Appen

ASX Release

22 May 2026

2026 Annual General Meeting - CEO address and presentation

Appen Limited (ASX: APX) provides the attached CEO address and presentation to be delivered at today's Annual General Meeting commencing at 10.00am AEST.

Authorised for release by the CEO and Managing Director of Appen Limited.

For further information, please contact:
Investor Relations
[email protected]
+612 9468 6300

About Appen

Appen is a global market leader in data for the AI Lifecycle. With 30 years of experience in data sourcing, data annotation, and model evaluation by humans, we enable organisations to launch the world's most innovative artificial intelligence systems.

Our expertise includes a global crowd of more than 1 million skilled contractors who speak over 500 languages¹, in over 200 countries², as well as our advanced AI data platform. Our products and services give leaders in technology, automotive, financial services, retail, healthcare, and governments the confidence to launch world-class AI products.

Founded in 1996, Appen has customers and offices globally.

Appen Limited, 9 Help Street, Chatswood, NSW 2067, Australia – ACN 138 878 298

¹ Self-reported.
² Self-reported, includes territories.

Appen

CEO's Address and Presentation
2026 Annual General Meeting
22 May 2026

Good morning everyone, and thank you for joining us today.

I will cover three areas this morning. First, the AI market and Appen's role in it. Second, our FY25 results. Third, our growth strategy and guidance for FY26.

Appen's Role in the AI Ecosystem

Let me start with the market.

Appen plays a critical role in the AI ecosystem by providing high quality data that is used to build and monitor AI models.

Model Performance Driven by Bespoke, Human-Generated Data

AI development depends on three inputs: algorithms, compute, and data.

Algorithms are the way that models are configured. The approach is highly complex, and is how researchers are able to define new techniques for model development. You can think of this as the blueprint for AI development

Compute what is required to train and run the models. This is the raw horsepower that it is large demand across model builders.

The third ingredient is data. These are the examples that are used to inform the models how to behave, and it increasingly has the largest influence on how models perform.

Appen's role in the AI ecosystem is to provide the world's leading AI builders with the high-quality, human-generated data they need to build and improve their models.

To understand why that role matters, it helps to look at what data is actually available to model builders.

There are three types.

Public data is the text, images, and audio that has been collected from the internet. The world's leading AI labs have largely exhausted this source. It is already captured in existing models, and training on it provides little ability to differentiate.

Appen

Synthetic data, generated by other AI models, carries a different set of risks. Models trained on it can struggle to solve novel or real-world situations, and in some cases it leads to model collapse, where model quality degrades over successive generations.

That leaves real-world, human-generated data. This is the input that brings genuine expertise, diverse perspectives, and real-world nuance into model performance. It is the only data type that can push a model into new capabilities and new domains.

This is not a temporary gap. Every new capability a model needs, from reasoning and domain expertise to multi-lingual fluency and physical world understanding, requires fresh human data to unlock it.

Appen Creates High Quality Data That Is Used to Build the World's Best AI

Appen's role is to create High Quality Data That Is Used to Build the World's Best AI

The models that power the products billions of people use — the assistants, the search engines, the coding tools, the translation services — were built with data that companies like Appen provided.

Appen Has Evolved to Support Complex Data Requirements

AI training is not a singular task. The techniques and approaches required to build and improve a model are varied, numerous, and constantly changing as the field evolves.

Teaching a model to write better code requires human experts reviewing its outputs and explaining what is right or wrong.

Teaching a model to understand Japanese or Arabic requires native speakers providing examples in those languages.

Teaching a robot to pick up objects requires humans demonstrating the task and having every movement carefully labelled.

Each application requires a different approach, different skills, and different quality standards. And as AI evolves, new techniques emerge that we need to support.

Appen supports the full set of human data techniques used in AI development today. We have deep expertise in delivering high-quality data across each of them. And we continue to evolve our capabilities as the demands of our customers change. That ability to move with the market is one of the things that sets us apart.

Appen

AI Megatrends Driving Demand for Appen's Services

Three structural trends are driving sustained demand in the AI data market.

First is consumer AI globalisation. The global digital advertising market is approximately $740 billion. As AI-powered products are deployed across different countries and cultures, they need human data to reflect local nuance. That work requires Appen's global workforce.

Second is enterprise AI adoption. McKinsey estimates $2.6 to $4.4 trillion in annual value potential from enterprise AI use cases. Embedding domain expertise into enterprise models requires specialist human data.

Third is new form factors. Humanoid and industrial robotics is projected to be a $60 to $100 billion market by the early 2030s. Models that interact with the real world need training data that reflects it.

The important point is that data demand scales with model deployments. As AI adoption grows, so does the demand for what we do.

Appen Is Winning in Complex Areas of Generative AI

Let me give some concrete examples of where we are winning in FY26.

The first example is supporting robotics. We annotated egocentric human video to build datasets for physical AI training, delivering over 50,000 units of data across annotation and evaluation workflows.

The second is related to verifying coding outputs from AI models. We worked with a leading coding foundation model to evaluate coding outputs, identifying vulnerabilities and errors. We have a direct relationship with their research and technical leaders, and the program has expanded multiple times since an initial signing in Q1.

The third is related to enterprise AI solutions. We developed reinforcement learning environments across 20 domains, delivering 1,000 synthetic artifacts, with domain experts validating outputs to remove errors and hallucinations.

The fourth is AI for music. We evaluated the ability for a model to generate music, spanning multiple countries and genres.

The final example is supporting the development of a multi-lingual speech model. We delivered a data across approximately 60 languages for a leading AI lab. That program has since been extended into its next phase.

What is also worth noting is that physical AI, coding models, and reinforcement learning are among the fastest-growing and most heavily invested areas in AI right now.

Appen

The fact that we are not only winning initial projects in these areas but expanding them is a strong signal. It tells us that our customers value what we deliver, and that we are well positioned to grow with these markets as they scale.

Appen Is Well Positioned to Capture Growth at a Global Scale

We have a deep track record supporting the globalisation of AI models for consumer-facing applications. We have been doing this for 30 years, and the next wave of large-scale AI data demand is in areas where that experience is directly relevant.

We are combining that 30-year legacy with new technical capabilities required for the next phase of AI. In the last 12 months we have added more than 20 industry experts with deep technical backgrounds across AI and machine learning. These are people who understand our customers' problems from the inside, and they are directly contributing to the new and expanding projects we are winning.

Our market position in China is one of strength. Appen China revenue is significantly larger than the leading local Chinese listed competitor, and we are the vendor of choice for top Chinese technology companies and model builders.

Finally, we are built for scale. Our products and processes are built to adapt with customer requirements, and technology is central to how we continue to drive operational efficiencies.

FY25 Results

Now let's turn to our financial results for FY25.

FY25 Highlights

FY25 was a year of meaningful progress for Appen.

At the group level, we delivered $230.8 million in revenue, up 4.5% on FY24 when you exclude the impact of Google.

Growth was driven predominantly by new project wins and expansions in generative AI. Generative AI is the clear growth driver for the market and a positive signal that we are executing well against our strategy.

On profitability, we delivered $12.2 million in underlying EBITDA, excluding FX.

Appen

The full-year margin was 5.3%, with a strong end to the year with Q4 coming in at 18.2% EBITDA margin.

We also continue to capture operational efficiencies through technology innovation and automation.

Now looking at our two segments:

On Appen Global—the full year revenue was $127.9m, which was down YoY. However we ended the year on a high note, with Q4 revenue of $41.4 million, up 56% on Q3, and EBITDA of $10.2 million at a 24.6% margin.

Appen China had an exceptional year, delivering $102.9 million in revenue, up 75% YoY, with EBITDA up 640% to $10.6m. Growth being driven by new and expanding generative AI-related projects.

Across the group, 44.1% of Q4 revenue came from GenAI—up from 34.8% a year ago.

Finally, we closed the year with US$59.8 million in cash on hand. We continue to drive operational efficiencies across the businesses. In FY25 we realised $10 million in annualised cost efficiencies through technology and automation

In summary, FY25 was a year that demonstrated real progress — revenue growth, improving margins and exceptional performance from Appen China.

FY25 Group Financial Performance

As mentioned group revenue was $230.8 million for the full year. Gross margin improved to 40.3%, up 100 basis points on FY24, driven by a higher mix of generative AI projects across both divisions.

Underlying EBITDA before FX was $12.2 million, up 251% on the prior year.

Appen Global: Strong finish in Q4

Appen Global revenue for FY25 was $127.9 million, down 21% on FY24 excluding Google, reflecting lower project volumes in Q1 through Q3.

Q4 told a different story. Revenue of $41.4 million was up 56% on Q3, driven by new project wins. Underlying EBITDA for Q4 was $10.2 million, reflecting a 24.6% EBITDA margin for the quarter.

We also executed $10 million in annualised cost efficiencies across Appen Global through automation and operational improvement.

Appen

Appen China: $102.9M Revenue for FY25, 75% Year on Year Growth

Appen China grew 75% in FY25, delivering $102.9 million in revenue. Growth was driven by new and expanding LLM-related projects, including work supporting the international expansion of Chinese technology customers.

The profitability improvement was significant. Underlying EBITDA before FX was $10.6 million for the full year, up 640% on FY24, reflecting a 10.3% EBITDA margin. Q4 FY25 delivered a record $4.3 million, at a 13.5% margin.

Growth Strategy

With that financial context, let me turn to our growth strategy and what we are focused on in FY26.

2026 Focus: Continued Execution to Capture Growth in Our Core Market

Our FY26 strategy is focused on four priorities.

First, data quality. This remains the north star across our operations, technology, and talent. Quality is what drives repeat business and program expansion with our customers.

Second, customer growth. We are focused on market segments with the highest account potential, predominantly hyperscalers and newer foundation model builders. We have more than 20 industry experts added to our team in the last 12 months.

Third, new data segments. We are expanding into additional modalities and techniques through co-innovation with customers. Coding is a high-priority growth area for FY26.

Fourth, operational efficiency. We are continuing to embed AI-led efficiencies across our operations, which I will cover on the next slide.

Strong Focus on AI Enabled Operational Efficiencies

On operational efficiency, the progress we have made is tangible and I want to walk through it.

In project delivery, we have built custom tools to access data faster, deployed AI quality agents to improve annotation accuracy, and automated workflows to reduce time to delivery.

Appen

In workforce operations, we use AI-assisted interviewing to validate contributor performance and provide real-time feedback, improving onboarding speed and contributor quality.

Across the organisation, company-wide AI usage and increasingly agentic operations are reducing operating costs and accelerating our output.

We continue to invest heavily in how we use AI across every part of the business, and we see significant potential for further efficiencies in many areas. This remains an active and high-priority area of focus for us.

FY26 Outlook and Guidance

I will now cover our outlook and guidance for FY26.

FY26 Outlook & Guidance

We remain confident in the AI data market and in Appen's ability to meaningfully contribute to the development of leading foundation models.

We continue to see positive signals on LLM-related growth from both Appen Global and Appen China customers. Tight cost controls remain in place, in keeping with our focus on managing costs in line with the revenue opportunity.

As in previous years, Appen Global revenue is predominantly project-based and seasonality skews revenue to H2. Shareholders should expect that pattern to continue.

Considering this, Appen reaffirms the following FY26 guidance: revenue of $270 million to $300 million, and underlying EBITDA margin before FX of approximately 5 to 10%.

Thank You / Closing

The AI data market is large and growing. We are laser focussed on the fundamentals of delivering high quality data for our clients, and evolving as their data needs change.

I'm very confident in our ability to capture growth in this fast paced market.

Thank you for your continued support. I will now hand back to Vanessa.

2 CEO Address

Ryan Kolln

Appen's role in the AI ecosystem

Appen

Model performance driven by bespoke, human-generated data

AI building blocks
Compute	Data	Algorithms and Research
Public data	Appen
Real-world data	Synthetic data
Largely exhausted, already captured in existing models
Provides little ability to differentiate	Unique data that enables new AI approaches
Brings real-world human data perspectives and expertise into model performance	Reliant on other models to produce
Does not solve new or novel situations
Can result in model collapse

Appen creates high quality data that is used to build the world's best AI

Appen

Appen has evolved to support complex data requirements

LLM Training Data
Supervised Fine Tuning (SFT), RLHF & preference annotation, Chain of Thought (CoT) trajectory labeling, pre-training data production

Multimodal Data
Text-to-image annotation, aesthetic scoring, video labeling, embodied intelligence data

Speech & Audio
Text-to-speech collection, Automated Speech Recognition (ASR) evaluation, dialect & multilanguage audio across 60+ languages

Domain Expert Annotation
Medical, scientific (PhD-level math, biology), legal & financial

Model Evaluation
LLM & Vision Language Models (VLM) benchmarking, GUI agent eval, search quality

Computer Vision & Physical AI
Autonomous driving, robotics, smart home, AR/VR, embodied AI

RL environments
Data and environments to support reinforcement learning training techniques

Off-the-Shelf Datasets
Pre-built datasets: dialects, minor languages, image editing, world model data

Platform & Tooling
On-premise annotation platform deployments & SaaS licensing for enterprise AI teams

AI megatrends driving demand for Appen's services

Consumer AI globalisation

KEY ECONOMIC DRIVER

~$740B global digital advertising market¹

IMPLICATIONS FOR HUMAN DATA

Human data needed to align models with multi-country cultural nuance

Enterprise AI adoption

KEY ECONOMIC DRIVER

$2.6T-$4.4T annual potential across enterprise use cases²

IMPLICATIONS FOR HUMAN DATA

Human data needed to incorporate domain and enterprise expertise into AI models

New form factors and applications

KEY ECONOMIC DRIVER

Humanoid and industrial robotics could be $60-100B market by early 2030s³

IMPLICATIONS FOR HUMAN DATA

Human data needed to provide examples and evaluate models interacting with real world

Data demand scales with model deployments

Dentsu 2026 forecast spend on digital advertising 2. McKinsey: Economic potential of Generative AI 3. Morgan Stanley.

Appen

Appen is winning in complex areas of Generative AI

Examples of Appen projects started in 2026

| New team in existing frontier lab customer

Robotics annotation
- Annotation of egocentric human videos to create action-labelled datasets for training Physical AI
- Evaluation of teleoperated robot performance by scoring task execution across key dimensions such as motion efficiency, grasp accuracy, and object manipulation
- 50,000+ units of data across the annotation and evaluation workflows so far | First project with foundation model builder

Coding vulnerability testing
- Worked with leading coding foundation model to evaluate coding outputs
- Direct relationship with researchers and technical leaders
- Multiple project expansions since initial signing in Q1 | First project with foundation model builder

Reinforcement learning environments
- Leveraged Appen RL environment pipeline to create diverse tasks across 20 domains
- Delivered 1,000 synthetic artifacts covering 20 domains
- Leveraged domain experts identify and fix hallucinations, duplications, and structural issues in preferred model-generated documents. | Expansion of existing project

Multi-modal evals, focus on music generation
- Large scale comparison of music model generation
- Project spanning multi-countries to bring global diversity into model feedback | Program extension with AI lab

Multi-lingual audio model
- Supporting leading research team with multi-lingual language model
- Successfully delivered first phase across ~60 languages to support model launch
- Program recently extended into next phase |
| --- | --- | --- | --- | --- |

Appen is well positioned to capture growth at a global scale

Deep track record supporting the globalization of B2C based AI models

Next wave of large AI data demand is in Appen's area of expertise.

Newer competitors struggle with large scale global operation.

Infusing 30-year expertise with new tech-forward capabilities

Combining Appen's 30-year legacy and expertise with deep technical capabilities required for the next phase of AI.

Strong market position in China

Appen China revenue significantly larger than established local competitor¹.

Vendor of choice for top Chinese technology companies and model builders.

Scalable and robust technology platforms

Products and processes built to scale and adapt with customer requirements.

Technology key to continued operational efficiencies.

Data Ocean (688787.S5:SHH)

Appen

FY25 results

Appen

FY25 highlights

$230.8 million revenue in FY25

Up 4.5% on FY24, excluding the impact of Google
Strong end to the year, predominantly from new projects and expansions in generative AI related projects

$12.2 million underlying EBITDA¹ (ex FX) in FY25

5.3% EBITDA margin for the year, with a strong Q4 of 18.2%
Gross margin improvement driven by greater mix of generative AI projects
Operational efficiencies achieved via technology innovation and automation

Append Global turnaround continues

$127.9m revenue for FY25, down 21%² vs FY24
$5.8m underlying EBITDA (before FX) for FY25, down 36.5% on FY24
Strong Q4 driven by ongoing success in winning generative AI related projects
20+ industry experts added in last 12 months

Appen China growth acceleration & momentum

$102.9m FY25 revenue, up 75% on FY24
$10.6m underlying EBITDA (before FX) for FY25, up 640% on FY24, reflecting 10.3% EBITDA margin for the full year
Growth predominantly driven by new and expanding LLM related projects

Continued success in high-growth Generative AI related projects

Majority of revenue growth and margin expansion from new and expanding generative AI projects
44.1% of Q4 FY25 revenue from GenAI, up from 34.8% in Q4 FY24

Strong cash balance remains

Cash on hand as at 31 December 2025 of $59.8m (AUD 89.5 m³)
$10.0m annualised cost efficiencies (net of talent upgrades) executed in FY25, achieved via tech innovation and automation
Underlying EBITDA excludes restructure costs, transaction costs, and acquisition-related and one-time share-based payment expense.
Excludes the FY24 impact of Google contract termination

FY25 Group financial performance

Group revenue (\$M)¹

$230.8 million revenue for FY25, up 4.5% on FY24¹
Q4 FY25 growth on prior quarter for both Appen Global and Appen China

Group Gross Margin² %

40.3% gross margin for FY25, up 100 bps on FY24
Margin improvement driven by growth in high priority generative AI projects across both Appen Global and Appen China

Group underlying EBITDA³ before FX (\$M)

$12.2 million underlying EBITDA before FX for FY25, up 251% on FY24
5.3% EBITDA margin for the year, with a strong Q4 of 18.2%
Q4 FY25 highest EBITDA quarter since 2021, driven by strong project wins and ongoing efficiencies in the business

1 Excludes the FY24 impact of Google contract termination.

2 Gross margin refers to revenue less crowd expenses.

3 Underlying EBITDA excludes restructure costs, transaction costs, and acquisition-related and one-time share-based payment expense.

Appen

Appen Global: strong finish in Q4, 24.6% EBITDA margin

Appen Global revenue ($M)

$41.4 million revenue for Q4 FY25, up 56% on Q3 FY25
$127.9 million revenue for FY25, down 21%¹ vs FY24 due to lower volumes than expected for Q1 FY25 to Q3 FY25

Appen Global EBITDA ($M)

$5.8 million underlying EBITDA (before FX) for FY25, down 36.5% on FY24
Expanding gross margins from an increase in generative AI projects
$10 million annualised cost efficiencies achieved across Appen Global
Strong Q4 FY25 performance, delivering $10.2 million underlying EBITDA (before FX), reflecting a 24.6% EBITDA margin for the quarter
Excludes the FY24 impact of Google contract termination.

Appen

Appen China: $102.9m revenue for FY25 - 75% year on year growth

Appen China revenue ($M)

$102.9 million revenue for FY25, up 75% on FY24
Growth predominantly driven by new and expanding LLM related projects, including supporting international expansion for Chinese tech customers
Growth continued throughout Q4 FY25, with December annualised revenue exceeding $135 million
Strong market position continues, Appen China revenue significantly larger than an established local Chinese listed competitor¹

Appen China EBITDA ($M)

$10.6 million underlying EBITDA (before FX) for FY25, up 640% on FY24, reflecting 10.3% EBITDA margin for the full year
Record profit performance in Q4 FY25, delivering $4.3 million underlying EBITDA (before FX), reflecting a 13.5% EBITDA margin for the quarter
Improving gross margins from greater mix of higher-margin generative AI projects and increased revenue from high-margin prebuilt datasets
Capturing scaling efficiencies due to tight opex controls as revenue expands
Data Ocean (688787.S5:SHH)

Appen

Growth Strategy

Appen

2026 focus: continued execution to capture growth in our core market

1. Data quality

Relentless pursuit of high data quality, the north star for all areas of operations, technology and talent

2. Customer growth

Hyper go-to-market focus on market segments with highest account potential, predominantly hyper-scalers and newer foundation model builders

3. New data segments

Expand into data modalities and techniques through co-innovation with customers. High focus on complex domains including coding

4. Operational efficiency

Continued AI-led efficiencies across operations

Appen

Strong focus on AI enabled operational efficiencies

	Sample initiatives	Benefits
Project automations	AI quality agents Custom data platforms Custom MCPs built to access data	Faster project outcomes and increased data quality
Workforce automations	Automated ticket responses AI interviews to validate performance Real-time performance feedback	Improved contributor experience and onboarding speed
Internal efficiencies	AI led product development Company-wide Claude usage Increasingly agentic operations	Faster operations with reduced operating costs

Appen

FY26 Outlook and Guidance

Appen

FY26 outlook & guidance

The Company remains confident on the AI data market, and the potential for Appen to meaningfully contribute to the development of leading foundation models.

The Company continues to see positive signals on LLM related growth including from Appen Global and Appen China customers.

Tight cost controls remain in place, in keeping with the Company's focus on managing costs in line with the revenue opportunity.

As in previous years, Appen Global revenue continues to be mostly derived from project-based work and seasonality skews revenue to H2.

Considering this, Appen reaffirms the following FY26 group guidance:

Revenue of $270 - $300 million; and

Underlying EBITDA¹ (before FX) margin of ~5-10%

Underlying EBITDA excludes restructure costs, transaction costs, and acquisition-related and one-time share-based payment expense.

Appen

Thank you

Appen

AI assistant

APPEN LIMITED — AGM Information 2026

64403_rns_2026-05-21_76efd486-774e-4cc5-aa52-a6b407656a1f.pdf

ASX Release

2026 Annual General Meeting - CEO address and presentation

About Appen

Appen's Role in the AI Ecosystem

Model Performance Driven by Bespoke, Human-Generated Data

Appen

Appen Creates High Quality Data That Is Used to Build the World's Best AI

Appen Has Evolved to Support Complex Data Requirements

Appen

AI Megatrends Driving Demand for Appen's Services

Appen Is Winning in Complex Areas of Generative AI

Appen

Appen Is Well Positioned to Capture Growth at a Global Scale

FY25 Results

FY25 Highlights

Appen

FY25 Group Financial Performance

Appen Global: Strong finish in Q4

Appen

Growth Strategy

2026 Focus: Continued Execution to Capture Growth in Our Core Market

Strong Focus on AI Enabled Operational Efficiencies

Appen

FY26 Outlook and Guidance

FY26 Outlook & Guidance

Thank You / Closing

2 CEO Address

Appen's role in the AI ecosystem

Model performance driven by bespoke, human-generated data

Appen creates high quality data that is used to build the world's best AI

Appen has evolved to support complex data requirements

AI megatrends driving demand for Appen's services

Consumer AI globalisation

Enterprise AI adoption

New form factors and applications

Appen is winning in complex areas of Generative AI

Appen is well positioned to capture growth at a global scale

FY25 highlights

$230.8 million revenue in FY25

$12.2 million underlying EBITDA¹ (ex FX) in FY25

Append Global turnaround continues

Appen China growth acceleration & momentum

Continued success in high-growth Generative AI related projects

Strong cash balance remains

FY25 Group financial performance

Appen Global: strong finish in Q4, 24.6% EBITDA margin

Appen Global revenue ($M)

Appen Global EBITDA ($M)

Appen China: $102.9m revenue for FY25 - 75% year on year growth

Appen China revenue ($M)

Appen China EBITDA ($M)

Growth Strategy

Appen

2026 focus: continued execution to capture growth in our core market

1. Data quality

2. Customer growth

3. New data segments

4. Operational efficiency

Strong focus on AI enabled operational efficiencies

FY26 Outlook and Guidance

FY26 outlook & guidance

More from APPEN LIMITED

APPEN LIMITED AGM Information 2026