampling bit or other type, whether core is oriented and if so, by what method, etc.).	- 13 holes totalling 818.0 m, with hole lengths ranging from 31.5–110.0 m (median 60.0 m).The 10 fully processed holes (GADD-2025-03, GADD-2025-05, GADD-2025-06, GADD-2025-07, GADD-2025-08, GADD-2025-11, GADD-2025-12, GADD-2025-13, GADD-2025-14, GADD-2025-15) carry full geological and geotechnical logs.The 3 additional holes (GADD-2025-28, GADD-2025-29, GADD-2025-30) are drilled and pXRF-assayed; geological logging, geotechnical recovery and laboratory assay are deferred to the standard core-sampling step in this campaign workflow. The unsawn core for these three holes was photographically reviewed by the Competent Person and appears competent across all three holes.All holes are essentially vertical: first-station downhole readings span dips of -86.5° to -0.0° across the 10 surveyed holes, with full-hole survey records (7–13 readings per 100 m on the surveyed holes) recording dip range -84.8° to -90.0°.Holes intersect the gently west-dipping mineralised package (10–20° W) at a high angle to bedding (typically 70–80°).Core is not oriented – oriented core is not required for a gently-dipping stratabound target where intercept angle is controlled by bedding geometry.
Drill sample recovery	Method of recording and assessing core and chip sample recoveries and results assessed.Measures taken to maximise sample recovery and ensure representative nature of the samples.Whether a relationship exists between sample recovery and grade and whether sample bias may have occurred due to preferential loss/gain of fine/coarse material.	Core recovery within the mineralised package (CLP + CDP + CAP + CMP) averages 95.7% by weighted mean across 293.04 m of geotech-logged mineralised intervals.Per-domain recovery is essentially flat across the four Mineral Resource Estimate (MRE) domains: CAP 95.79%, CMP 95.24%, CLP 95.37%, CDP 95.09% – no domain-specific recovery bias has been identified, and no grade-versus-recovery correlation has been identified.CAP, the principal ore domain, accounts for 74% of the mineralised metres logged.Aggregate dataset recovery (mineralised, waste, weathering and unclassified material combined) is 94.7% across 643.90 m.Recovery data are derived from 219 logged intervals across the 10 fully processed drill holes, audited against the SRK interpreted domain model.

Table 1 - JORC Code 2012

Criteria	JORC Code explanation	Commentary
Logging	■ Whether core and chip samples have been geologically and geotechnically logged to a level of detail to support appropriate Mineral Resource estimation, mining studies and metallurgical studies.■ Whether logging is qualitative or quantitative in nature. Core (or costean, channel, etc.) photography.■ The total length and percentage of the relevant intersections logged.	■ Six mineralised intervals record sub-90% recovery: two unexplained intervals at GADD-2025-12 in pure CAP rock (a hole-wide mechanical-recovery pattern, with the hole running 91.6% overall); two fault-controlled intervals at GADD-2025-06 (a logged fault 3 m up-hole) and GADD-2025-07 (a complex fault zone west of the main domain); and two marginal intervals at GADD-2025-14 in weak (R1) CMP rock. Net recovery loss across the six intervals is 1.59 m, or 0.5% of mineralised metres logged.■ No mineralised interval records over 100% recovery.■ The three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) carry no geotechnical log; recovery will be logged at the standard core-sampling step. The unsawn core was photographically reviewed by the Competent Person and appears competent across all three holes.■ Recovery is not recorded for the trench samples.■ Geological logging covers the 10 fully processed drill holes and all 5 trenches: 67 intervals at the major-lithology level across 643.7 m, recording lithology, sub-lithology, grain size where observed, and a free-text description.■ Estimation domains for the Mineral Resource are derived from the SRK-interpreted lithology dataset – 177 intervals across all 13 drill holes and 5 trenches, refining the field log to a seven-unit stratigraphy (CLN dolomitic nummulitic limestone hanging wall, CLC dolomitic limestone, CLP, CDP, CAP, CMP, CMR mudstone footwall) that supports the four-unit mineralised package definition (CLP + CDP + CAP + CMP).■ Field-stage pXRF point measurements verify lithological logs; pXRF effectively discriminates the dolomitic limestone caprock, the phosphorite sub-units and the Cretaceous mudstone footwall.■ Logged lithology is cross-verified against laboratory major-oxide chemistry on receipt of assays, providing geochemical Quality Assurance / Quality Control (QAQC) of the field log and improving the consistency of lithostratigraphic domains in the Mineral Resource.■ Geotechnical logging covers 219 intervals across the 10 fully processed drill holes (643.90 m), recording recovery, rock quality designation

Criteria	JORC Code explanation	Commentary
		(RQD), rock strength (R1-R3), deformation style and deformation intensity. Strength codes are dominated by R2 (52%) and R3 (36%).Structural logging covers 117 point observations across the same 10 drill holes – fractures (65), veins (33), faults (10) and karst (9). Alpha angle (between structure plane and core axis) is recorded on 87% of records; beta angle is null on all records, as expected for unoriented core.Core boxes are marked with box number, core depths, driller's blocks and sample depths; core has been photographed.The three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) carry no geological, geotechnical or structural log; all three datasets will be logged at the standard core-sampling step.
Sub-sampling techniques and sample preparation	If core, whether cut or sawn and whether quarter, half or all core taken.If non-core, whether riffled, tube sampled, rotary split, etc. and whether sampled wet or dry.For all sample types, the nature, quality and appropriateness of the sample preparation technique.Quality control procedures adopted for all sub-sampling stages to maximise representivity of samples.Measures taken to ensure that the sampling is representative of the in situ material collected, including for instance results for field duplicate/second-half sampling.Whether sample sizes are appropriate to the grain size of the material being sampled.	HQ core was half-cored using a diamond saw, with sampling guided by pXRF point measurements taken at intervals down each metre. Cutting commences in low-grade overburden several metres above the mineralised package to allow for mining dilution and continues through the phosphate sequence to the underlying mudstone footwall. Samples are taken to the closest lithological boundary, then in increments of approximately 1 m within uniform lithology.Trench channel samples (~20 kg/m) follow the same nominal 1 m sample length and to lithological-boundary scheme.All samples were bagged and dispatched to ALS Spain for preparation and analysis. The entire submitted sample was crushed to -2 mm using a crusher/rotary splitter combination (70% passing 2 mm); a 250 g split is then pulverised to better than 85% passing 75 μm; pulps are sealed in double heat-sealed, air-evacuated plastic bags.Sample types comprise 319 half-core records and 164 full-core records (predominantly trench channels). Sample interval averages 0.98 m (median 1.00 m, range 0.4–1.5 m).Sample grain size at KM ranges from coarse coprolitic phosphate (CDP) through medium- to fine-grained phospharenite (CAP) to micritic marl (CMP); the 1 m nominal sample length is representative of this style of stratabound sedimentary phosphorite.QAQC samples were inserted into the laboratory analytical run: 24 certified reference materials (CRMs – GPO13, GPO14 and GPO17 in approximately equal proportions; 4.97% insertion rate) and 18 coarse

Criteria	JORC Code explanation	Commentary
		blanks (3.73% insertion rate); both within typical 3–5% targets. Three field duplicate pairs (0.62% insertion rate) are recorded, materially below the typical 5–10% target.
Quality of assay data and laboratory tests	■ The nature, quality and appropriateness of the assaying and laboratory procedures used and whether the technique is considered partial or total. ■ For geophysical tools, spectrometers, handheld XRF instruments, etc., the parameters used in determining the analysis including instrument make and model, reading times, calibrations factors applied and their derivation, etc. ■ Nature of quality control procedures adopted (e.g. standards, blanks, duplicates, external laboratory checks) and whether acceptable levels of accuracy (i.e. lack of bias) and precision have been established.	■ Laboratory analysis was conducted at ALS Chemex (Spain), returning 483 records across nine drill holes and five trenches. ■ Major oxides (Al2O3, CaO, Fe2O3, K2O, MgO, P2O5, SiO2, LOI) were analysed by ME-XRFO6m – lithium-metaborate fusion followed by XRF – a total-extraction method appropriate for phosphorite chemistry. Trace elements (As, Cd, Cr, Pb, U, Zn) were analysed by ME-ICP61 – HF-HNO3-HClO4 acid digestion followed by ICP-AES. ■ Oxide closure across the eight reported major oxides is acceptable: mean sum 97.83%, range 93.01–100.77%; nine records sum below 95% and no records sum above 105%. ■ Single-laboratory assignment across all KM samples eliminates inter-laboratory bias. ■ A total of 118 pXRF point-set records cover the three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) where laboratory assays are deferred. Readings were taken with a Hitachi X-MET8000 Expert Geo handheld unit in Mining LE FP mode at 60 seconds per reading. The unit is calibrated to laboratory CRMs prepared by Geostats Pty Ltd from regional sedimentary phosphorite, and is checked against a CRM approximately every 10 readings during routine use. Readings were taken with the instrument perpendicular to and in contact with the core in the core tray; data are downloaded daily by a single qualified technician. ■ pXRF records have been used in the MRE in three drill holes where sampling was pending. pXRF was validated against laboratory assays in adjacent holes and CRM standards inserted into the pXRF sequence. ■ QAQC sample inserts in the laboratory analytical run comprise 24 CRMs (4.97%; GPO13, GPO14 and GPO17) and 18 blanks (3.73%) – both within typical insertion targets. Three field duplicate pairs (0.62%) are recorded, materially below the typical 5–10% target and flagged for resolution.

Criteria	JORC Code explanation	Commentary
Verification of sampling and assaying	■ The verification of significant intersections by either independent or alternative company personnel. ■ The use of twinned holes. ■ Documentation of primary data, data entry procedures, data verification, data storage (physical and electronic) protocols. ■ Discuss any adjustment to assay data.	■ Verification of the KM exploration database is provided through an SRK data audit covering collar, survey, geology, geotechnical, structural, laboratory assay, pXRF, density and QAQC tables. ■ The audit confirms zero gaps, zero overlaps and zero internal duplicates within the depth-keyed drill hole and trench datasets. Geotech-to-SRK-lithology apportionment reconciles to 643.90 m logged-in versus 643.90 m apportioned-out with zero drift on the join. ■ pXRF records on the three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) are field-stage measurements and are treated as provisional pending laboratory assays. ■ No twinned holes are present at KM. The case for twinning is reduced by the dense drill-and-trench network within a single mineralised domain, and by consistent vertical grade profiles observed between adjacent drill holes spaced 40–50 m apart. Three drill holes completed after the initial geological model was drafted have been incorporated into the resource estimate without materially changing the contained resource at the prospect scale, providing reconciliation-style support for the geological interpretation.
Location of data points	■ Accuracy and quality of surveys used to locate drill holes (collar and down-hole surveys), trenches, mine workings and other locations used in Mineral Resource estimation. ■ Specification of the grid system used. ■ Quality and adequacy of topographic control.	■ All 10 fully processed drill hole collars are surveyed by DGPS. The 3 additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) carry handheld-GPS coordinates pending DGPS upgrade. ■ All 5 trench collars are DGPS-surveyed. ■ All coordinates are in Universal Transverse Mercator (UTM) Zone 32N, World Geodetic System 1984 (WGS84) datum (European Petroleum Survey Group code EPSG:32632); elevations are in metres above mean sea level. ■ Topographic control derives from an airborne LiDAR survey accurate to approximately ±0.3 m, supplemented by DGPS outcrop-contact traverses for surface geological mapping. The native LiDAR point cloud was resampled to a 10 m grid digital elevation model (DEM) for the Mineral Resource model. ■ Downhole deviation surveys are available for all 10 fully processed drill holes (76 station records taken at nominal 10 m intervals – 7 to 13 records per 100 m of hole). The 3 additional drill holes carry a single

Criteria	JORC Code explanation	Commentary
Data spacing and distribution	■ Data spacing for reporting of Exploration Results. ■ Whether the data spacing and distribution is sufficient to establish the degree of geological and grade continuity appropriate for the Mineral Resource and Ore Reserve estimation procedure(s) and classifications applied. ■ Whether sample compositing has been applied.	■ Drill hole spacing within the KM deposit ranges from approximately 40 m to 50 m in the more densely drilled portions of the deposit to +100 m at the deposit margins; the dataset comprises 13 drill holes and 5 trenches across a surface extent of approximately 579 m east-west by 441 m north-south. ■ Spacing density is reinforced by the dense surface outcrop of the mineralised package, channel sampling at five trenches across the deposit, and the 10 m LiDAR topographic surface. ■ The data spacing is sufficient to establish geological and grade continuity at a level supporting Indicated and Inferred classification. ■ Three drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) drilled after the initial geological model was drafted have been incorporated into the resource estimate without materially changing the contained resource at the prospect scale, providing reconciliation-style support for the adopted spacing density.
Orientation of data in relation to geological structure	■ Whether the orientation of sampling achieves unbiased sampling of possible structures and the extent to which this is known, considering the deposit type. ■ If the relationship between the drilling orientation and the orientation of key mineralised structures is considered to have introduced a sampling bias, this should be assessed and reported if material.	■ The KM mineralised horizon is a tabular, gently west-dipping (10–20° W) stratabound phosphorite hosted in a single fault-bounded domain. ■ Drilling comprises 13 essentially vertical diamond drill holes (full-hole dip range -84.8° to -90.0°). Channel trenches were cut perpendicular to stratigraphy and corrected for slope during sample collection. ■ Vertical drill holes intersect the gently west-dipping bedding at a high angle (approximately 70–80° to the bedding plane), giving apparent downhole intercept lengths within ~6% of true thickness. ■ Channel-trench thicknesses represent true thickness. ■ Faults at KM include a north-south striking, east-dipping (~30°) western bounding fault, an east-west trending southern bounding fault, and two east-west internal faults (Fault_3 and Fault_6) which segment the deposit into the Northern, Central and Southern sub-blocks. Internal faults exhibit predominantly strike-slip displacement with a subordinate normal component; closely spaced drilling has confirmed that mineralisation is sharply sheared and cleanly offset across these structures, with no along-fault thickening. ■ No drilling-orientation bias to sampling has been identified.

Criteria	JORC Code explanation	Commentary
Sample security	■ The measures taken to ensure sample security.	■ Core is not oriented; alpha angle (between structure plane and core axis) is the primary orientation measurement on structural records. ■ Drill core boxes are transported on the same day to PhosCo's core processing facility at Rohia under controlled conditions using a 4×4 field transport system under security supervision. ■ Core is in the custody of the drilling contractor until transported to the Rohia facility, at which point control transfers to PhosCo. Core is logged, sampled and photographed at Rohia. ■ Following mechanical sample preparation, samples are split into 800 g to 1 kg sizes and placed in paper bags, that are placed into 50 L plastic barrels for secure transport. ■ Prepared samples are transported under a documented chain of custody to the airport and dispatched to the ALS laboratory in Spain. ■ pXRF field analysis is conducted on a Hitachi X-MET8000 Expert Geo unit at Rohia; data are downloaded daily by a single qualified technician. ■ Database integrity and the single-operator/single-laboratory dataset structure support chain-of-custody confidence.
Audits or reviews	■ The results of any audits or reviews of sampling techniques and data.	■ Audits and reviews supporting the KM Mineral Resource dataset: - SRK 2026 current-program audits – database audit; core recovery audit; specific gravity (SG) (bulk density) audit; geological model workflow; grade validation summary. - SRK 2026 Gassat Geochemical Review – provides the lithogeochemical framework, cross-prospect geochemical profiling and the basis for the SRK-interpreted domain codes used at KM. ■ No external third party audit of the current KM Mineral Resource has been undertaken as at the date of this report.

Section 2 Reporting of Exploration Results

(Criteria listed in section 1 also apply to this section.)

Criteria	JORC Code explanation	Commentary
Mineral tenement and land tenure status	■ Type, reference name/number, location and ownership including agreements or material issues with third parties such as joint ventures, partnerships, overriding royalties, native title interests, historical sites, wilderness or national park and environmental settings. ■ The security of the tenure held at the time of reporting along with any known impediments to obtaining a licence to operate in the area.	■ The Gassat Project (formerly Chaketma) is held 100% by Himilco Pty Ltd, a wholly owned subsidiary of PhosCo Ltd (ASX:PHO). ■ The current exploration permit was granted on 6 March 2025 for a 3-year term, covering approximately 56 km2 across eight phosphate prospects: KEL, GSK, SAB, KEA, DOH, GEZ, KH and KM. KM is the northernmost prospect on the permit. ■ The Chaketma exploration licence was first granted in 2010; held by Chaketma Phosphates SA (CPSA – a joint venture of Celamin Ltd and Tunisian Mining Services) from 2011 to 2015; and transferred to Himilco in 2025. ■ The Company and licence are in good standing with the Tunisian Ministry of Industry, Mines and Energy. ■ At this stage of the exploration licence no joint venture, partnership or off-take agreements are in place, and no private net smelter return royalties have been identified. The project is held under an exclusive exploration licence; no mining (exploitation) rights have been granted. ■ Land use across the KM area is predominantly agricultural and pastoral. No known environmental, heritage or land-tenure issues are currently considered blocking factors, subject to confirmation through the required regulatory studies (environmental impact assessment or environmental and social impact assessment (EIA/ESIA), land and heritage assessments). ■ Any future transition to mining will require relevant mining exploitation permits and regulatory approvals under the Tunisian Mining Code.
Exploration done by other parties	■ Acknowledgment and appraisal of exploration by other parties.	■ Modern exploration at KM has been conducted entirely by Himilco/PhosCo in the 2025 campaign; no CPSA-era drilling, trenching or sampling was undertaken at the KM prospect. ■ Earlier regional studies of the Tunisian phosphates by the Centre d'Études et de Recherches des Phosphates Minéraux (CERPHOS) and the Compagnie des Phosphates de Gafsa (CPG) covered the broader Gassat licence area as part of regional reconnaissance programs; copies

Criteria	JORC Code explanation	Commentary
		of these studies have not been obtained by PhosCo and they do not contribute data to the current Mineral Resource Estimate. ■ No historical workings (open pits, galleries or test trenches) are recorded at KM. ■ No third party exploration datasets, technical reports or standalone studies specific to the KM prospect have been identified or obtained.
Geology	■ Deposit type, geological setting and style of mineralisation.	■ KM is a stratiform Eocene marine phosphorite deposit of the North African phosphogenic province: ■ Mineralisation is hosted in the Chouabine Formation (Late Palaeocene to Early Ypresian) of the Metlaoui Group, deposited in a shallow restricted marine basin around the Kasserine Island palaeohigh under upwelling-driven nutrient supply. ■ Deposit type: marine sedimentary phosphorite of the Gafsa-Metlaoui style. ■ Structural setting: KM sits on the western edge of the Rouhia Basin – a northwest-southeast collapse graben disrupting the northeast-southwest Hercynian trend of the Tunisian Atlas – and forms a single, northeast-trending, fault-bounded graben. ■ The Chouabine Formation at KM is a seven-unit lithostratigraphic package (base to top): CMR mudstone footwall; CMP marly phospharenite; CAP apatitic phospharenite (primary ore); CDP coprolitic phospharudite; CLP phosphatic limestone (transitional upper); CLC dolomitic limestone; CLN dolomitic nummulitic limestone hanging wall. ■ The mineralised package (CLP + CDP + CAP + CMP) has a median preserved thickness of approximately 30 m, reaching a maximum of 56.4 m in GADD-2025-03 – substantially thicker than the other Chaketma prospects. ■ The Mineral Resource is defined within a single coherent fault-bounded domain. The deposit is bounded to the west by an east-dipping (“30”) north-south striking listric normal fault, to the south by an east-west trending fault, and is internally segmented by two east-west faults (Fault_3 and Fault_6) into the Northern, Central and Southern sub-blocks; stratigraphic continuity across the sub-blocks supports modelling and estimation as a single, geometrically uncut volume. ■ Francolite (carbonate-fluorapatite) is the dominant phosphate mineral.

Criteria	JORC Code explanation	Commentary
Drill hole Information	■ A summary of all information material to the understanding of the exploration results including a tabulation of the following information for all Material drillholes: - easting and northing of the drillhole collar - elevation or RL (Reduced Level - elevation above sea level in metres) of the drillhole collar - dip and azimuth of the hole - downhole length and interception depth - hole length. ■ If the exclusion of this information is justified on the basis that the information is not Material and this exclusion does not detract from the understanding of the report, the Competent Person should clearly explain why this is the case.	■ The KM dataset comprises 13 diamond drill holes (GADD-2025 series, 818.0 m total) and 5 channel trenches (GAT-01 to GAT-05). ■ All coordinates are in UTM Zone 32N (WGS84; EPSG:32632); elevations are in metres above mean sea level. ■ All 13 drill holes are essentially vertical: full-hole downhole-survey records on the 10 fully processed holes show a dip range of -84.8° to -90.0°. ■ Ten fully processed drill hole collars are surveyed by DGPS; the three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) carry handheld-GPS coordinates pending DGPS upgrade. ■ All five trench collars are DGPS-surveyed. ■ The three additional drill holes carry a placeholder downhole-survey record at depth 0 m only; full downhole survey is scheduled for the standard core-sampling step. ■ A complete drill hole and trench schedule (collar XYZ, total depth, dip, azimuth and analytical-data coverage) is presented in the companion public report.
Data aggregation methods	■ In reporting Exploration Results, weighting averaging techniques, maximum and/or minimum grade truncations (e.g. cutting of high grades) and cut-off grades are usually Material and should be stated. ■ Where aggregate intercepts incorporate short lengths of high grade results and longer lengths of low grade results, the procedure used for such aggregation should be stated and some typical examples of such aggregations should be shown in detail. ■ The assumptions used for any reporting of metal equivalent values should be clearly stated.	■ Not applicable. This Mineral Resource Estimate supersedes the underlying Exploration Results; data aggregation methods relevant to length-weighted intercept reporting are not material to the disclosure (Joint Ore Reserves Committee (JORC) Code 2012, Clause 27).
Relationship between mineralisation widths and intercept lengths	■ These relationships are particularly important in the reporting of Exploration Results. ■ If the geometry of the mineralisation with respect to the drillhole angle is known, its nature should be reported. ■ If it is not known and only the downhole lengths are reported, there should be a clear statement to this effect (e.g. ‘down hole length, true width not known’).	■ The KM mineralised package is gently west-dipping (10-20° W). ■ All drilling is essentially vertical (full-hole dip range -84.8° to -90.0° across 13 holes). ■ Vertical drilling into the gently-dipping bedding gives apparent downhole intercept lengths within approximately 6% of true thickness across the deposit. ■ Channel trenches were cut perpendicular to strike and corrected for slope during sample collection. Reported channel thicknesses represent true thickness.

Criteria	JORC Code explanation	Commentary
		■ The stratigraphic block-model framework reconciles intercept length to true thickness during estimation; no apparent-versus-true residual flows through to reported block tonnes. ■ No artificial width inflation arises from the combination of drilling orientation and mineralisation geometry.
Diagrams	■ Appropriate maps and sections (with scales) and tabulations of intercepts should be included for any significant discovery being reported These should include, but not be limited to a plan view of drillhole collar locations and appropriate sectional views.	■ Not applicable. Diagrams supporting the KM Mineral Resource Estimate are presented in the companion public report; the criterion is not material to this Table 1 (JORC Code 2012, Clause 27).
Balanced reporting	■ Where comprehensive reporting of all Exploration Results is not practicable, representative reporting of both low and high grades and/or widths should be practiced to avoid misleading reporting of Exploration Results.	■ Not applicable. This Mineral Resource Estimate supersedes the underlying Exploration Results; the criterion is not material to this disclosure (JORC Code 2012, Clause 27).
Other substantive exploration data	■ Other exploration data, if meaningful and material, should be reported including (but not limited to): geological observations; geophysical survey results; geochemical survey results; bulk samples – size and method of treatment; metallurgical test results; bulk density, groundwater, geotechnical and rock characteristics; potential deleterious or contaminating substances.	■ Not applicable. Substantive supporting exploration data are described in the relevant Section 1 and Section 3 criteria of this Table 1; no further disclosure is required (JORC Code 2012, Clause 27).
Further work	■ The nature and scale of planned further work (e.g. tests for lateral extensions or depth extensions or large-scale step-out drilling). ■ Diagrams clearly highlighting the areas of possible extensions, including the main geological interpretations and future drilling areas, provided this information is not commercially sensitive.	■ Completion of laboratory assays, geological/geotechnical/structural logging and DGPS/downhole-survey upgrades on the three additional drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) at the standard core-sampling step. ■ A dedicated laboratory and QAQC audit closing out KM-specific QAQC-performance documentation, including resolution of the field-duplicate underpopulation. ■ A trench-versus-drill hole geochemical bias check by major oxide within each estimation domain. ■ Forward-looking exploration and study work: ■ Sighter metallurgical testwork is underway, focused on the mineralogy and metallurgical characteristics of the KM phosphate sub-units to inform the optimal processing route.

Criteria	JORC Code explanation	Commentary
		■ Definition of the eastern margin geometry, currently constrained primarily by surface mapping and outcrop traverses with limited drilling control. ■ Finalisation of the resource-model treatment for the SAU (saprolite) and CLR (clay regolith) weathering horizons present in two southern drillholes. ■ Subsequent drilling at KM and across the wider Gassat licence is contingent on the outcomes of the maiden Mineral Resource Estimate and the integrated Scoping Study update.

Section 3 Estimation and Reporting of Mineral Resources

(Criteria listed in section 1, and where relevant in section 2, also apply to this section.)

Criteria	JORC Code explanation	Commentary
Database integrity	■ Measures taken to ensure that data has not been corrupted by, for example, transcription or keying errors, between its initial collection and its use for Mineral Resource estimation purposes.
■ Data validation procedures used.	■ The KM exploration database has been audited by SRK across collar, survey, geology, geotechnical, structural, laboratory assay, pXRF, density and QAQC tables.
■ Audit findings: zero gaps, zero overlaps and zero internal duplicates within the drill hole and trench datasets.
■ Oxide closure across the eight reported major oxides is acceptable: mean sum 97.83%, range 93.01–100.77%; nine records sum below 95% and no records sum above 105%.
■ Field duplicate QAQC reveals excellent reconciliation between parent and duplicate assays with extremely low nuggets (near zero), reflecting the bulk stratiform nature of the material.
■ pXRF measurements have been compared against laboratory assays and possess a small negative bias against laboratory geochemistry, indicating that the pXRF-supported intervals represent a conservative grade measurement.
■ Data are directly fed from the audited dataset into the geological model and grade estimation.
Site visits	■ Comment on any site visits undertaken by the Competent Person and the outcome of those visits.
■ If no site visits have been undertaken indicate why this is the case.	■ Oliver Willetts, MAusIMM(CP – Geology) 312940, undertook multiple site visits to the Gassat licence area between 2011 and 2013 and is familiar with the deposit geology and dataset. Mr Willetts was also responsible for compiling KM data into the project database. Mr Willetts has not visited the KM prospect, nor witnessed exploration and sample preparation procedures. Mr Willetts acts as Competent Person for estimation and reporting of Mineral Resources.
■ Aymen Arfaoui, MAusIMM is currently employed by PhosCo as Exploration Manager for the Gassat project. Mr Arfaoui designed the KM exploration program, has overseen field data acquisition and sampling at KM, and visits site regularly. He has witnessed both the CPSA and Himilco/PhosCo (GADD) exploration programs across the licence. Mr Arfaoui acts as Competent Person for reporting of exploration data, sampling procedures and results.

Criteria	JORC Code explanation	Commentary
Geological interpretation	Confidence in (or conversely, the uncertainty of) the geological interpretation of the mineral deposit. Nature of the data used and of any assumptions made. The effect, if any, of alternative interpretations on Mineral Resource estimation. The use of geology in guiding and controlling Mineral Resource estimation. The factors affecting continuity both of grade and geology.	The KM Mineral Resource is defined within a single coherent fault-bounded domain. The Chouabine Formation phosphorite package is hosted within tilt-block of the Rouhia graben. Tabular mineralisation approximates a gentle syncline folded about an east-west axis (like the adjacent Gassaa Kebira prospect) and is internally segmented by two east-west faults (Fault_3 and Fault_6) into the Northern, Central and Southern sub-blocks. Stratigraphic continuity is preserved across the internal east-west faults: drilling and outcrop mapping demonstrate that mineralisation is sharply sheared and cleanly offset, with no along-fault thickening or grade discontinuity. Estimation runs across the internal faults through a continuous, geometrically uncut block model. The four-unit mineralised package (CLP + CDP + CAP + CMP) is robustly defined by drilling, lithogeochemical interpretation against laboratory major-oxide chemistry, and surface mapping of contacts. Geological confidence is high in the Central sub-block, where drilling is at 50–100 m spacing and three infill drill holes (GADD-2025-28, GADD-2025-29, GADD-2025-30) confirmed the pre-existing block model with less than 0.2% movement on contained CAP P2O5. Confidence is appropriate-to-classification in the Northern sub-block (drilling at 80 m to 100 m spacing supplemented by surface trenches) and in the Southern sub-block (mineralisation outcropping at surface, mapped contacts verified in the structural model). GADD-2025-08 sits to the west of the western bounding fault and intersected only CLC and CMR – no mineralised stratigraphy. The hole is excluded from the modelled domain volume but retained in the database for completeness. Multiple realisations were produced during geological modelling and grade estimation. Testing included: different domain boundary configurations, fault interactions; grid layering and trend directionality, variogram parameters and interpolant configurations. These serve as alternate interpretations in the optimisation of the model and resource estimate.

Criteria	JORC Code explanation	Commentary
Dimensions	■ The extent and variability of the Mineral Resource expressed as length (along strike or otherwise), plan width, and depth below surface to the upper and lower limits of the Mineral Resource.	■ The KM Mineral Resource occupies a footprint of approximately 450 m east-west by 700 m north-south. ■ The mineralised package (CLP + CDP + CAP + CMP) has a median preserved thickness of approximately 30 m, reaching a maximum of 56.4 m in GADD-2025-03 – thicker than the other Gassat prospects, indicating better local preservation and trap configuration. ■ The deposit lies at shallow depth: surface to base of mineralisation is typically less than 100 m, with several drill hole intercepts commencing at or near surface. ■ The block model spans the deposit at 10 m × 10 m horizontal cell resolution. Vertical cell resolution is set per-domain on a structurally-conformant basis: 1 m floor-following steps in the principal CAP ore domain (referenced upward from the basal CAP-CMP contact) and proportional layering in the CLN, CLC, CLP, CDP, CMP and CMR domains. ■ Resource depth, lateral continuity and footprint are constrained by 13 diamond drill holes, 5 channel trenches and surface-mapped outcrop contacts integrated into the structural model.
Estimation and modelling techniques	■ The nature and appropriateness of the estimation technique(s) applied and key assumptions, including treatment of extreme grade values, domaining, interpolation parameters and maximum distance of extrapolation from data points. If a computer assisted estimation method was chosen, include a description of computer software and parameters used. ■ The availability of check estimates, previous estimates and/or mine production records and whether the Mineral Resource estimate takes appropriate account of such data. ■ The assumptions made regarding recovery of by-products. ■ Estimation of deleterious elements or other non-grade variables of economic significance (e.g. sulphur for acid mine drainage characterisation). ■ In the case of block model interpolation, the block size in relation to the average sample spacing and the search employed. ■ Any assumptions behind modelling of selective mining units. ■ Any assumptions about correlation between variables.	■ Structural and grade-domain model. A seven-unit stratigraphic model (CLN, CLC, CLP, CDP, CAP, CMP, CMR) was built using minimum-curvature horizon interpolation against well markers from the SRK-interpreted lithology dataset. The four mineralised domains (CLP, CDP, CAP, CMP) are the grade-estimation domains. Faults enter the model as bounding geometry and sub-block dividers (Fault_3 and Fault_6 on east-west trends; western and southern bounding faults defining domain limits) but do not break horizon interpolation or grid cells. Estimation runs through a continuous, geometrically uncut grid. ■ Block model. 10 m × 10 m horizontal cell resolution. Vertical cell structure is domain-conformant: 1 m floor-following steps in CAP referenced upward from the basal CAP-CMP contact (preserving depositionally banded internal grade structure); proportional layering in the other six domains sized to mean unit thickness. ■ Composites. Variable-length blocked composites generated by intersecting raw downhole assay intervals against the local block-model layer geometry, weighted by cell-overlap fraction. Composite lengths

Criteria	JORC Code explanation	Commentary
	Description of how the geological interpretation was used to control the resource estimates. Discussion of basis for using or not using grade cutting or capping. The process of validation, the checking process used, the comparison of model data to drillhole data, and use of reconciliation data if available.	follow the structural-model layering rather than a regularised downhole length. Variography. Per-domain, per-oxide variograms manually fitted against blocked-composite experimental variograms. Manual fitting was preferred to automatic fitting given per-domain composite counts (CLP n = 34, CDP n = 31, CMP n = 84, CAP n = 407). The CAP P2O5 variogram is spherical with ranges 120 m × 80 m × 3.5 m, azimuth 0°. Estimation method. Ordinary Kriging in IJK (block-model index) space – anisotropy and ranges fitted to the structurally-conformant layer geometry rather than world coordinates. There is a maximum of 20 composites contributing to each block estimate. Seven major oxides estimated independently per domain (P2O5, MgO, CaO, SiO2, Al2O3, Fe2O3, K2O). Negative kriged estimates near sparse-data domain edges (uncommon, concentrated on trace oxides) zero-clipped post-estimation. Validation. Composite-versus-block-model grade validation completed for all four estimated domains. All four domains pass at the 10% bias threshold on P2O5; variance ratios sit below 1.0, reflecting the expected information-effect smoothing for ordinary kriging. Local validation against composites was undertaken using swath plots alongside visual validation of grade alignment on section.
Moisture	Whether the tonnages are estimated on a dry basis or with natural moisture, and the method of determination of the moisture content.	Tonnages are reported on an air-dried in situ basis. Bulk density is determined from sealed air-dried cores by Archimedes water displacement. Natural in situ moisture is not a material tonnage factor at MRE stage.
Cut-off parameters	The basis of the adopted cut-off grade(s) or quality parameters applied.	Mineral Resource is reported at a cut-off grade of 8% P2O5, applied to the mineralised domain package (CLP + CDP + CAP + CMP) within the classified KM Mineral Resource volume. The 8% cut-off is consistent with the natural grade-population break between carbonate/marl waste (predominantly below 5% P2O5) and the mineralised phosphorite package (predominantly above 10% P2O5 through the core, grading to 5–15% at the upper and lower contacts). MgO is the principal product-specification deleterious element for phosphate concentrate. MgO at KM is at its lowest in the principal CAP ore domain (~3.2% mean within the cut-off) and is elevated in the

Criteria	JORC Code explanation	Commentary
		dolomite-cemented CDP unit (~7.2%) and in the lower phosphatic-marl CMP (~5.4%). The 8% P2O5 cut-off effectively screens higher MgO concentrations at the estimation domain scale; processing-route assumptions and product-spec compliance are anticipated outputs of in-progress sighter metallurgical testwork. ■ The cut-off captures the full mineralised package including gradational contact fringes without incorporating waste material. ■ The 8% value is consistent with Arethuse's 2016 wireframing practice at KEL and is lower than the 10% cut-off applied by Geos Mining to the 2013 KEL and GK Mineral Resource Estimates. Both KEL and GSK are part of the same depositional system as KM. ■ The more detailed geochemical/lithofacies control from the SRK 2026 geochemical review enables confident distinction between mineralisation and waste at the gradational CLP and CMP contacts.
Mining factors or assumptions	■ Assumptions made regarding possible mining methods, minimum mining dimensions and internal (or, if applicable, external) mining dilution. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential mining methods, but the assumptions made regarding mining methods and parameters when estimating Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the mining assumptions made.	■ KM Mineral Resource is reported with open pit mining as the assumed extraction method. ■ Open pit amenability is supported by the shallow mineralisation depth, outcrop along flanks, and low strip ratio (0-1.6; averaging 0.4), consistent with the 2022 Scoping Study assumptions at the neighbouring KEL and GK. ■ Continuous tabular stratiform geometry is amenable to bulk or selective mining. ■ No specific pit optimisation has been run against the KM Mineral Resource Estimate at this stage. Pit design and mining-factor testing are anticipated during the Scoping Study update integrating KM and SAB alongside the existing KEL and GK resources. ■ Mining-related dilution, ore loss and minimum-mining-width considerations are outside the scope of the Mineral Resource Statement.
Metallurgical factors or assumptions	■ The basis for assumptions or predictions regarding metallurgical amenability. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential metallurgical methods, but the assumptions regarding metallurgical treatment processes and parameters made when reporting Mineral Resources may not always be rigorous. Where this is	■ Metallurgical assumptions currently rely on project-level testwork rather than KM-specific testwork. ■ The 2017 Pilot Plant Report established the baseline reverse flotation flowsheet for the Chaketma phosphorite and demonstrated production of a saleable concentrate from KEL and GK feed with acceptable recovery, grade and deleterious-element characteristics.

Criteria	JORC Code explanation	Commentary
	the case, this should be reported with an explanation of the basis of the metallurgical assumptions made.	■ The 2017 Chaketma Pre-feasibility Study (PFS) integrated this flowsheet into a full process design; the 2022 Scoping Study confirmed continuing applicability. ■ KM phosphorite is geologically continuous with the Chaketma feed material – same Chouabine lithostratigraphy, same domain framework, same dominant mineralogy (francolite in a calcareous matrix) – and is assumed compatible with the established flowsheet at estimation stage. ■ Sighter metallurgical testwork on KM core is currently underway, and focused on the mineralogy and metallurgical characteristics of the KM phosphate sub-units to inform the optimal processing route. ■ Deleterious-element levels at KM are comparable with other Tunisian phosphate ores and CPG operating ores: ■ MgO is of greatest commercial attention because it can affect beneficiation at concentrations exceeding 6%. MgO averages 3.81% across all domains in the classified Mineral Resources. ■ CAP (3.22% MgO) accounts for ~80% of the Mineral Resources; CMP (5.4% MgO) is borderline and accounts for ~15% of the Mineral Resources; CLP (8.7%) and CDP (7.2%) contain more dolomite and account for ~5% of the Mineral Resources. ■ SRK considers blending of higher MgO domains with CAP represents a pathway to meet processing specifications. Alternately, CLP and CDP might represent a potential run of mine direct application product. ■ Cadmium levels are elevated and correlate with increasing P2O5 grades (length-weighted averages: CLP: 7.3 ppm, CDP: 21.5 ppm, CAP: 39 ppm, CMP: 19.1 ppm). Decadmiation of phosphoric acid concentrate is one method to control cadmium levels and produce a marketable product.
Environmental factors or assumptions	Assumptions made regarding possible waste and process residue disposal options. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider the potential environmental impacts of the mining and processing operation. While at this stage the determination of potential environmental impacts, particularly for a greenfields project, may not always be well advanced, the status of early consideration of these potential environmental impacts should be reported. Where these	■ Environmental framework assumptions rely on the broader Chaketma/Gassat Project context established in the 2022 Scoping Study and 2017 PFS. ■ The Chaketma/Gassat Project area is not identified as environmentally sensitive at regional scale – no national park, Ramsar wetland, World Heritage or equivalent designation overlaps the permit. The terrain is semi-arid with no major perennial watercourses across the KM area.

Criteria	JORC Code explanation	Commentary
	aspects have not been considered this should be reported with an explanation of the environmental assumptions made.	■ Conceptual waste rock and tailings management assumptions are established at project level in the 2017 PFS. KM-specific characterisation has not been conducted. ■ The area hosts occasional rural population with local employment participation an engagement theme since the CPSA era. No pastoralist or Indigenous community land tenure conflicts are identified. ■ The Tunisian Mining Code sets the EIA pathway for conversion of exploration tenements to a mining licence; the framework was positively assessed in the 2022 Scoping Study. ■ No environmental factor identified to date would materially preclude eventual economic extraction.
Bulk density	■ Whether assumed or determined. If assumed, the basis for the assumptions. If determined, the method used, whether wet or dry, the frequency of the measurements, the nature, size and representativeness of the samples. ■ The bulk density for bulk material must have been measured by methods that adequately account for void spaces (vugs, porosity, etc.), moisture and differences between rock and alteration zones within the deposit. ■ Discuss assumptions for bulk density estimates used in the evaluation process of the different materials.	■ Bulk density is applied as a single per-domain mean (in t/m3) across the four mineralised domains: - CLP: 2.59 t/m3 - CDP: 2.58 t/m3 - CAP: 2.67 t/m3 - CMP: 2.72 t/m3. ■ Density values are derived from drill core measurements at the SRK Rohia core-processing facility. Sealed air-dried cores are weighed and their volume determined by Archimedes water displacement; bulk density is reported on an air-dried in situ basis. ■ Each measurement was screened in a two-stage audit: a physical-plausibility check (rejecting records below 2.0 t/m3 or above 3.5 t/m3, with intermediate values flagged for review); and a stoichiometric-envelope check that compares each measurement against a predicted bulk-density range derived from co-located XRF major-oxide chemistry. The chemistry-based prediction decomposes each sample into mineral components (apatite, dolomite, calcite, siliciclastic), computes a zero-porosity grain density, and applies a per-domain porosity envelope. Measurements outside the envelope are excluded from the clean dataset. ■ The per-domain mean was selected over spatial interpolation and over chemistry-based regression. Spatial interpolation does not add precision because the within-domain coefficient of variation is low (around 5%)

Criteria	JORC Code explanation	Commentary
		and spatial structure is weak at the available sample density. Within-domain density-versus-single-oxide R2is below 0.27 across all domains with adequate support, indicating that single-oxide regression cannot reliably predict density. The constant-per-domain approach matches that adopted at SAB. ■ Sample support: CAP n = 40 across eight holes (the strongest support); CMP n = 9; CDP n = 6 (pooled with the geochemically-equivalent SAB CLP population per domain-equivalence analysis); CLP n = 2 (the most thinly supported domain).
Classification	■ The basis for the classification of the Mineral Resources into varying confidence categories. ■ Whether appropriate account has been taken of all relevant factors (i.e. relative confidence in tonnage/grade estimations, reliability of input data, confidence in continuity of geology and metal values, quality, quantity and distribution of the data). ■ Whether the result appropriately reflects the Competent Person's view of the deposit.	■ The KM Mineral Resource is classified as Indicated and Inferred: ■ Indicated classification is applied across the full mineralised domain volume except where downgraded by one of two structural criteria (see Inferred below). ■ This reflects the high degree of internal geological and grade consistency observed between points of observation and across east-west faults and the robustness of the geological model to the addition of the three latest drill holes. ■ Inferred classification is applied to the periphery of the deposit on two grounds: ■ A 25 m hard buffer along the north-south bounding fault wireframes, reflecting positional uncertainty in the wireframe (single direct fault intercept in GADD-2025-15) and a sharp but finite transition zone observed across the Gassat field as the north-south graben-controlling faults terminate mineralisation. ■ A classification downgrade polygon capturing the more widely-spaced peripheral parts of the eastern margins of the deposit on data-density grounds, including the eastern margin where pXRF-only support from GADD-2025-29 and GADD-2025-30 dominates. ■ Classification is applied uniformly to all four estimation domains (CLP, CDP, CAP, CMP), reflecting the operational reality that all four mineralised domains are mined together.

Criteria	JORC Code explanation	Commentary
		■ The Central sub-block carries the densest drilling (50–100 m spacing), reinforced by GADD-2025-28, GADD-2025-29 and GADD-2025-30 infill drilling that confirmed the pre-existing block model with less than 0.2% movement on contained CAP P2O5. The Southern sub-block carries surface outcrop of mineralisation with mapped contacts integrated into the structural model. The Northern sub-block carries somewhat sparser drilling (80–100 m spacing) reinforced by surface trenches. ■ East-west internal faults (Fault_3 and Fault_6) are not buffered – drilling and mapping demonstrate that mineralisation is sharply sheared and cleanly offset across these structures, with grade continuity preserved on either side. ■ The classification framework supports the geometry of the deposit and the level of confidence in tonnage and grade.
Audits or reviews	■ The results of any audits or reviews of Mineral Resource estimates.	■ The estimation has not been independently reviewed by a third party. No external audit of the current KM Mineral Resource has been undertaken as at the date of this report. ■ SRK has undertaken an internal peer review in line with company quality management protocol.
Discussion of relative accuracy/confidence	■ Where appropriate, a statement of the relative accuracy and confidence level in the Mineral Resource estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the resource within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors that could affect the relative accuracy and confidence of the estimate. ■ The statement should specify whether it relates to global or local estimates, and, if local, state the relevant tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used. ■ These statements of relative accuracy and confidence of the estimate should be compared with production data, where available.	■ The level of confidence supports the Indicated and Inferred classification categories assigned. ■ Empirical confidence indicators: ■ Composite-versus-block-model grade validation passes the 10% bias threshold on P2O5 across all four mineralised domains, with variance ratios below 1.0 – reflecting the expected information-effect smoothing for ordinary kriging. ■ Pre-infill/post-infill movement on contained CAP P2O5 following the GADD-2025-28, GADD-2025-29 and GADD-2025-30 drilling was less than 0.2%; the infill drilling confirmed the pre-existing block model in stratigraphic structure and grade. ■ Parallel-estimate testing of the headline grade (laboratory + pXRF) versus a laboratory-only alternative produced a deposit-scale movement on contained CAP P2O5 of less than 0.05%, demonstrating that the pXRF-supported drill holes do not materially change the resource at deposit scale.

Criteria	JORC Code explanation	Commentary
		■ Field duplicate QAQC reveals excellent reconciliation between parent and duplicate assays with extremely low nuggets (near zero), reflecting the bulk stratiform nature of the material. ■ Recovery within the mineralised package averages 95.7% across geotech-logged drill metres, with no domain-specific recovery bias and no grade-versus-recovery correlation.
		Sources of residual uncertainty: ■ Three drill holes (GADD-2025-28, GADD-2025-29 and GADD-2025-30) are pXRF-only at this stage; while the parallel-estimate result demonstrates immateriality at deposit scale, the eastern margin where these holes provide preferential support has been classified Inferred via the `rescat_poly` polygon downgrade as a structural hedge. ■ Position uncertainty in the north-south bounding fault wireframe has been managed via the 25 m hard buffer downgrade. Additional fault intercepts in subsequent drilling could permit refinement of the buffer width. ■ The eastern margin of the deposit is currently constrained primarily by surface mapping with limited drilling control. Future drilling east of GADD-2025-14, GADD-2025-29 and GADD-2025-30 could permit upgrade of the polygon-downgraded periphery to Indicated. ■ The Mineral Resource is reported on a global basis at the prospect scale; local-scale tonnage and grade in any specific area of the deposit may differ from the global estimate.

Table 1 – JORC Code 2012

SAB (Sidi Ali Ben Oum Ezzine) – Table 1b

JORC Code, 2012 Edition – Table 1

Section 1 Sampling Techniques and Data

(Criteria in this section apply to all succeeding sections.)

Criteria	JORC Code explanation	Commentary
Sampling techniques	• Nature and quality of sampling (e.g. cut channels, random chips, or specific specialised industry standard measurement tools appropriate to the minerals under investigation, such as downhole gamma sondes, or handheld XRF instruments, etc.).
These examples should not be taken as limiting the broad meaning of sampling.
• Include reference to measures taken to ensure sample representivity and the appropriate calibration of any measurement tools or systems used.
• Aspects of the determination of mineralisation that are Material to the Public Report.
• In cases where ‘industry standard’ work has been done, this would be relatively simple (e.g. ‘reverse circulation drilling was used to obtain 1 m samples from which 3 kg was pulverised to produce a 30 g charge for fire assay’). In other cases, more explanation may be required, such as where there is coarse gold that has inherent sampling problems. Unusual commodities or mineralisation types (e.g. submarine nodules) may warrant disclosure of detailed information.	Four modes of sampling contribute to the SAB Mineral Resource:
• HQ diamond core from nine holes drilled by Chaketma Phosphates SA (CPSA – a joint venture between Celamin Ltd and Tunisian Mining Services) between 2011 and 2013 (referred to as CHDD holes)
• HQ diamond core from six holes drilled by Himilco Pty Ltd (a wholly owned subsidiary of PhosCo Ltd) in 2025 represents a targeted infill of the CPSA-era dataset in Block 6 and Block 7 (referred to as GADD-2025 holes)
• Channel samples from 33 surface trenches cut by CPSA in 2012 and 2013.
• Two trenches (CHT070, CHT071) were mechanically excavated across the full mineralised package; the remainder are narrow diamond-saw channels.

The dataset comprises 443 channel, 312 half-core and 12 quarter-core samples – 767 major-oxide assays across 799.3 m of sampled interval.
• Half-core samples were collected at nominal 1 m lengths (range 0.2–1.55 m), sampling to lithological contacts.
• CHDD samples were jaw-crushed to 2 mm and riffle-split to ~500 g on site by TMS personnel.
• Twelve quarter-core samples from surficial overburden of the early CHDD holes are present, along with additional quarter-core samples prepared for the June 2015 independent re-assay program. |
| Drilling techniques | • Drill type (e.g. core, reverse circulation, open-hole hammer, rotary air blast, auger, Bangka, sonic, etc.) and details (e.g. core diameter, triple or standard tube, depth of diamond tails, face-sampling bit or other type, whether core is oriented and if so, by what method, etc.). | All SAB drilling is HQ-diameter diamond core:
• Nine CHDD holes (2011–13) totalling 509.4 m include seven holes (372.2 m) within the Mineral Resource estimate (MRE). |

Criteria	JORC Code explanation	Commentary
		CHDD drilling was carried out by Tunisian Mining Services (TMS) using a Longyear DB 520 drill rig with a double-tube core barrel and normal-circulation bentonite/water fluids. CHDD holes comprise four vertical and five angled. Within the MRE: four vertical CHDD holes (all 2011 campaign – one each in blocks B2, B4, B5 and B6) and three angled CHDD holes at -70°/-75° dipping 35-50° northeast (2012 campaign – one in B6, two in B7). The two angled CHDD holes in Block 3 external to the MRE (2013 campaign) are retained in the database for transparency. All 18 CHDD collar dips reconcile with downhole survey dips. Six GADD-2025 holes totalling 434.0 m are infill across Block 6 and Block 7: GADD-2025 drilling was conducted by Géoinfra. Drill holes GADD-2025-09 and GADD-2025-10 were completed using HQ single-tube (single core barrel) configuration with a Casagrande C6 rig, with an approximate hole diameter of 96 mm. Drill holes GADD-2025-16, GADD-2025-17, GADD-2025-18 and GADD-2025-19 were completed using HQ double-tube (double core barrel) configuration with a JDL-280 rig, producing core with an approximate diameter of 63.5 mm. Downhole surveys (54 shots across six holes) show maximum deviation of 2.0° from collar dip (range -88.0° to -89.7°), confirming the vertical orientation of the 2025 drilling. Core is not oriented — oriented core is not required for a gently-dipping stratabound target where the intercept angle is controlled by bedding. Holes typically intersect the gently-dipping (10-26° SW) mineralisation at a high angle.
Drill sample recovery	Method of recording and assessing core and chip sample recoveries and results assessed. Measures taken to maximise sample recovery and ensure representative nature of the samples. Whether a relationship exists between sample recovery and grade and whether sample bias may have occurred due to preferential loss/gain of fine/coarse material.	Core recovery data are derived from 342 geotechnical intervals across all 15 drill holes, audited against the SRK-interpreted domain model. Aggregate recovery across the dataset is 95.1%. Recovery within the mineralised package (CLP + CAP + CMP) averages 96.7% across 191.28 m; the primary ore domain CAP averages 97.5% – the best-recovered domain in the dataset – with CLP at 97.1% and CMP at 94.3%. No grade-versus-recovery correlation is evident, and no systematic loss of fine high-grade material has been identified.

Criteria	JORC Code explanation	Commentary
Logging	■ Whether core and chip samples have been geologically and geotechnically logged to a level of detail to support appropriate Mineral Resource estimation, mining studies and metallurgical studies.■ Whether logging is qualitative or quantitative in nature. Core (or costean, channel, etc.) photography.■ The total length and percentage of the relevant intersections logged.	■ Three flagged mineralised intervals below 90% recovery occur within known fault zones and are localised.■ No mineralised interval records over 100% recovery. Recovery data are not available for trenched samples.■ Geological logging covers 373 intervals totalling 1,365.5 m across all 48 drill holes and trenches – 100% coverage of logged length with no gaps or overlaps. Logging used a simplified lithological-group scheme covering the limestone roof, basal mudstone and the three internal phosphate layers.■ Geological logging is qualitative to semi-quantitative, capturing lithology (primary code), sub-lithology and grain size where observed; grain size is recorded for 90 intervals (24%), concentrated in the phosphatic package.■ Geotechnical logging covers 342 intervals across all 15 drill holes, recording recovery, rock quality and rock strength. Structural logging covers 321 point observations, dominantly fractures.■ Holes were logged in their entirety; core boxes are marked with box number, core depths, driller's blocks and sample depths; most core boxes were photographed; contacts between overburden and ore footwall are well defined.■ Geological logging is cross-verified against laboratory major-oxide chemistry following receipt of assays. This ensures geochemical quality assurance and quality control (QAQC) of logged lithology, improving the consistency of lithostratigraphic domains in the MRE.■ Core photography was completed as part of standard logging procedures.Photographs are stored in a centralised file system that includes original drill hole logs.■ Core is photographed in dry condition, after geological description; 100% of drilled core has been photographed.
Sub-sampling techniques and sample preparation	■ If core, whether cut or sawn and whether quarter, half or all core taken.■ If non-core, whether riffled, tube sampled, rotary split, etc. and whether sampled wet or dry.■ For all sample types, the nature, quality and appropriateness of the sample preparation technique.■ Quality control procedures adopted for all sub-sampling stages to maximise representivity of samples.	■ CHDD cores were sawn in half on consistent sides; half-core samples were jaw-crushed on site to 2 mm and riffle-split to approximately 500 g for laboratory dispatch. Final pulverising occurred at the analytical laboratory.■ CHT channel samples (approximately 20 kg/m) were jaw-crushed in their entirety to 2 mm and riffle-split to a 500 g sub-sample.■ GADD-2025 half-core was crushed and riffle-split on site, then dispatched to ALS Spain where the entire submitted sample was crushed to -2 mm via crusher/rotary splitter (70% passing 2 mm); a 250 g split was pulverised to better than 85% passing 75 μm, and the pulps were sealed in double heat-sealed air-evacuated bags.

Criteria	JORC Code explanation	Commentary
	■ Measures taken to ensure that the sampling is representative of the in situ material collected, including for instance results for field duplicate/second-half sampling. ■ Whether sample sizes are appropriate to the grain size of the material being sampled.	■ Sample grain sizes at SAB span coarse coprolitic phosphorite through fine phospharenite to micritic marl; the 1 m nominal sample length is representative for this style of stratabound deposit and consistent with exploration practices used previously at other Gassat prospects. ■ No evidence of grain-size-driven sampling bias has been identified. ■ A QAQC program incorporating commercial pulverised certified reference materials (CRMs), coarse non-certified blanks, and first-division (coarse-reject) duplicates, was progressively developed through the CPSA-era campaigns; field duplicates were introduced partway through the campaigns but were not implemented in the 2013 KEL work. ■ For the GADD sampling campaign, a total of 102 original samples were submitted to ALS Spain for analysis. ■ A total of 14 QAQC samples were inserted, comprising blanks and CRMs (standards) (approximately 7 blanks and 7 standards). ■ This represents an overall QAQC insertion rate of approximately 12%.
Quality of assay data and laboratory tests	■ The nature, quality and appropriateness of the assaying and laboratory procedures used and whether the technique is considered partial or total. ■ For geophysical tools, spectrometers, handheld XRF instruments, etc., the parameters used in determining the analysis including instrument make and model, reading times, calibrations factors applied and their derivation, etc. ■ Nature of quality control procedures adopted (e.g. standards, blanks, duplicates, external laboratory checks) and whether acceptable levels of accuracy (i.e. lack of bias) and precision have been established.	■ Laboratory chemistry comprises 767 major-oxide samples reporting Al2O3, CaO, Fe2O3, K2O, MgO, Na2O, P2O5, SiO2and LOI by XRF, and 354 trace-element samples reporting As, Cd, F, Pb, U, Zn by ICP-OES/ICP-AES. ■ X-ray fluorescence (XRF) is a total-extraction industry-standard method for phosphorite chemistry. Oxide sum statistics (excluding Na2O where absent) – minimum 91.4%, maximum 101.0%, mean 95.9% – are within acceptable bounds. ■ All CPSA-era CHDD and CHT samples (2011–13) were analysed at Al Amri Laboratories (KSA); all GADD-2025 samples were analysed at ALS Spain.
		Campaign
		Laboratory
		Methods
		2011 CHDD
		Al Amri
		XRF major oxide only (batch ALL-11-00605.0)
		2012 CHDD + CHT
		Al Amri
		XRF major oxide (7 batches, .0) + ICP-OES multi-element (5 batches, .1, introduced mid-2012)
		2013 CHDD
		Al Amri
		XRF major oxide + ICP-OES multi-element (batch ALL-13-00052)
		2025 GADD
		ALS
		ME-XRFO6m (Li-metaborate fusion XRF) + ME-ICP61 (HF-HNO3-HClO4 + ICP-AES 34 elements)

QAQC programme architecture:

Criteria	JORC Code explanation	Commentary
		■ A QAQC programme covered three control classes inserted into the SAB sample stream: certified reference materials (CRMs) with staged P2O5values spanning approximately 13–27%, on-site limestone blanks, and field duplicates (quarter-core or parallel-channel splits) for combined sampling and analytical precision. ■ SAB-prospect control populations comprise 26 CRM inserts, 49 blank inserts and 11 field-duplicate pairs across the three campaigns (CHDD 2011–13, CHT, GADD-2025).
		Analytical accuracy: ■ Accuracy on the reporting variables P2O5and CaO is fit for purpose, with critical-failure rates (
		Contamination control: ■ Blank performance is fit for purpose for the GADD-2025 campaign (mean 0.055% P2O5; maximum 0.068%). ■ Historic CHDD blanks and the early CHT blanks (CHT006–CHT038) are statistically indistinguishable populations sourced from on-site limestone with low-level natural phosphorite background (mean 0.15–0.17% P2O5, maximum 0.26%; zero exceedances above 0.5% across n = 26 inserts). These blanks are unfit at the 0.05% tolerance but remain diagnostic for material cross-contamination at any threshold above 0.5%. ■ Blank P2O5in the late-CHT campaign (CHT039–CHT051, n = 13) changes to elevated and more variable values (one insert at 2.08% P2O5, three above 1%); the cause is not formally established (possible blank-material source change or genuine cross-contamination) and is treated as a monitoring limitation for those specific trenches rather than evidence of contamination in routine assays. ■ At thresholds above 0.5% P2O5, 36 of 42 historic blanks (86%) are within tolerance, with non-compliance confined to the late-CHT window. ■ CHT assays show no material bias and acceptable scatter against drill core assays in the principal mineralised domains. This provides independent assurance — outside the blank stream — that the CHT routine assays are not materially contaminated, supporting their inclusion in the MRE despite the blank-material limitation.
		Analytical precision:

Criteria	JORC Code explanation	Commentary
		■ Field-duplicate precision at ore-relevant grade (P2O5>5%) is excellent: 7 of 7 ore-grade pairs satisfy a Half Absolute Relative Difference (HARD) threshold of 20% for P2O5, with maximum HARD of 1.18% and median 0.52%. The ore-grade pool spans three campaigns and three blocks (B1, B6, B7), providing cross-campaign and cross-block evidence that combined sampling and analytical precision at resource grade is consistently good across the SAB sample stream.
		Inter-laboratory check assaying:■ Inter-laboratory check assaying in June 2015 re-assayed 46 CPSA-era quarter-core samples at ALS Spain across nine drill holes, providing cross-laboratory corroboration of the Al Amri stream.■ The Arethuse KEL Mineral Resource Estimate reported a 7% bias in favour of ALS between the two laboratories for major-oxide XRF. Because all SAB CHDD and CHT samples were analysed at Al Amri and all GADD-2025 samples at ALS, the bias potentially projects onto SAB as a systematic offset: CPSA-era major-oxide grades may be approximately 7% lower than if assayed at ALS.■ No retrospective bias correction has been applied. The bias is confined to major-element XRF and does not transfer mechanically to trace-element ICP comparisons. Its direction is conservative for reported grades; cross-campaign comparisons within Block 6 and Block 7 (where CHDD and GADD coexist) may embed a laboratory offset and will be addressed in §3.14 Discussion of relative accuracy/confidence.
Verification of sampling and assaying	■ The verification of significant intersections by either independent or alternative company personnel.■ The use of twinned holes.■ Documentation of primary data, data entry procedures, data verification, data storage (physical and electronic) protocols.■ Discuss any adjustment to assay data.	■ Database construction minimises human handling: field logs and original laboratory certificates were transferred via software to the Gassat database (hosted on Azure).■ Verification for this MRE was conducted through an SRK data audit covering collar, survey, geology, geotechnical, structural, assay, density and QAQC tables.■ The audit resolved a CHDD collar-versus-survey dip discrepancy across five holes.■ No duplicate sample records or sample-depth overlaps were found.■ No twinned holes are present at SAB; the case for twinning is reduced by the dense outcrop/trench/drill network and the consistent vertical grade profile observed between points of observation.■ Geochemical bias between exposed trench channel samples (CHT) versus sub-surface drill hole samples (CHDD, GADD) was assessed per major oxide within each estimation domain (CLP, CAP, CMP) to assess potential surface weathering impacts. The CAP primary-ore domain shows no material P2O5bias between sample modes, while transitional units (CLP, CMP) show minor Fe2O3variation consistent

Criteria	JORC Code explanation	Commentary
Location of data points	Accuracy and quality of surveys used to locate drill holes (collar and down-hole surveys), trenches, mine workings and other locations used in Mineral Resource estimation. Specification of the grid system used. Quality and adequacy of topographic control.	with expected surface weathering. No correction was required, and the MRE treats trench and drill samples as equivalent within each lithology domain. All 15 drill hole collars were surveyed by differential GPS (DGPS) All 33 trench collars were DGPS-surveyed at inflection points project-wide, except 2 at SAB (CHT045, CHT046) which carry handheld-GPS coordinates and have estimated azimuths of 090° – disclosed as reduced-confidence data points. Within the MRE blocks, DGPS collar residuals against the 10 m LiDAR DEM are typically ±1–2 m, consistent with the DEM cell size and expected interpolation error on moderate slopes; the median residual across all 48 sites is 0.57 m. All coordinates are in UTM Zone 32N (WGS84; EPSG:32632). Topographic control derives from an airborne LiDAR survey at approximately 0.2 m native XY sampling. Surveys were resampled to a 10 m grid DEM for the MRE – the 10 m resolution preserves the required topographic profile detail for survey reconciliation and model clipping. The native point cloud is retained as project archive. DGPS outcrop-contact traverses supplement the DEM for geological mapping. Reconciliation of trench surveys against the 10 m LiDAR DEM resulted in the adjustment of five MRE-relevant trenches: CHT045 and CHT046 (B7, handheld-GPS) projected with azimuth of 090° and DEM-draped at 1 m intervals; CHT038 (B7) DGPS stations snapped to DEM and the survey string extended along final bearing; CHT032 (B2) and CHT071 (B6) DGPS station reduced levels (RLs) snapped to DEM values. Lateral DGPS-LiDAR offset (2–3 m) was observed at CHT070 (B6) and CHT041 (B7). DGPS coordinates were retained because both trenches are incised into steep cliff faces which are not honoured by the draped profile which suppressed real topographic relief that needs to be preserved in the model. Downhole surveys are available for all GADD-2025 holes (54 shots); CHDD holes rely on downhole projection along an azimuth and dip measured from the drill rig. Hole depths are shallow (<100 m) and excessive deviation is not regarded as a material factor for the estimate. Downhole deviation surveys were conducted using a Camteq Proshot Dual CTPS200 camera probe instrument. Measurements were taken at 10 m intervals along the drill hole.

Criteria	JORC Code explanation	Commentary
Data spacing and distribution	Data spacing for reporting of Exploration Results. Whether the data spacing and distribution is sufficient to establish the degree of geological and grade continuity appropriate for the Mineral Resource and Ore Reserve estimation procedure(s) and classifications applied. Whether sample compositing has been applied.	The survey data were used to monitor and correct drill hole deviation, ensuring accurate spatial positioning of the drill holes for geological interpretation and resource evaluation. Drill and trench spacing in the classified MRE areas is of the order of 60–150 m, supported by continuous outcrop mapping and 10 m LiDAR. The 2025 GADD-2025 HQ diamond campaign targeted Block 6 and Block 7 as infill to the historical CPSA-era dataset, improving data density in the two largest blocks (containing 83% of the Mineral Resource). The data spacing is sufficient to establish geological and grade continuity consistent with Indicated and Inferred classification Spacing and continuity at SAB are broadly comparable with – and in places tighter than – the data density supporting the 2013 KEL North (180 m drill/90 m channel average, Inferred) and 2013 GSK (400 m drill/350 m channel average, Inferred) Mineral Resources (Geos Mining, 2013).
Orientation of data in relation to geological structure	Whether the orientation of sampling achieves unbiased sampling of possible structures and the extent to which this is known, considering the deposit type. If the relationship between the drilling orientation and the orientation of key mineralised structures is considered to have introduced a sampling bias, this should be assessed and reported if material.	The SAB mineralised horizon is a tabular, gently-dipping stratabound phosphorite (bedding 10–26° SW, block-dependent). Drilling comprises vertical GADD-2025 (six holes, 2025 campaign) and vertical CHDDs (four holes, 2011 campaign), plus inclined CHDDs (five holes at dip -70°/-75°, azimuths 35–50° NE; 2012–13 campaigns). Channel trenches are cut perpendicular to strike and corrected for slope. Vertical drill holes intersect the 10–26° SW-dipping bedding at 64–80° (high angle to the bedding plane). Inclined CHDDs drilled northeast into the southwest-dipping bedding produce shallower intersections (~44° at the steepest Block 6/Block 7 bedding dip). Three inclined CHDD holes fall within MRE blocks (one in Block 6, two in Block 7); the remaining two CHDD holes are in excluded Block 3. Trenches on shallow slopes can introduce lateral bias of sampling within lithological domains, due to a component of along-strike sampling. However, the stratigraphic estimation method and use of proportionally scaled blocked wells accommodates this bias. Faults are predominantly sub-vertical and sub-parallel to drilling direction; outcrop mapping is the principal tool for fault location, as faulting at SAB tends to reduce rather than over-thicken intercepts. No drilling-orientation bias to sampling has been identified.

Criteria	JORC Code explanation	Commentary
Sample security	■ The measures taken to ensure sample security.	Historical (CPSA) exploration data: ■ CPSA-era sample preparation (jaw-crushing and riffle-splitting) was conducted on site by TMS personnel and witnessed by the Competent Person during the Geos Mining site visit of September 2012. ■ CPSA-era materials from 2011–15 (historical samples) including pulp samples, coarse rejects, and paper documentation, are currently stored under the custody of the Tunisian Office National des Mines (ONM) in a secure facility consistent with the Tunisian Mining Code. ■ Himilco has direct access to the project data management platform (OMN system), which has been operational since 2011, without the need to route through ONM for data retrieval or management. Field sampling workflow (2025): ■ Drill core boxes are transported on the same day to the PhosCo depot in Rohia under controlled conditions using a 4x4 field transport system under security supervision. ■ For the current program, GADD-2025 core is in the custody of the drilling contractor until transported to PhosCo's core processing facility at Rohia, at which point control transfers to the Company. ■ Core is logged, sampled and photographed. ■ Following mechanical sample preparation, samples are split into 800 g to 1 kg samples and placed in paper sample bags, which are then placed into 50 L plastic barrels for secure transport. ■ Prepared samples are transported under a documented chain of custody to the airport and subsequently dispatched to the ALS laboratory in Spain. ■ Database integrity supports chain-of-custody confidence.
Audits or reviews	■ The results of any audits or reviews of sampling techniques and data.	Multiple audits and reviews support the SAB Mineral Resource dataset: ■ CPSA-era reviews: Geos Mining MREs at KEL North (March 2013) and GSK (September 2013), with Competent Person Oliver Willetts, MAusIMM; Geos Mining KEL Resource Upgrade Path (January 2014); ■ Arethuse Geology preliminary and detailed audits (January, February and June 2015 – the June 2015 audit including check-logging of 15 holes and re-assay of 46 quarter-core samples at ALS Spain); ■ Arethuse Geology KEL Mineral Resource Estimate (March 2016).

Criteria	JORC Code explanation	Commentary
		■ SRK was engaged in late 2021/early 2022 to recompile the historical CPSA drilling and assay data into the relational Gassat database. ■ Current-program audits (SRK, 2026)include: Database Audit v1.5; Survey-DEM Reconciliation v1.3; SG (Bulk Density) Audit v1.0; Core Recovery Audit v1.0; Grade Validation Summary; ■ SRK 2026 Geochemical Review reviewed the lithogeochemical aspects of the CPSA-era SAB, KEL and GSK prospect data.

Section 2 Reporting of Exploration Results

(Criteria listed in section 1 also apply to this section.)

Criteria	JORC Code explanation	Commentary
Mineral tenement and land tenure status	■ Type, reference name/number, location and ownership including agreements or material issues with third parties such as joint ventures, partnerships, overriding royalties, native title interests, historical sites, wilderness or national park and environmental settings. ■ The security of the tenure held at the time of reporting along with any known impediments to obtaining a licence to operate in the area.	■ The Gassat (Chaketma) Project is held 100% by Himilco Pty Ltd, a wholly owned subsidiary of PhosCo Ltd (ASX:PHO). ■ The current exploration permit was granted on 6 March 2025 for a 3-year term, covering approximately 56 km² across eight phosphate prospects: KEL, GK/GSK, SAB, KEA, DOH, GEZ, KH and KM. ■ The Chaketma exploration licence was first attributed in 2010; held by CPSA (a Celamin/TMS joint venture) from 2011–15; and transferred to Himilco in 2025. ■ The Company is in good standing with the Tunisian Ministry of Industry, Mines and Energy. ■ At this stage of the exploration licence, no joint venture, partnership, or off-take agreements are in place, and no private net smelter return royalties have been identified. The project is held and operated under an exclusive exploration licence with no mining (exploitation) rights granted. ■ Land use is predominantly agricultural and pastoral, with no material constraints identified at this stage. No known environmental, heritage, or land tenure issues are currently considered to be blocking factors, subject to confirmation through required regulatory studies (environmental impact assessments/environmental and social impact assessments, land, and heritage assessments). ■ Any future transition to mining will remain subject to the granting of the relevant mining exploitation permits and regulatory approvals.
Exploration done by other parties	■ Acknowledgment and appraisal of exploration by other parties.	Modern exploration at Gassat has been conducted by two successive operators: ■ CPSA (2011–15): completed the foundational drilling and trenching programs across the project, including the SAB CHDD and CHT campaigns. CPSA’s work underpins the KEL (Geos Mining, 2013; Arethuse, 2016) and GK (Geos Mining, 2013) Mineral Resources. ■ Himilco/PhosCo (2025–present): completed the 2025 GADD-2025 infill drilling at SAB.

Criteria	JORC Code explanation	Commentary
		■ Earlier regional work by CERPHOS and CPG has been referenced in the literature but is not available to PhosCo and does not contribute data to the current MRE (PhosCo KM Table 1 2026).■ Exploration work undertaken by Compagnie des Phosphates de Gafsa and CERPHOS was conducted as part of regional-scale phosphate exploration programs covering the broader Gassaat project area.■ These historical studies were not specifically targeted at the SAB sector, but rather formed part of a wider regional investigation.■ Within the SAB area itself, only limited historical workings are recorded, consisting of two trenches and one underground gallery. These features have been subsequently re-sampled and re-analysed as part of the current exploration program.■ No additional dedicated third party exploration datasets, detailed reports, or standalone studies specific to the SAB sector have been identified or obtained to date.
Geology	■ Deposit type, geological setting and style of mineralisation.	SAB is a stratiform Eocene marine phosphorite deposit of the North African phosphogenic province:■ Mineralisation is hosted in the Chouabine Formation (Late Palaeocene to Early Ypresian) of the Metlaoui Group, deposited in a shallow restricted marine basin around the Kasserine Island palaeohigh under upwelling-driven nutrient supply.■ Deposit type: marine sedimentary phosphorite, consistent with the Gafsa-Metlaoui deposits of central and southern Tunisia.■ Structural setting: SAB sits within the Rouhia Basin – a northwest to southeast collapse graben disrupting the northeast-southwest Hercynian trend of the Tunisian Atlas – interpreted as a pull-apart basin between the KEL and GSK massifs with pronounced east-west to northwest-southeast faulting.■ Chouabine Formation at SAB is a five-unit lithostratigraphic package (base to top): CMR mudstone footwall; CMP marly phospharenite; CAP apatitic phospharenite (primary ore); CLP phosphatic limestone; CLN dolomitic nummulitic limestone hangingwall.■ Package thickness averages 5–15 m across the MRE blocks (maximum 18 m at Block 6).■ Mineralisation is stratabound, compositionally stable (median CAP P2O5 21.9–23.1% across blocks) and laterally continuous.■ Francolite (carbonate-fluorapatite) is the dominant phosphate mineral.

Criteria	JORC Code explanation	Commentary
		■ The Mineral Resource is defined across five fault-bounded blocks (Block 2, Block 4, Block 5, Block 6, Block 7), each a discrete half-graben or sub-basin compartment 200 m to 400 m across. ■ Block 1 and Block 3 are excluded on exploration data and internal structural grounds respectively.
Drill hole Information	■ A summary of all information material to the understanding of the exploration results including a tabulation of the following information for all Material drillholes: - easting and northing of the drillhole collar - elevation or RL (Reduced Level – elevation above sea level in metres) of the drillhole collar - dip and azimuth of the hole - downhole length and interception depth - hole length. ■ If the exclusion of this information is justified on the basis that the information is not Material and this exclusion does not detract from the understanding of the report, the Competent Person should clearly explain why this is the case.	■ The SAB dataset comprises 15 drill holes (943.4 m) and 33 channel trenches (442.6 m) across the prospect. ■ Of these, 13 drill holes (806.2 m) and 25 trenches (381.0 m) fall within the five MRE blocks. ■ All coordinates are in WGS84 UTM Zone 32N (EPSG:32632); elevations are in metres above mean sea level. ■ Six GADD-2025 holes are vertical: orientation confirmed by 54 downhole survey records across the six holes (dip range -88.0° to -89.7°; maximum deviation 2.0° from nominal -90°). ■ Nine CHDD holes comprise four vertical (CHDD008, CHDD009, CHDD010, CHDD012 – all 2011 campaign) and five angled (CHDD041, CHDD046, CHDD048 at dip -70°, azimuths 40–50° NE; CHDD049, CHDD051 at dip -75°, azimuths 35–50° NE – 2012 and 2013 campaigns). ■ Of the 25 MRE trenches, 23 are DGPS-surveyed; CHT045 and CHT046 carry handheld-GPS coordinates and estimated 090° azimuths. ■ Total length for each trench records the sampled length from origin; physical trench traverses measured from DGPS XYZ profile surveys may be longer where sampling was terminated at package boundaries.
Data aggregation methods	■ In reporting Exploration Results, weighting averaging techniques, maximum and/or minimum grade truncations (e.g. cutting of high grades) and cut-off grades are usually Material and should be stated. ■ Where aggregate intercepts incorporate short lengths of high grade results and longer lengths of low grade results, the procedure used for such aggregation should be stated and some typical examples of such aggregations should be shown in detail. ■ The assumptions used for any reporting of metal equivalent values should be clearly stated.	■ Not applicable. This Mineral Resource Estimate supersedes the underlying Exploration Results; data aggregation methods relevant to length-weighted intercept reporting are not material to the disclosure (Joint Ore Reserves Committee (JORC) Code 2012, Clause 27).

Criteria	JORC Code explanation	Commentary
Relationship between mineralisation widths and intercept lengths	■ These relationships are particularly important in the reporting of Exploration Results. ■ If the geometry of the mineralisation with respect to the drillhole angle is known, its nature should be reported. ■ If it is not known and only the downhole lengths are reported, there should be a clear statement to this effect (e.g. ‘down hole length, true width not known’).	■ The SAB mineralised package is gently dipping (10-26° SW, block-dependent). ■ Drilling comprises vertical GADD-2025 holes (six holes, 2025 campaign); vertical CHDD holes (four holes, 2011 campaign); and angled CHDD holes (five holes, 2012-13 campaigns, dip -70° to -75° at azimuths 35-50° NE). Of the five angled CHDD holes, three are in MRE blocks (one in Block 6, two in Block 7) and two are in the excluded Block 3. ■ Vertical drilling in the flatter blocks (Block 2, Block 4, Block 5; below ~20° dip) gives apparent downhole thickness within ~6% of true thickness. ■ Vertical drilling in the steeper blocks (Block 6, Block 7; 25-26° dip) gives apparent thickness exceeding true thickness by ~10%. ■ Angled CHDD holes drilled northeast into the southwest-dipping bedding intersect at shallower angles to the bedding plane (~44° at the steepest Block 6/Block 7 bedding), giving apparent intercept lengths substantially greater than true thickness for the three MRE-relevant intercepts. ■ The geometric effect is resolved by the stratigraphic block-model framework, where model cells respect bedding geometry; intercepts are reconciled to true thickness during estimation and do not flow through to reported block tonnes as a residual bias. ■ Channel trenches are cut perpendicular to strike and corrected for slope during sample collection (PhosCo KM Table 1 2026). Reported channel thicknesses represent true thickness. ■ No artificial width inflation arises from the combination of drilling orientation and mineralisation geometry.
Diagrams	■ Appropriate maps and sections (with scales) and tabulations of intercepts should be included for any significant discovery being reported These should include, but not be limited to a plan view of drillhole collar locations and appropriate sectional views.	■ Not applicable. Diagrams supporting the SAB Mineral Resource Estimate are presented in the companion public report; the criterion is not material to this Table 1 (JORC Code 2012, Clause 27).
Balanced reporting	■ Where comprehensive reporting of all Exploration Results is not practicable, representative reporting of both low and high grades and/or widths should be practiced to avoid misleading reporting of Exploration Results.	■ Not applicable. This Mineral Resource Estimate supersedes the underlying Exploration Results; the criterion is not material to this disclosure (JORC Code 2012, Clause 27).
Other substantive exploration data	■ Other exploration data, if meaningful and material, should be reported including (but not limited to): geological observations; geophysical survey results; geochemical survey results; bulk	■ HiveMap digital geological mapping from high-resolution aerial imagery combined with the 10 m LiDAR DEM provides continuous outcrop-contact traces for the CLN caprock and CAP exposure along block flanks (SAB Geological Model Workflow v1.1).

Criteria	JORC Code explanation	Commentary
	samples – size and method of treatment; metallurgical test results; bulk density, groundwater, geotechnical and rock characteristics; potential deleterious or contaminating substances.	■ SRK's PHS002 geochemical review (January 2026) provides lithofacies characterisation, cross-prospect geochemical profiling and the basis for the SRK-interpreted domain codes. ■ A total of 94 Archimedes water-displacement specific gravity (SG) measurements across nine drill holes support per-lithology bulk density assignment. ■ Trace-element geochemistry for 354 SAB samples indicates cadmium levels comparable with other Tunisian ores and CPG operations, uranium within reasonable levels, and arsenic, lead and zinc at low levels (PhosCo KM Table 1 2026).
Further work	■ The nature and scale of planned further work (e.g. tests for lateral extensions or depth extensions or large-scale step-out drilling). ■ Diagrams clearly highlighting the areas of possible extensions, including the main geological interpretations and future drilling areas, provided this information is not commercially sensitive.	■ SAB-specific metallurgical testwork – sighter-level flotation testing analogous to the PhosCo KM program – targeted confirmation of SAB-specific concentrate grade and recovery. PHS002 noted SAB-specific geochemical distinctions (coarser phospharudite, sharper CLN contact, lower maximum Na2O) that warrant targeted testing. ■ Given the shallow mineralisation, SAB is a candidate for prioritised inclusion in the mine plan. ■ At present, no specific exploration program has been defined or scheduled for the SAB sector. ■ Any decision regarding further exploration work will be contingent upon the results of the evaluation of the Mineral Resources which will determine whether additional exploration is warranted and will inform any future definition of scope, sequencing, and timing of work at SAB.

Section 3 Estimation and Reporting of Mineral Resources

(Criteria listed in section 1, and where relevant in section 2, also apply to this section.)

Criteria	JORC Code explanation	Commentary
Database integrity	■ Measures taken to ensure that data has not been corrupted by, for example, transcription or keying errors, between its initial collection and its use for Mineral Resource estimation purposes. ■ Data validation procedures used.	■ Exploration data are held in a dedicated and secure Microsoft SQL server relational database. ■ The database was originally constructed by Geos Mining under CPSA using automated data import workflows to compile field logs and original laboratory certificates, minimising human handling of assay data. ■ SRK recompiled the CPSA drilling and assay data into the current relational structure in 2021–22 and maintains ongoing custody of the database. ■ Data integrity was audited by SRK for this MRE. ■ There were no duplicate Sample ID records and no sample-depth overlaps identified. All 48 sites reconcile to downstream tables within ±0.5 m.
Site visits	■ Comment on any site visits undertaken by the Competent Person and the outcome of those visits. ■ If no site visits have been undertaken indicate why this is the case.	■ Oliver Willetts, MAusIMM(CP – Geology) 312940, undertook multiple site visits between 2011 and 2013 and witnessed CPSA sampling, drill sites and geological logging. He has visited each of the SAB resource blocks in the MRE. Mr Willetts acts as Competent Person for estimation and reporting of Mineral Resources. ■ Aymen Arfaoui (MAusIMM: ID) is currently employed by PhosCo as Exploration Manager for the Gassat project. Mr Arfaoui designed the SAB exploration program, and has overseen field data acquisition and sampling. Mr Arfaoui visits site regularly and has witnessed both the CPSA and GADD exploration programs. Mr Arfaoui acts as Competent Person for reporting of exploration data, sampling procedures and results.
Geological interpretation	■ Confidence in (or conversely, the uncertainty of) the geological interpretation of the mineral deposit. ■ Nature of the data used and of any assumptions made. ■ The effect, if any, of alternative interpretations on Mineral Resource estimation. ■ The use of geology in guiding and controlling Mineral Resource estimation. ■ The factors affecting continuity both of grade and geology.	■ Geological model is constructed in tNavigator v25.4 using 15 HQ diamond drill holes (943.4 m), 33 channel trenches (442.6 m), HiveMap digital geological surface mapping, the 10 m LiDAR DEM, DGPS outcrop-contact traverses, 767 major-oxide assays and 354 trace-element assays. ■ Mineralisation in each MRE block is limited by interpreted structures representing discrete fault-bounded sub-basins within the Rouhia Graben extensional setting. ■ Block 1 and Block 3 are excluded from the MRE on grounds of insufficient internal data and structural disruption respectively – the two drill holes and eight trenches in those blocks remain in the database for transparency and sample QAQC purposes, but are excluded from the MRE associated analyses.

Criteria	JORC Code explanation	Commentary
		■ The five-unit Chouabine Formation stratigraphy (CLN → CLP → CAP → CMP → CMR) is laterally continuous and strongly stratabound across all five MRE blocks. ■ Assayed oxide lithogeochemistry was used to interpret and to further refine the logged stratigraphic unit contact positions from logged observations. ■ Stratigraphic unit contacts are interpolated as horizon DTM surfaces and validated to ensure optimal honouring of source data. ■ Fault segments lacking confident cross-fault correlation are employed as classification boundaries. Tonnes in fault blocks that currently contain only trenches, or lack data entirely, are downgraded to reflect lower local resource confidence. ■ A 3D stratigraphic geology model exists for each estimation block that includes the geochemically-refined stratigraphic sequence alongside internal and bounding faults. This serves as the input for grade estimation. ■ Minor uncertainty in geometry and locations of interpreted faults in areas of limited exposure is reflected in classification-boundary decisions rather than in the MRE tonnage. ■ No alternative interpretation that would materially change the reported Mineral Resource has been identified.
Dimensions	■ The extent and variability of the Mineral Resource expressed as length (along strike or otherwise), plan width, and depth below surface to the upper and lower limits of the Mineral Resource.	■ The SAB Mineral Resource occupies five fault-bounded blocks across a prospect extent of approximately 2.0 km north-northeast–south-southwest by 1.5 km east-northeast–west-southwest. ■ Mineralised-package thickness varies from 5–10 m in the smaller blocks (Block 4, Block 2, Block 5) to 8–18 m in the larger blocks (Block 6, Block 7, where the maximum for the prospect is reached). ■ Overburden from topographic surface to the top of CLN ranges from zero at the outcropping flanks to a maximum of the order of several tens of metres at the deepest drilled points, supporting a low strip ratio (averaging 2.6, reaching a maximum average of 4.2 at Block 7) consistent with the 2022 Chaketma Scoping Study assumptions at KEL and GK.
Estimation and modelling techniques	■ The nature and appropriateness of the estimation technique(s) applied and key assumptions, including treatment of extreme grade values, domaining, interpolation parameters and maximum distance of extrapolation from data points. If a computer assisted estimation method was chosen, include a description of computer software and parameters used.	Resource estimation was undertaken using tNavigator v25.4 within a stratigraphic block model covering stratigraphic modelling, compositing, variography and grade interpolation. ■ Stratigraphic model: Block-by-block grids on a 10 m × 10 m horizontal cell size, layer structure respecting the five-unit Chouabine stratigraphy. Horizon surfaces were

Criteria	JORC Code explanation	Commentary
	■ The availability of check estimates, previous estimates and/or mine production records and whether the Mineral Resource estimate takes appropriate account of such data. ■ The assumptions made regarding recovery of by-products. ■ Estimation of deleterious elements or other non-grade variables of economic significance (e.g. sulphur for acid mine drainage characterisation). ■ In the case of block model interpolation, the block size in relation to the average sample spacing and the search employed. ■ Any assumptions behind modelling of selective mining units. ■ Any assumptions about correlation between variables. ■ Description of how the geological interpretation was used to control the resource estimates. ■ Discussion of basis for using or not using grade cutting or capping. ■ The process of validation, the checking process used, the comparison of model data to drillhole data, and use of reconciliation data if available.	interpolated by the Minimum Curvature method. Phosphatic stratigraphic units (CLP, CAP, CMP) serve as estimation domains. ■ Compositing: Irregular, proportionally scaled downhole or along-channel composites, domain-respecting, were used, with composite breaks at every lithological contact. There are a total of 645 composites from 13 MRE drill holes and 25 MRE trenches. ■ Variography: Composites are grouped into spatial domains. Variograms were then calculated for each oxide for the various domain combinations. ■ Spatial groups are partitioned into East (Block 2, Block 4, Block 5) and West (Block 6, Block 7) to respect observed differences in along-strike grade continuity. ■ Sills are calculated for each oxide per estimation domain within each spatial group. ■ Nugget contributions were assessed using field duplicate pairs (to maintain lateral, stratigraphic relationships compatible with the modelling search space) and sit below 3% of the total population variance. ■ Modelled ranges were assessed from groups of variograms, displayed together and grouped according to geochemical fractions: a phosphate-carbonate fraction (P2O5,CaO), a dolomitic fraction (MgO) and a siliciclastic fraction (SiO2, Al2O3, Fe2O3and K2O) consistent with the depositional system and oxide correlations. This method proved more robust where individual oxide variograms were poorly structured (mainly CMP domain). ■ Grade interpolation: Block grade estimation via Ordinary Kriging was applied for P2O5, MgO, CaO, SiO2, Al2O3, Fe2O3and K2O. Search neighbourhoods are oriented parallel to stratigraphic bedding using an IJK coordinate system (stratigraphic) non-orthogonal, proportional grid. ■ This modelling strategy honours stratigraphic contacts precisely, decreasing dilution and grade mixing between domains. Lateral grade continuity and vertical grade variation are both effectively reproduced. ■ Block model: There are approximately 183,000 cells in the mineralised domain package across the five MRE blocks. Block-model estimates are validated against composites. Block grade estimation is conducted along stratigraphic layering (IJK coordinate space) into a non-orthogonal grid. ■ No cutting or capping was applied to composite grades. No metal equivalents were reported.

Criteria	JORC Code explanation	Commentary
Moisture	■ Whether the tonnages are estimated on a dry basis or with natural moisture, and the method of determination of the moisture content.	■ Tonnages are reported on an air-dried in situ basis. ■ Bulk density is determined from sealed air-dried cores by Archimedes water displacement. ■ Natural in situ moisture is not a material tonnage factor at MRE stage.
Cut-off parameters	■ The basis of the adopted cut-off grade(s) or quality parameters applied.	■ Mineral Resource is reported at a cut-off grade of 8% P2O5, applied to the mineralised domain package (CLP + CAP + CMP) within the classified MRE blocks. ■ The 8% cut-off is consistent with the natural grade-population break between carbonate/marl waste (predominantly below 5% P2O5) and the mineralised phosphorite package (predominantly above 10% P2O5 through the core, grading to 5–15% at the upper and lower contacts). ■ The cut-off captures the full mineralised package including gradational contact fringes without incorporating waste material. It effectively excludes high MgO hanging wall waste from the Mineral Resources that would be sub-economic to process. ■ The 8% value is consistent with Arethuse's 2016 wireframing practice at KEL and is lower than the 10% cut-off applied by Geos Mining to the 2013 KEL and GK MREs. ■ The more detailed geochemical/lithofacies control from the PHS002 geochemical review enables confident distinction between mineralisation and waste at the gradational CLP-CMP contacts.
Mining factors or assumptions	■ Assumptions made regarding possible mining methods, minimum mining dimensions and internal (or, if applicable, external) mining dilution. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential mining methods, but the assumptions made regarding mining methods and parameters when estimating Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the mining assumptions made.	■ The SAB Mineral Resource is reported with open pit mining as the assumed extraction method. ■ Open pit amenability is supported by the shallow mineralisation depth, outcrop along flanks, and low strip ratio, consistent with the 2022 Scoping Study assumptions at the neighbouring KEL and GK prospects. ■ Continuous tabular stratiform geometry is amenable to bulk or selective mining. ■ No specific pit optimisation has been run against the SAB MRE at this stage. Pit design and mining-factor testing are anticipated during the Scoping Study update integrating SAB and KM alongside the existing KEL and GK resources. ■ Mining-related dilution, ore loss and minimum-mining-width considerations are outside the scope of the Mineral Resource statement.
Metallurgical factors or assumptions	■ The basis for assumptions or predictions regarding metallurgical amenability. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider potential metallurgical methods, but the	■ Metallurgical assumptions currently rely on project-level testwork conducted by CPSA at the KEL and GSK prospects rather than SAB-specific testwork. ■ The 2017 Pilot Plant Report established the baseline reverse flotation flowsheet for the Chaketma phosphorite and demonstrated production of a saleable concentrate

Criteria	JORC Code explanation	Commentary
	assumptions regarding metallurgical treatment processes and parameters made when reporting Mineral Resources may not always be rigorous. Where this is the case, this should be reported with an explanation of the basis of the metallurgical assumptions made.	from KEL and GK feed with acceptable recovery, grade and deleterious-element characteristics. ■ The 2017 Chaketma Pre-feasibility Study (PFS) integrated this flowsheet into a full process design; the 2022 Scoping Study confirmed continuing applicability. ■ SAB phosphorite is geologically continuous with the Chaketma feed material – same Chouabine lithostratigraphy, same domain framework, same dominant mineralogy (francolite in a calcareous matrix) – and is assumed compatible with the established flowsheet at MRE stage. ■ The PHS002 geochemical review notes several SAB-specific characteristics that warrant targeted confirmation in SAB-specific testwork – coarser phospharudite component, sharper CLN contact, and lower maximum Na2O – however, none of these are considered to preclude flowsheet compatibility at MRE stage. ■ Deleterious-element levels at SAB are comparable with other Tunisian phosphate ores and CPG operating ores: ■ MgO is of greatest commercial attention because it can affect beneficiation at concentrations exceeding 6%. MgO averages 5.15% across all domains in the classified Mineral Resources. ■ CAP (3.48% MgO) accounts for ~65% of the Mineral Resources and sits below the 6% threshold; CMP (7.99% MgO) is elevated and accounts for ~26% of the Mineral Resources; CLP (9.11% MgO) contains more dolomite and accounts for ~9% of the Mineral Resources. ■ SRK considers blending of higher MgO domains with CAP represents a pathway to meet processing specifications. Alternately, high impurity regions of CLP and CMP might represent a potential run of mine direct application product. ■ Cadmium levels are elevated and correlate with increasing P2O5 grades (length weighted averages: CLP: 9 ppm, CAP: 49 ppm, CMP: 16 ppm). Decadmiation of phosphoric acid concentrate is one method to control cadmium levels and produce a marketable product.
Environmental factors or assumptions	Assumptions made regarding possible waste and process residue disposal options. It is always necessary as part of the process of determining reasonable prospects for eventual economic extraction to consider the potential environmental impacts of the mining and processing operation. While at this stage the determination of potential environmental impacts, particularly for a greenfields project, may not always be well advanced, the status	Environmental framework assumptions rely on the broader Chaketma Project context established in the 2022 Scoping Study and 2017 PFS. ■ The Chaketma Project area is not identified as environmentally sensitive at regional scale – no national park, Ramsar wetland, World Heritage nor equivalent designation overlaps the permit. The terrain is semi-arid with no major perennial watercourses across SAB.

Criteria	JORC Code explanation	Commentary
	of early consideration of these potential environmental impacts should be reported. Where these aspects have not been considered this should be reported with an explanation of the environmental assumptions made.	Conceptual waste rock and tailings management assumptions are established at project level in the 2017 PFS. SAB-specific characterisation has not been conducted.The area hosts occasional rural population with local-employment participation an engagement theme since the CPSA era. No pastoralist or Indigenous community land tenure conflicts are identified.The Tunisian Mining Code sets the environmental impact assessment pathway for conversion of exploration tenements to mining licences; the framework was positively assessed in the 2022 Scoping Study.No environmental factor identified to date would materially preclude eventual economic extraction.
Bulk density	Whether assumed or determined. If assumed, the basis for the assumptions. If determined, the method used, whether wet or dry, the frequency of the measurements, the nature, size and representativeness of the samples.The bulk density for bulk material must have been measured by methods that adequately account for void spaces (vugs, porosity, etc.), moisture and differences between rock and alteration zones within the deposit.Discuss assumptions for bulk density estimates used in the evaluation process of the different materials.	Bulk density is assigned per lithology from 94 Archimedes water displacement SG measurements across nine SAB drill holes, conducted by TMS personnel on dried unsealed core under procedure PA (TMS, 2013).SG measurements were audited by SRK using a two-stage screen combining physical-plausibility thresholds by lithology and per-sample stoichiometric limits determined using XRF oxide chemistry.SG anomalies were removed from the sample population and did not contribute to the MRE SG determination.Adopted per-lithology bulk density values are: CLP 2.61 t/m3; CAP 2.69 t/m3;CMP 2.70 t/m3.The tight range is consistent with the compositional range of francolite + calcite + dolomite + minor quartz at SAB, and with bulk densities adopted at KEL (Geos Mining, 2013; Arethuse 2016) and GK (Geos Mining, 2013) which ranged from 2.65 t/m3 to 3.18 t/m3 across equivalent domains.
Classification	The basis for the classification of the Mineral Resources into varying confidence categories Whether appropriate account has been taken of all relevant factors (i.e. relative confidence in tonnage/grade estimations, reliability of input data, confidence in continuity of geology and metal values, quality, quantity and distribution of the data).Whether the result appropriately reflects the Competent Person's view of the deposit.	The SAB Mineral Resource is classified into Indicated and Inferred categories under JORC (2012).Mineral Resources are only classified in resource blocks that contain both drilling and outcrop trenches containing economic phosphorite intercepts.Drilling samples the full stratigraphic sequence whereas trench sampling preferences outcropping portions of mineralisation. This is why both sampling modes are required in combination.Resource classification relies on local assessment of grade and geological continuity at each block. The classification considers drilling, trenching, outcrop-mapping, variography, database-integrity and estimation-quality evidence.

Criteria	JORC Code explanation	Commentary
		■ Classification boundaries reflect the Competent Person's view that geological variability carries greater project risk than grade variability – which is quite consistent laterally. This relates to the stable depositional environment of the material combined with the complex structural setting. ■ Indicated classification is applied in each block where drilling and trenching data density is sufficient to establish both geological and grade continuity with confidence. ■ Smooth contours were derived from P2O5kriging variance for the primary CAP domain in Block 6 and Block 7; Block 2, Block 4 and Block 5 are small enough to be classified in their entirety. Practically, this translates to a classification radius around sample locations of 50–80 m. ■ Faulting is acknowledged during classification; internal faults require sampling in each independent fault block and across structures, while bounding faults terminate mineralisation entirely. These structures are clearly identified from digital surface mapping. ■ Inferred classification is applied where geological and grade continuity is established to lower confidence, typically as lower local data density regions (or peripheries) within an otherwise well-controlled block. ■ Smooth contours were derived from P2O5kriging variance for the primary CAP domain in Block 6 and Block 7. Practically, this translates to a classification radius around sample locations out to 120 m. ■ Internal fault blocks containing only trenching were classified as Inferred at Block 4, Block 6 and Block 7. ■ Exploration Targets are defined for material outside the Mineral Resources in Block 6 and Block 7. ■ Fault-bounded compartments within Block 6 and Block 7 where the internal data density is insufficient to support Mineral Resource classification are reported as Exploration Targets.
Audits or reviews	■ The results of any audits or reviews of Mineral Resource estimates.	■ The estimation has not been independently reviewed by a third party. No external audit of the current KM Mineral Resource has been undertaken as at the date of this report.

Criteria	JORC Code explanation	Commentary
		■ SRK has undertaken an internal peer review in line with company quality management protocol..
Discussion of relative accuracy/confidence	■ Where appropriate, a statement of the relative accuracy and confidence level in the Mineral Resource estimate using an approach or procedure deemed appropriate by the Competent Person. For example, the application of statistical or geostatistical procedures to quantify the relative accuracy of the resource within stated confidence limits, or, if such an approach is not deemed appropriate, a qualitative discussion of the factors that could affect the relative accuracy and confidence of the estimate. ■ The statement should specify whether it relates to global or local estimates, and, if local, state the relevant tonnages, which should be relevant to technical and economic evaluation. Documentation should include assumptions made and the procedures used. ■ These statements of relative accuracy and confidence of the estimate should be compared with production data, where available.	■ The accuracy and confidence statements relate to global estimates at the prospect scale. ■ The SAB Mineral Resource is of appropriate accuracy and confidence to support the reported Indicated and Inferred classifications and the supporting Exploration Target. ■ Block-model validation against composites shows block model versus composite mean P2O5 agreement within ±2.3% across all classified blocks for the CAP domain – the primary ore contributor. CMP and CLP validation is acceptable across all blocks at ±10% reflecting reduced sample counts in these domains. ■ Trench-versus-drill grade-bias checks confirmed no material grade bias exists between the two sampling modes. ■ Production reconciliation is not applicable (this is a maiden Mineral Resource). ■ The reported accuracy and confidence is consistent with the Indicated and Inferred classifications assigned.

ASX Release 5 May 2026
Page 58

JORC Code, 2012 Edition – Table 1a - KM Prospect

Appendix B – KM Mineral Resource estimate summary

GASAAT PHOSPHATE PROJECT

Maiden Mineral Resource Estimate

Project overview

KM is the northernmost of eight phosphate prospects within the 56 km² Gassat (Chaketma) Project permit, situated in the Rouhia Basin of central-western Tunisia. The current exploration permit was granted on 6 March 2025 for a 3-year term and is held 100% by Himilco Pty Ltd (Himilco), a wholly owned subsidiary of PhosCo Ltd (ASX:PHO). The Chaketma exploration licence was first granted in 2010 and was held by Chaketma Phosphates SA (CPSA - a Celamin/Tunisian Mining Services joint venture) between 2011 and 2015 prior to transfer to Himilco in 2025.

Exploration database

Modern exploration at KM has been conducted entirely by Himilco/PhosCo in a single exploration campaign between 2025 and 2026.

The estimation dataset (shown in Figure KM.1) comprises 13 HQ-diameter (63.5 mm) diamond drill holes (GADD-2025 series, 818.0 m total) and 5 surface channel trenches (GAT-01 to GAT-05), all completed in a single 2025 campaign by Himilco/PhosCo. All coordinates are in WGS84 UTM Zone 32N (EPSG:32632); data are sourced from the Gassat project database.

Sampling was conducted on a half-core basis at nominal 1 m intervals broken to lithological boundaries. Sample preparation (crushing, splitting, pulverising) was undertaken at the core logging facility; full preparation-chain detail is provided in the supporting JORC Table 1 (Section 1).

All drill holes are essentially vertical (full-hole dip range -84.8° to -90.0°), intersecting the gently west-dipping (10–20° W) mineralised package at a high angle and giving apparent intercept lengths within ~6% of true thickness. Trenches were mechanically excavated perpendicular to stratigraphy and corrected for slope during sample collection; channel widths return true thickness directly.

Geochemical coverage comprises 483 laboratory major-oxide records (assayed at ALS Spain, ME-XRFO6m: P₂O₅, CaO, MgO, SiO₂, Al₂O₃, Fe₂O₃, K₂O, Na₂O, LOI¹) across nine drill holes and all five trenches, plus 118 portable x-ray fluorescence (pXRF) point-set records on three drill holes (GADD-2025-28, GADD-2025-29, GADD-2025-30) where laboratory assays are currently pending. Trace elements (As, Cd, Cr, Pb, U, Zn) are supplied through the same laboratory records by ME-ICP61. The estimation lithostratigraphy model is made from the SRK-interpreted lithology dataset, refined from field logs to the seven-unit stratigraphy via lithogeochemical cross-verification.

SRK audited and validated the database prior to estimation.

¹ LOI – loss on ignition

Figure KM.1: Plan view of KM stratigraphic domains. exploration and cross-sections

Sources: SRK 2026

Geological Setting

KM is a stratiform Eocene marine phosphorite deposit of the Gafsa-Metlaoui style, forming part of the broader North African phosphogenic province. Mineralisation is hosted within the Chouabine Formation (Late Palaeocene to Early Ypresian) of the Metlaoui Group, deposited in a shallow restricted marine basin developed around the Kasserine Island palaeohigh under upwelling-driven nutrient supply (Garnit et al., 2017). Francolite (carbonate-fluorapatite) is the dominant phosphate mineral.

The Chouabine Formation at KM exhibits a seven-unit lithostratigraphic package, from base (deepest, oldest) to top (shallowest, youngest):

The mineralised package has a median preserved thickness of approximately $30\mathrm{m}$ and reaches a maximum of $56.4\mathrm{m}$ in GADD-2025-03 – substantially thicker than the other Chaketma prospects, indicating better local preservation and trap configuration.

Structurally, KM sits on the western edge of the Rouhia Basin and forms a single, northeast-trending, fault-bounded graben. The deposit is bounded to the west by an east-dipping ( $\sim 30^{\circ}$ ) north-south striking listric normal fault, and to the south by an east-west trending fault. Internally, two east-west faults segment the

deposit into the Northern, Central and Southern sub-blocks, with strike-slip and normal displacement. Closely spaced drilling has confirmed that mineralisation is sharply sheared and cleanly offset across these internal east-west structures with no fault thickening; stratigraphic continuity across the sub-blocks supports modelling and estimation as a laterally continuous volume.

Geology model

Geological modelling and grade estimation were undertaken in tNavigator v25.4. Seven stratigraphic horizon surfaces were built from the SRK lithogeochemical dataset to create laterally continuous stratigraphy across the deposit's extent.

Trench data (GAT-01 to GAT-05) and stratigraphic contacts drafted from high resolution LiDAR² digital elevation model (DEM) topography and aerial imagery were used for outcrop and shallow-section control, allowing confident control of unit volumes along the eastern and northern flanks. Faults were digitally mapped from the aerial surveys and were used to structurally constrain modelled units.

A 3D stratigraphic grid model (shown in plan view in Figure KM.1 and in cross section in Figure KM.2) was produced from the horizon surfaces extending 700 mN × 450 mE. The grid is structurally conformant and non-orthogonal – vertical cell thickness follows the local layer geometry which underpins a structural flattening coordinate transform to improve lateral continuity during grade estimation, a critical property of the KM mineralisation.

Figure KM.2: Geological model cross section showing stratigraphy

Sources: SRK Consulting 2026
Notes: Cross-section location shown in Figure KM.1

Grade estimation

Grade estimation was performed in IJK space – the structurally flattened coordinate system established by the non-orthogonal grid. This method was selected to maximise the observed lateral along-bedding grade continuity witnessed between intercepts, and also to preserve downhole cross-bedding grade variability to produce a representative estimate (see Figure KM.3).

Page 60

Estimation runs were performed per domain across the four mineralised units (CLP, CDP, CAP, CMP) and per oxide across the seven major-oxide suite (P₂O₅, MgO, CaO, SiO₂, Al₂O₃, Fe₂O₃, K₂O), giving 28 independent estimates. The CLN, CLC and CMR units were not estimated and are treated as waste.

Compositing within each domain generated multiple variable-length intervals (averaging approximately 1 m length) according to the local vertical grid cell heights at each drill hole and trench location. Composite lengths therefore inherit the layer geometry of the geological model rather than being regularised to a fixed downhole length. Per-domain composite support reflects the relative thickness and lateral extent of each unit.

Block grades were interpolated by Ordinary Kriging, using variogram models fitted manually per domain and per oxide against the variable composite length experimental variograms. The CAP P₂O₅ variogram (the primary resource domain) is spherical with ranges 120 m × 80 m × 3.5 m with a principal direction of 0°. Modelled nuggets are extremely low, effectively zero in most cases. A maximum of 20 composites contribute to each block estimate. Negative kriged estimates (uncommon, concentrated on trace oxides) have been set to zero post-estimation.

Bulk density is applied as a single per-domain mean (CLP 2.59 t/m³, CDP 2.58 t/m³, CAP 2.67 t/m³, CMP 2.72 t/m³) sourced from sealed Archimedes water-displacement measurements. SRK reviewed measured densities in a two-stage process to confirm initial physical-plausibility, then subjecting measurements to stoichiometric evaluation using XRF major-oxide chemistry.

The Mineral Resource is reported above an 8% P₂O₅ per-cell cut-off, with topographic clipping to the LiDAR DEM applied post-estimation to retain the contribution of surface trench data and outcrop samples to estimation before final volume reporting.

Estimate validation

Block model grade estimates were validated globally against composite inputs across the four estimated domains (CLP, CDP, CAP, CMP) for the seven-oxide suite. Local validation was undertaken using swath plots along easting, northing and elevation, alongside visual inspection of grade alignment on cross section.

Key findings:

In all four mineralised domains, the global composite mean P₂O₅ differs from the block model mean P₂O₅ by less than 10%.
Block grade variances are lower than composite grade variances in all four domains – the expected smoothing pattern for kriged estimates and consistent with the modelled variogram structure.
Pre-infill/post-infill differences in estimated contained CAP P₂O₅ following the GADD-2025-28, GADD-2025-29 and GADD-2025-30 drilling were less than 0.2%; the infill drilling confirmed the pre-existing block model in stratigraphic structure and grade, demonstrating the robustness of the geological model and interpretation.
Testing the effect of pXRF on the global estimated grade confirmed a global difference on contained CAP P₂O₅ of less than 0.05% between the headline grade estimate (assay + pXRF) and an assay only run. These additional estimates confirmed that the pXRF-supported drill holes do not materially change the resource at deposit scale.
Field duplicate quality assurance and quality control (QAQC) reveals excellent reconciliation between parent and duplicate assays (globally for Gassat and at locally at KM – although the duplicate population is currently low) with extremely low nuggets (near zero), consistent with the bulk stratiform character of the material.
Core recovery within the mineralised package averages 95.7% across geotech-logged drill metres, with no domain-specific recovery bias and no grade-versus-recovery correlation.

The validation supports the use of the block model for Mineral Resource reporting at the Indicated and Inferred classification levels.

Figure KM.3: Cross section showing stratigraphy, $\mathsf{P}_2\mathsf{O}_5$ , block grades and modelled structures

Sources: SRK Consulting 2026
Notes: Cross-section location shown in Figure KM.1

Classification and reporting

The KM Mineral Resource is classified as Indicated and Inferred in accordance with the JORC Code (2012 Edition). Classification is applied uniformly to all four estimation domains, reflecting the operational reality that they would be potentially mined together within the same vertical sequence (Figure KM.4).

High lateral grade continuity is observed between points of observation, within each stratigraphic domain extending across fault blocks, over the full prospect.

The primary factor affecting geological continuity is the structural configuration of mineralisation. Drilling and mapping confirm that mineralisation is sharply sheared and cleanly offset across east-west internal faults with grade continuity preserved either side (e.g. GADD-2025-28). This eliminates the requirement for buffering or classification downgrades near east-west faulting.

Indicated classification reflects the high degree of internal geological and grade consistency observed between points of observation, the demonstrated continuity of mineralisation across the east-west internal faults (Fault_3 and Fault_6), and the robustness of the geological model to the addition of the three latest infill drill holes (when the new drilling was added to the existing model, $< 0.2\%$ movement resulted on contained $\mathrm{CAPP_2O_5}$ ). Indicated regions of the model are defined where drilling, trenching or mapped mineralisation in outcrop are spaced $100\mathrm{m}$ apart or less.

Graben-bounding faults along the western, eastern and southern peripheries locally control mineralisation thickness. These structures result in downgrade of mineralisation to Inferred using two different approaches according to location:

West/south: a $25\mathrm{m}$ hard buffer along the north-south bounding fault wireframes, reflecting positional uncertainty in the wireframe (single direct intercept in GADD-2025-15) and the sharp but finite transition zone observed as north-south graben-controlling faults terminate mineralisation.
East: a classification downgrade polygon covering the lower confidence eastern margin (east of GADD-2025-14 and GADD-2025-29/ GADD-2025-30; the southeastern tip of the North sub block), routed deliberately between GADD-2025-29 and GADD-2025-30 to leave both holes available for

future laboratory assay driven re-evaluation. The polygon also captures the pXRF-influenced eastern margin where GADD-2025-29 and GADD-2025-30 dominate grade support.

Inferred Mineral Resources are classified out to a maximum 150 m from points of observation, although most blocks are classified between 100 m and 120 m.

Figure KM.4: KM Mineral Resource classification

Sources: SRK, 2026

Reasonable Prospects for Eventual Economic Extraction

The KM Mineral Resource is reported on an open pit basis. Open pit amenability is supported by shallow mineralisation depth (surface to base of mineralisation typically less than $100\mathrm{m}$ , with several drill hole intercepts commencing at or near surface) and a low strip ratio (range from 0 to 1.6, average 0.4 – see Figure KM.5), consistent with the 2022 Scoping Study assumptions at the neighbouring KEL and GSK deposits. The continuous tabular stratiform geometry is amenable to bulk or selective mining.

The $8\%$ $\mathrm{P}_2\mathrm{O}_5$ reporting cut-off was selected to capture the natural grade-population break between waste (carbonate/marl predominantly below $5\%$ $\mathrm{P}_2\mathrm{O}_5$ ) and the mineralised phosphorite package (predominantly above $10\%$ $\mathrm{P}_2\mathrm{O}_5$ through its core). The cut-off captures the full mineralised package, including gradational contacts, without incorporating waste material, and concurrently screens out the higher-MgO portions of the package described above.

Adjacent deposits within the same depositional system support the 8% value: it matches the cut-off applied in the Arethuse (2016) KEL Mineral Resource Estimate and is lower than the 10% cut-off applied by Geos Mining to the 2013 KEL and GSK Mineral Resource Estimates, reflecting the more detailed geochemical and stratigraphic control supplied by the SRK 2026 geochemical review which enables confident distinction between mineralisation and waste at the gradational hanging wall CLP and footwall CMP contacts.

KM phosphorite is geologically continuous with the wider Gasaat feed material from the KEL, SAB and GK deposits and is assumed compatible with the reverse flotation flowsheet established in the 2017 Pilot Plant Report and integrated into the 2017 Chaketma Pre-feasibility Study and the 2022 Scoping Study update. Sighter metallurgical testwork on KM core is currently underway and is focused on the mineralogy and metallurgical characteristics of the KM phosphate sub-units to inform the optimal processing route.

MgO is the controlling deleterious element for Gassat phosphorite under reverse-flotation processing and averages 5.16% across the classified Mineral Resource. The CAP unit (3.48% MgO) sits within typical flowsheet specifications and accounts for 65% of the Mineral Resource; CMP and CLP carry higher dolomite content. Selective mining of CAP, feed sequencing or blending each represent viable pathways to manage MgO within the Chaketma flowsheet performance envelope demonstrated in the 2017 Pilot Plant Report. Higher-MgO material may alternatively be amenable to direct application phosphate product routes, which carry less stringent MgO tolerance than flotation feed.

Cadmium levels are elevated and correlate with increasing P₂O₅ grade (length-weighted averages: CLP 7.3 ppm, CDP 21.5 ppm, CAP 39 ppm, CMP 19.1 ppm); decadmiation of phosphoric acid concentrate is one established pathway to control cadmium and produce a marketable product.

The Gassat permit area is not identified as environmentally sensitive at regional scale. Site-specific mining, metallurgical and infrastructure studies at KM remain to be completed and are anticipated under the Scoping Study update integrating KM and SAB alongside the existing KEL and GK resources.

ASX Release 5 May 2026
Page 64

Figure KM.5: Strip ratio for combined KM estimation domains

Sources: SRK, 2026

Recommendations

The following actions are recommended to advance the KM Mineral Resources:

Complete laboratory assays, geological/geotechnical/structural logging, and DGPS surveys on the three pXRF-only drill holes (GADD-2025-28, GADD-2025-29, GADD-2025-30). Receipt of laboratory assays for GADD-2025-29 and/or GADD-2025-30 is expected to result in a local upgrade in the Mineral Resource classification over the corresponding area.
Advance sighter metallurgical testwork on KM core to inform the optimal processing route and product-specification compliance.
Undertake additional definition along the eastern margin of the Central block, east and north of GADD-2025-14, GADD-2025-29 and GADD-2025-30. This region is the principal candidate for classification upgrade and requires a small shallow infill campaign. Additional drilling north of GADD-2025-13/GADD-2025-15 is a parallel candidate for upgrade in the North sub-block.
Undertake additional exploration over the wider KM prospect, especially to the west of the current Mineral Resources and to the immediate north of hole GADD-2025-08, which was excluded from the resource estimate due to failure to intercept mineralisation.
Increase the field duplicate QAQC sample population to at least 5% of total sampling in future exploration programs.

ASX Release 5 May 2026
Page 66

ASX Release 5 May 2026
Page 67

Appendix C – SAB Mineral Resource estimate summary

GASAAT PHOSPHATE PROJECT

Maiden Mineral Resource Estimate

Project overview

The Gassat Project hosts eight phosphate prospects across an approximately 56 km² exploration permit in central-western Tunisia, approximately 200 km southwest of Tunis. The project is held 100% by Himilco Pty Ltd (Himlico), a wholly owned subsidiary of PhosCo Ltd (ASX:PHO).

The SAB prospect is situated between the larger Kef El Louz (KEL) and Gassaa Kebira (GSK) deposits which were the subject of historical Mineral Resource Estimates by previous operators (Geos Mining, 2013; Arethuse, 2016).

The current exploration permit was granted on 6 March 2025 for a 3-year term. The Gassat (formerly Chaketma) exploration licence was first attributed in 2010 and held by Chaketma Phosphates SA (CPSA – a Celamin and Tunisian Mining Services joint venture) between 2011 and 2015; the licence transferred to Himilco in 2025.

Exploration database

Modern exploration at SAB has been conducted by two successive operators in three campaign periods between 2011 and 2025 (Figure SAB.1). CPSA completed the foundational diamond drilling (CHDD series, 2011–13) and surface trenching (CHT series, 2012–13); Himilco completed a targeted infill diamond drilling program (GADD-2025 series) across Block 6 and Block 7 in 2025 (Table SAB.1).

Table SAB.1: SAB exploration data by resource block

Block	CHDD	GADD	CHT	Total
Block 2	1	-	5	6 sites
Block 4	1	-	3	4 sites
Block 5	1	-	4	5 sites
Block 6	2	1	4	7 sites
Block 7	2	5	9	16 sites
MRE total	7 / 372.2 m	6 / 434.0 m	25 / 381.0 m	38 / 1,187.2 m

Source: SRK, 2026

All SAB drilling is HQ-diameter diamond core. The six GADD-2025 holes are vertical and intersect the 10–26° southwest-dipping bedding at 64–80° (high angle to the bedding plane). The nine CHDD holes comprise four vertical (2011 campaign – one each in Block 2, Block 4, Block 5, Block 6) and five angled at -70°/-75° dipping northeast (2012–13 campaigns – one in Block 6, two in Block 7, plus two in the excluded Block 3). Intercepts are reconciled to true thickness during estimation.

Channel trenches are cut perpendicular to strike along outcropping. Two trenches in Block 6 (CHT070 and CHT071) were mechanically excavated across the full mineralised package; the remainder of the trenches are narrower diamond-saw channels.

A QAQC program incorporating commercial pulverised certified reference materials (CRMs), coarse non-certified blanks, and first-division (coarse-reject) duplicates, was progressively developed through the CPSA-era campaigns and continued through the GADD-2025 program.

Sampling was conducted on a half-core basis at nominal 1 m intervals broken to lithological contacts. Sample preparation (jaw-crushing to 2 mm, riffle-splitting to ~500 g) was undertaken on site, with final pulverising at the assay laboratory; full preparation-chain detail is provided in the supporting JORC Table 1 (Section 1).

Figure SAB.1: Plan view of SAB block locations, stratigraphic domains and exploration

Source: SRK, 2026
Notes: Cross section locations shown by A-AFigure SAB.2 and BB` Figure SAB.3

Geochemical coverage comprises 767 laboratory major-oxide assays (P₂O₅, CaO, MgO, SiO₂, Al₂O₃, Fe₂O₃, K₂O, Na₂O, LOI³ by XRF⁴ – the total-extraction industry-standard for phosphorite chemistry) and 354 laboratory trace-element assays (As, Cd, F, Pb, U, Zn by ICP-OES/ICP-AES). All CPSA-era samples were assayed at Al Amri Laboratories (Kingdom of Saudi Arabia) and all GADD-2025 samples at ALS Chemex (Spain) using methods ME-XRFO6m and ME-ICP61.

A 7% inter-laboratory bias in favour of ALS for major-oxide XRF was reported by Arethuse (2016) for the adjacent KEL deposit. CPSA-era SAB assays share the same analytical laboratory and potentially the same bias whereby CHDD and CHT sample major-oxide grades may be approximately 7% lower than if assayed at ALS – a direction conservative for reported grades.

The estimation lithostratigraphy model is built from the SRK-interpreted lithology dataset, refined from field logs to the five-unit Chouabine stratigraphy via lithogeochemical cross-verification. Core recovery within the mineralised package averages 96.7% across 191.3 m, with the primary CAP unit recovery at 97.5%.

SRK audited and validated the database prior to estimation and found no material issues.

ASX Release 5 May 2026
Page 68

Geological setting

SAB is a stratiform Eocene marine phosphorite deposit of the North African phosphogenic province. Mineralisation is hosted within the Chouabine Formation (Late Palaeocene to Early Ypresian) of the Metlaoui Group, deposited in a shallow restricted marine basin around the Kasserine Island palaeohigh under upwelling-driven nutrient supply (Garnit et al., 2017). Francolite (carbonate-fluorapatite) is the dominant phosphate mineral, hosted in a calcareous matrix consistent with the wider Gafsa-Metlaoui phosphorite province of central and southern Tunisia.

The Chouabine Formation at SAB comprises a five-unit lithostratigraphic package: the CMR mudstone footwall, overlain by the CMP marly phospharenite, the CAP apatitic phospharenite (the primary ore unit), the CLP phosphatic limestone, and capped by the CLN dolomitic nummulitic limestone hanging wall. The CAP, CMP and CLP units are mineralised; the CMR and CLN units are treated as waste. The mineralised package averages 5–15 m thickness across the MRE blocks and reaches a maximum of 18 m in Block 6 – substantially thicker than at the neighbouring KEL and GSK deposits. Median CAP P₂O₅ grades range narrowly between 21.9% and 23.1% across the five MRE blocks, reflecting strong stratabound compositional stability.

SAB sits within the Rouhia Graben – a northwest-southeast collapse graben disrupting the regional northeast-southwest Hercynian trend of the Tunisian Atlas – interpreted as a pull-apart basin between the KEL and GSK massifs and dissected by northwest-southeast, northeast-southwest and east-west trending normal and transtensional faults. Five fault-bounded sub-basin compartments host the Mineral Resource (Block 2, Block 4, Block 5, Block 6, Block 7), each 200–400 m across. Bounding faults terminate the mineralised package; internal faults preserve continuity of the package across the structures. Block 1 and Block 3 have been excluded from the Mineral Resource on grounds of insufficient internal data and structural disruption respectively.

Geology model

The geological model is constructed in tNavigator v25.4 (Rock Flow Dynamics) for each of the five MRE blocks. Stratigraphic horizons are interpolated as digital terrain model (DTM) surfaces respecting the five-unit Chouabine package, and refined by lithogeochemical cross-verification of the logged lithology dataset against laboratory major-oxide chemistry to produce the SRK-interpreted lithostratigraphy used in domaining.

Surface inputs comprise HiveMap digital geological mapping of contacts and structures from high-resolution aerial imagery, the 10 m LiDAR DEM, and DGPS outcrop-contact traverses. These inputs constrain the model along the outcropping flanks of the tilted blocks where stratigraphic contacts are directly observable. Within each block, mapped contact data are used to honour the surface geometry, supplemented by drill markers for sub-surface control.

Each block within the model contains a stratigraphic hierarchy designed to anchor off the most abundant stratigraphic data; the ranking system ensures conformation between layers (Figure SAB.2). Faults with sufficient internal data are used to cut the structural grid; faults lacking confident cross-fault correlation are used as classification boundaries between Mineral Resource and Exploration Target categories.

The 2025 GADD-2025 infill drilling program at SAB was executed in two stages following an early campaign geological model update. The first hole at Block 7 (GADD-2025-09) intersected previously unmapped faulting and returned no mineralised major-oxide assays – the hole did not provide useful grade information for the resource estimate. Detailed high-resolution aerial mapping of the overlying CLN limestone contact – a reliable surface indicator of the underlying phosphate extent across the Gassat project – was used to revise the Block 7 structural framework and to guide location of the remaining 2025 drill holes.

ASX Release 5 May 2026
Page 69

Four subsequent Block 7 holes were positioned against the revised model and successfully intersected mineralisation, materially expanding the Block 7 dataset. A single GADD-2025 hole at Block 6 was drilled to the south of the existing data and tied the central block to the surrounding trench data, extending the volume that satisfies the Indicated classification criteria.

Figure SAB.2: Geological model cross section A-A* (Block 2) showing stratigraphy

Source: SRK, 2026
Notes: Cross section location shown in Figure KM.1.

The five estimated resource blocks were each modelled individually, with a combined model extent spanning approximately $0.8\mathrm{km}$ north-south by $1.3\mathrm{km}$ east-west across the five MRE blocks. A $10\mathrm{m} \times 10\mathrm{m}$ horizontal cell size with proportional vertical layering was used to respect the stratigraphic package. The grid is structurally conformant – non-orthogonal – with cells dimensioned to local layer thickness rather than absolute elevation. This grid framework supports IJK estimation in the structurally flattened coordinate system, described under 'grade estimation' below.

Grade estimation

Estimation domains are the three mineralised units (CLP, CAP, CMP) within each of the five MRE blocks. Seven major oxides – $\mathsf{P}_2\mathsf{O}_5$ , MgO, CaO, $\mathsf{SiO}_2$ , $\mathsf{Al}_2\mathsf{O}_3$ , $\mathsf{Fe}_2\mathsf{O}_3$ and $\mathsf{K}_2\mathsf{O}$ – are estimated as independent variables within each domain. The CLN and CMR units are not estimated and are treated as waste.

Compositing was applied as variable-length blocked composites at each drill hole and trench location, dimensioned to local vertical grid cell heights and broken at every lithological contact. The total composite count is 645 across 13 MRE drill holes and 25 MRE trenches. Composite lengths inherit the local grid layer geometry rather than being regularised to a fixed downhole length, ensuring each composite represents the local layer's full vertical extent and remains compatible with the IJK estimation framework.

Composites are partitioned into Eastern (Block 2, Block 4, Block 5) and Western (Block 6, Block 7) spatial groups to respect observed differences in along-strike grade continuity during geostatistical analysis. The seven oxides cluster geochemically into a phosphate-carbonate fraction $(\mathsf{P}_2\mathsf{O}_5,\mathsf{CaO})$ , a dolomitic fraction $(\mathsf{MgO})$ and a siliciclastic fraction $(\mathsf{SiO}_2,\mathsf{Al}_2\mathsf{O}_3,\mathsf{Fe}_2\mathsf{O}_3,\mathsf{K}_2\mathsf{O})$ , consistent with the depositional system and previous geochemical studies.

Block grade estimates were obtained by Ordinary Kriging using isotropic single-structure spherical variograms fitted per-domain multiplied by per-oxide.

Modelled relative nuggets sit below 3% of population variance across all domain-oxide combinations, derived from project-wide field duplicate half-variance analysis. Fitted variogram structures support a single range per spatial domain and oxide fraction (75–150 m in CLP and CAP, and 75–100 m in CMP). Variogram ranges were taken from the major direction, which carried more data pairs because of the elongate basin geometry.

Search neighbourhoods are oriented parallel to stratigraphic bedding using the IJK grid, decreasing dilution and grade mixing across bedding within domains. Negative kriged grades – which were uncommon – are set to zero post-estimation.

Bulk density is assigned per lithology – CLP 2.61 t/m³, CAP 2.69 t/m³, CMP 2.70 t/m³ – from 94 Archimedes water-displacement measurements across nine SAB drill holes. SRK reviewed measured densities in a two-stage process to confirm the initial physical-plausibility, and then subjected measurements to stoichiometric evaluation using XRF major-oxide chemistry. The narrow observed range across mineralised lithologies is consistent with SAB mineralogy and aligns with bulk densities adopted at the adjacent KEL and GSK deposits.

The 8% P₂O₅ cut-off is applied on a per-cell basis after estimation, and the block model is clipped against the 10 m LiDAR DEM topographic surface post-estimation to retain trench and outcrop contributions to the resource volume.

Estimate validation

The block model was validated globally against input composites across all five MRE blocks and the three estimated domains (CLP, CAP, CMP) for the full seven-oxide suite. The validation supports the use of the block model for Mineral Resource reporting at the Indicated and Inferred classification levels.

Composite-versus-block grade differences for the primary CAP unit are within ±2.3% across all five classified blocks. Differences in the CLP and CMP units are larger but remain within ±10% across all blocks, consistent with the smaller composite counts in these thinner transitional units.
Block grade variances are lower than composite grade variances in all classified blocks for the CAP unit, with variance ratios of 0.4–0.8 – the expected smoothing pattern for kriged estimates and consistent with the modelled variogram structure.
Trench-versus-drill geochemical bias was assessed per major oxide within each estimation domain to test for surface weathering effects. The CAP unit shows no material P₂O₅ bias between trench and drill samples; minor Fe₂O₃ variation in the transitional CLP and CMP units is consistent with expected surface weathering. No correction was applied; trench and drill samples are treated as equivalent within each lithology domain.
Field duplicate variance across the wider Gassat dataset returned relative nuggets below 3% of population variance, indicating that measurement and sampling error contribute negligibly to total variance.
Core recovery within the mineralised package averages 96.7% across 191.3 m, with the CAP unit at 97.5%. Three sub-90% intervals within mineralised ground are confined to known fault zones and have not introduced a recovery-driven sampling bias.

Figure SAB.3: Cross section B-B (Block 6) showing stratigraphy, $\mathsf{P}_2\mathsf{O}_5$ block grades and modelled structures

Source: SRK, 2026
Notes: Cross section location shown in Figure KM.1.

Classification and reporting

The SAB Mineral Resource is classified as Indicated and Inferred in accordance with the JORC Code (2012 Edition). Classification is applied uniformly across the three estimated mineralised units (CLP, CAP, CMP), reflecting the operational reality that they would be potentially mined together within the same vertical sequence. The classification is shown in Figure SAB.4.

High lateral grade continuity is observed between points of observation within each stratigraphic domain, extending across internal fault blocks and over the full prospect. Classification boundaries reflect the Competent Person's view that geological variability – driven by the structurally complex setting – carries greater project risk than grade variability, which is laterally consistent across the deposit.

Mineral Resources are classified only in resource blocks containing both drilling and outcrop-trench data: drilling samples the full stratigraphic sequence, while trench sampling preferentially samples mineralisation along its exposed flank, and both modes are required in combination.

Indicated regions of the model are defined where drilling, trenching or mapped mineralisation in outcrop are spaced up to $100\mathrm{m}$ apart. Structural setting is also considered and the classification is occasionally extended beyond the data spacing within well-constrained internal fault compartments. Block 2, Block 4 and Block 5 outcrop on three flanks and are sufficiently small to be classified in their entirety where adequate data support exists.

Internal fault compartments containing only trench data and/or outcrop mapping have been classified Inferred at Block 4, Block 6 and Block 7, reflecting reduced confidence in grade continuity across structures.

Inferred Mineral Resources are classified out to a maximum $120\mathrm{m}$ from points of observation in Block 6 and Block 7. Beyond this distance, fault-bounded compartments lacking sufficient internal data are reported as Exploration Target.

Figure SAB.4: SAB Mineral Resource classification

Source: SRK, 2026

Exploration Target

The maiden SAB Mineral Resource is supported by a 1.0–1.5 Mt at $19 - 22\%$ $\mathsf{P}_2\mathsf{O}_5$ Exploration Target reported separately for fault-bounded compartments within Block 6 (0.7–1.0 Mt) and Block 7 (0.3–0.5 Mt) where internal data density is insufficient to support Mineral Resource classification (Figure SAB.4).

The potential quantity and grade of the Exploration Target is conceptual in nature: there has been insufficient exploration to estimate a Mineral Resource for this material and it is uncertain if further exploration will result in the estimation of a Mineral Resource.

The Exploration Target is based on actual Exploration Results – a narrow grade envelope reflecting compositional stability of the stratabound phosphorite (CAP grade varies only between $23.9\%$ and $24.2\%$ $\mathrm{P}_2\mathrm{O}_5$ across the five MRE blocks), and a wider tonnage envelope structural uncertainty extrapolating away from informing data.

The continuous mapped presence of the overlying CLN unit across the designated Exploration Target is a consistent stratigraphic indicator of the underlying mineralisation across the wider Gassat project and supports the assumed continuity.

No specific exploration program has been formally defined or scheduled at this stage; any decision regarding further work will be contingent on Mineral Resource model evaluation outcomes.

Reasonable prospects for eventual economic extraction

The SAB Mineral Resource is reported based on extraction by open-pit mining. Open-pit amenability is supported by shallow mineralisation depth, continuous tabular mineralisation, outcrop along the block flanks, and a low strip ratio averaging 2.6 (block tonnage-weighted) across all 5 blocks (Figure SAB.5).

SAB phosphorite is geologically continuous with the wider Gassat feed material from the Kef El Louz, Sidi Ali Ben Oum Ezzine and Gassaa Kebira deposits and is assumed compatible with the reverse flotation flowsheet established in the 2017 Pilot Plant Report and integrated into the 2017 Chaketma Pre-feasibility Study and the 2022 Scoping Study update. SRK recommends SAB-specific sighter-flotation testwork in the future.

MgO is the controlling deleterious element for Gassat phosphorite reverse flotation feed and averages $5.15\%$ across the classified Mineral Resource. The CAP unit $(3.48\% \mathrm{MgO})$ accounts for approximately $65\%$ of the Mineral Resource and is within typical flowsheet specifications; the CMP unit $(7.99\% \mathrm{MgO})$ is elevated at approximately $26\%$ of the Mineral Resource and would require blending; the CLP unit $(9.11\% \mathrm{MgO})$ carries a more pronounced dolomite component but represents only approximately $9\%$ of the Mineral Resource.

Figure SAB.5: SAB strip ratio for combined estimation domains

Source: SRK, 2026

Selective mining of CAP, feed sequencing or blending each represent viable pathways to manage MgO to meet processing specifications. Higher-MgO material may alternatively be amenable to direct application phosphate product routes, which carry less stringent MgO tolerance than flotation feed.

Cadmium content correlates with $\mathrm{P}_2\mathrm{O}_5$ grade (CLP 9 ppm; CAP 49 ppm; CMP 16 ppm length-weighted); decadmation of phosphoric acid concentrate is one established pathway to control cadmium and produce a marketable product. Other trace elements (uranium, arsenic, lead, zinc) are within ranges typical of Tunisian phosphorite.

The Gassat permit area is not identified as environmentally sensitive at regional scale. Site-specific mining, metallurgical and infrastructure studies at SAB remain to be completed and are anticipated under the Scoping Study update integrating KM and SAB alongside the existing KEL and GSK resources.

Recommendations

The following actions are recommended to advance the SAB Mineral Resources:

Advance SAB-specific sighter metallurgical testwork on retained core to confirm flotation behaviour and concentrate specification compliance, addressing the SAB-specific geochemical distinctions identified in the PHS002 geochemical review.
Undertake infill diamond drilling and channel trenching within the B6_EW1 fault compartment and beyond the Block 7 bounding structures to test geological continuity and convert Exploration Target material to Inferred Mineral Resource, contingent on resource model evaluation outcomes.
Undertake integrated mine planning work for SAB within the forthcoming Gasaat Scoping Study update, alongside KM and the existing KEL and GSK Mineral Resources, to test SAB's contribution to project economics.

ASX Release 5 May 2026
Page 75

AI assistant

PHOSCO LTD — Capital/Financing Update 2026

65559_rns_2026-05-04_ebeffac0-ae5c-47cb-bfa4-7da847d8d72e.pdf

Gasaat Phosphate Project

Maiden resources on new deposits boost scale and economic outlook

HIGHLIGHTS

SAB Mineral Resource Statement

Gasaat Project Geology

Next Steps

Previously released ASX Material References that relate to Gasaat Phosphate Project

COMPETENT PERSON'S STATEMENT

KM

SAB

Previously Reported Results

Gasaat Exploration History

KM (Kodiet Mssafin) Prospect – Table 1a

JORC Code, 2012 Edition – Table 1

Section 1 Sampling Techniques and Data

Section 2 Reporting of Exploration Results

Section 3 Estimation and Reporting of Mineral Resources

SAB (Sidi Ali Ben Oum Ezzine) – Table 1b

JORC Code, 2012 Edition – Table 1

Section 1 Sampling Techniques and Data

Section 2 Reporting of Exploration Results

Section 3 Estimation and Reporting of Mineral Resources

JORC Code, 2012 Edition – Table 1a - KM Prospect

Appendix B – KM Mineral Resource estimate summary

GASAAT PHOSPHATE PROJECT

Maiden Mineral Resource Estimate

Project overview

Exploration database

Geological Setting

Geology model

Grade estimation

Estimate validation

Key findings:

Classification and reporting

Reasonable Prospects for Eventual Economic Extraction

Recommendations

Appendix C – SAB Mineral Resource estimate summary

GASAAT PHOSPHATE PROJECT

Project overview

Exploration database

Geological setting

Geology model

Grade estimation

Estimate validation

Classification and reporting

Exploration Target

Reasonable prospects for eventual economic extraction

Recommendations

More from PHOSCO LTD

PHOSCO LTD Capital/Financing Update 2026