Hong Kong Exchanges and Clearing Limited and The Stock Exchange of Hong Kong Limited take no responsibility for the contents of this announcement, make no representation as to its accuracy or completeness and expressly disclaim any liability whatsoever for any loss howsoever arising from or in reliance upon the whole or any part of the contents of this announcement.

范式

PHANCY

Phancy Group Co., Ltd.

範式智能技術集團股份有限公司

(formerly known as “北京第四範式智能技術股份有限公司 Beijing Fourth Paradigm Technology Co., Ltd.”)

(A joint stock company incorporated in the People's Republic of China with limited liability)

(Stock Code: 6682)

ANNOUNCEMENT OF ANNUAL RESULTS FOR THE YEAR ENDED DECEMBER 31, 2025

The Board hereby announces the consolidated annual results of the Group for the year ended December 31, 2025, together with the comparative figures for the year ended December 31, 2024. The consolidated annual results of the Group for the Reporting Period have been audited by the auditor of the Company and reviewed by the Audit Committee.

FINANCIAL HIGHLIGHTS
	Year ended December 31,		Year-on-year change
	2025	2024
	Amount (RMB'000)	Amount (RMB'000)
Revenue	7,135,286	5,260,650	35.6%
Gross profit	2,483,436	2,244,829	10.6%
Non-IFRS measures:
Adjusted net profit/(loss)	6,330	(292,470)	Turnaround to profit
Adjusted net profit/(loss) attributable to owners of the Company	17,842	(264,991)	Turnaround to profit

BUSINESS REVIEW AND OUTLOOK

Overall Performance

In 2025, the Company delivered an encouraging set of results. Building on its first-ever achievement of adjusted net profitability, its operating revenue continued to experience rapid growth, reaching a record high.

In terms of revenue, our total revenue for the full year reached RMB7.135 billion, a year-on-year increase of 35.6%. This growth rate was significantly higher than the industry average and that of the same period last year, achieving a re-acceleration of revenue growth. This growth reflects the comprehensive and vigorous development across all of the Company's business lines, while benefiting from the continuous promotion of the national "AI+" action policy and the strong release of demand for domestic substitution.

In terms of profitability, the adjusted net profit attributable to the parent company reached RMB17.84 million, and the Company successfully achieved full-year profitability. This is an important turning point — it signifies that Phancy has entered a new cycle of parallel revenue growth and profitability improvement. Looking back over the past few years, the trend in profitability improvement has been unmistakable: losses narrowed significantly from 2022 to 2024, followed by the Company's first full-year profit in 2025.

In terms of customers, we have a total of over 1,000 contracted clients with deep penetration across more than 20 high-value industries such as energy, manufacturing, finance, retail and telecommunications, with a broad and solid customer base. As at the date of this announcement, the Company's orders on hand amounted to over RMB8.9 billion.

In respect of market position, according to the latest report from IDC, Phancy has ranked first in China's machine learning platform market share for seven consecutive years, representing the best testament from the industry to our technological strength and product competitiveness.

In terms of business structure, we have formally established the "three engines" that drive the Company's growth. This is not only a reclassification in our financial statements but also a deep reflection on our commercialization path in the AI 2.0 era. The AI Platform, as the core PaaS driver, contributed RMB6.552 billion in revenue in 2025, a year-on-year increase of 32.0%, accounting for 91.8% of the Group's total revenue and serving as the core pillar of the Company's revenue growth; the API business, empowered by Token, is the fastest-growing engine, with revenue reaching RMB79.9 million, an explosive year-on-year growth of 129.2%; the Agentic AI business, centered on a "result-as-a-service" model, achieved revenue of RMB503 million, a year-on-year increase of 93.2%, demonstrating strong growth momentum and the sustainable development of its business model. The three major segments, each with its own focus and mutually empowering, have jointly built a new multi-driver growth structure for Phancy.

I. AI Platform (PaaS, the Core Growth Driver)

Core Products: Centred on the Sage Platform, the platform encompasses a full-stack, end-to-end product matrix, including PhanthyCloud, HAMi vGPU, ModelHub XC, SageAIOS, Sage Ontology, SageSHIFT, and SageAIGS. This integrated matrix spans the entire AI process from model development and deployment to computing power support, serving as the cornerstone of the Company's business development. Notably, the platform has maintained the largest market share in China's machine learning platform sector for seven consecutive years.

Business Progress and Performance: Benefiting from the continued strong demand for domestic substitution and the implementation of the "AI+" initiative in China, this segment has emerged as a key pillar of the Company's revenue growth, accounting for most of the incremental revenue. In 2025, this business segment recorded revenue of RMB6.552 billion, representing a year-on-year increase of 32.0% and accounting for 91.8% of the Group's total revenue. This growth accelerated the Company's overall performance, enabling it to achieve single-quarter profitability for the first time in the third quarter of 2025. During the Reporting Period, the Company continued to drive the iterative enhancement of its products and technologies, with a particular focus on three core offerings: ModelHub XC, HAMi vGPU, and PhanthyCloud. Key developments are detailed as follows:

ModelHub XC: ModelHub XC was officially launched on September 22, 2025 as a core platform for the adaptation and management of information technology application innovation (ITAI, China's initiative to build an independent and controllable IT ecosystem) models. Its introduction also marked the launch of the ITAI community, ITAI models, and a suite of associated value-added services. ModelHub XC has now grown into the largest ITAI model community in China, effectively addressing key industry bottlenecks within the domestic computing power ecosystem, such as the need for separate adaptation of different models, time-intensive processes, and repetitive manual tasks. The product is powered by EngineX, an AI engine system specifically adapted for ITAI computing power. By adapting to the underlying algorithm architecture, it supports batch model processing and significantly enhances the efficiency of model adaptation. Leveraging the batch adaptation capabilities of EngineX, ModelHub XC has continuously accelerated model adaptation processes. The number of adapted and certified models has now surpassed 30,000, far exceeding initial projections. The product supports a broad spectrum of tasks such as text generation, visual understanding, and multi-modal Q&A, and encompasses mainstream model series such as MiniMax, Qwen, Hunyuan, GLM, Deepseek, Kimi, and Open Sora. The Company has initially planned to scale the number of adapted models to the hundred-thousand level within a year, a milestone it has already achieved ahead of schedule. Looking ahead, the Company

will continue to accelerate model adaptation, with the ultimate goal of expanding the ITAI computing community to a scale comparable with that of HuggingFace. In terms of hardware adaptation, ModelHub XC is fully compatible with mainstream domestic computing platforms, including Huawei Ascend, Cambricon, Iluvatar CoreX, Kunlunxin, MetaX, and Sunrise. In the future, the Company intends to further extend compatibility to encompass all mainstream ITAI computing platforms. In addition, the product offers a special value-added service for ITAI model adaptation. For users unfamiliar with the landscape of domestic computing power compatibility or requiring specific models to be adapted to designated domestic computing platforms, the service provides guaranteed adaptation. This ensures that clients using the service encounter no models that cannot be deployed, effectively addressing their practical application challenges. Meanwhile, ModelHub XC is closely integrated with PhanthyCloud, achieving seamless connection among model management, XC adaptation, and cloud services. Through PhanthyCloud, clients can readily access over 30,000 adapted models to rapidly deploy AI applications, further strengthening the integrated service capabilities across cloud, models, and computing power.

HAMi vGPU: As the Company's core product for GPU resource management, HAMi vGPU continuously upgrades its technological capabilities by leveraging HAMi-Core, a heterogeneous computing power virtualisation community initiated by Phancy and currently recognised as the most influential globally. In October 2025, HAMi vGPU achieved deep adaptation with Kubernetes Dynamic Resource Allocation (DRA) and launched the GPU DRA Driver, enabling integrated and fine-grained scheduling of GPU computing power and VRAM, and elevating cloud-native computing power management to a new level. The product effectively addresses the challenges of underutilisation and wasted GPU resources due to exclusive usage in traditional GPU deployments. It allows GPUs to be flexibly shared and scheduled like CPUs, supports parallel execution of multiple tasks on the same GPU node, and upgrades task allocation from a serial to a parallel scheduling approach. This significantly reduces task waiting time and enhances overall cluster throughput efficiency. In terms of resource application, HAMi vGPU supports the flexible selection of GPU configurations according to actual business needs, moving from fixed packages to customisable combinations. It enables fine-grained partitioning and pooling at the kernel level of down to $1\%$ of computing power and MB-level VRAM, thereby enhancing resource utilisation. In addition, launched in September 2025, the Virtual VRAM Expansion Card further extends the capabilities of HAMi vGPU by innovatively introducing extended VRAM technology. When used with HAMi vGPU, it can expand a single GPU's VRAM to 256GB (equivalent to ten RTX 4090 cards), overcoming traditional VRAM limitations without the need for hardware replacement. This innovation effectively removes hardware bottlenecks for large-scale AI tasks and meets high-performance computing demands of large model training and inference. The product is fully integrated with PhanthyCloud, creating seamless synergy with cloud services and ModelHub XC. Core capabilities, including GPU dynamic scheduling and VRAM expansion, are made accessible to customers via PhanthyCloud, enabling flexible, cloud-based invocation of computing resources. This approach not only reduces computing costs and operational workloads for customers but also enhances the integrated service capabilities of PhanthyCloud.

PhanthyCloud: As the core cloud service platform of the Company's full-stack AI PaaS matrix, PhanthyCloud works in close synergy with the Sage Platform and other key products to build an integrated service system across cloud, computing power, and AI capabilities, providing customers with convenient and efficient AI cloud service support. PhanthyCloud is now fully integrated with ModelHub XC and HAMi vGPU, enabling core capabilities such as ITAI model adaptation and GPU dynamic scheduling to be delivered as cloud services. This integration streamlines customers' deployment processes and reduces their hardware investment in AI deployment. Meanwhile, leveraging the Company's strong presence in the domestic computing power sector, PhanthyCloud is fully compatible with a wide range of mainstream domestic chips. It can provide secure, compliant, and fully controllable cloud services to core clients such as central and state-owned enterprises and financial institutions, precisely meeting their needs for transitioning to locally produced alternatives. In addition, PhanthyCloud is closely integrated with the Token business, enabling the cross-product utilisation of the Token-based payment mode. Clients can use Tokens to access cloud services, model resources, and computing power on demand, further enhancing flexibility and payment engagement, while supporting the large-scale growth of the Company's cloud service business.

Core Value: The Company aims to build an end-to-end, full-stack business system, foster deep synergy between the computing power business and the Sage Platform, provide customers with integrated services combining computing power and AI capabilities, and streamline the AI deployment process for customers, thereby lowering technical barriers. At the same time, through technological innovation and adaptation to domestic computing power, it supports the fully controllable development of domestic AI infrastructure, enabling enterprises to advance their AI-powered digital transformation.

Core Strengths and Business Structure:

The Company leads in hardware compatibility, supporting a full range of mainstream domestic chips, including Huawei Ascend, Cambricon, Iluvatar CoreX, Kunlunxin, MetaX, and Sunrise. The number of adapted and certified models on the platform currently exceeds 30,000, with plans to scale this to the hundred-thousand level within six months, and, ultimately, to cover all mainstream ITAI computing platforms on the market.
The Company is committed to building a fully controllable AI infrastructure adaptation system. HAMi-Core, a heterogeneous computing power virtualisation community initiated by Phancy and currently recognised as the most influential globally, has been fully integrated with DRA to upgrade cloud-native computing power management. This ensures business security and compliance, while aligning with domestic ITAI policy directives.
5 -

The Company precisely addresses the needs of core clients such as central and state-owned enterprises and financial institutions, seeking to transition from imported products to locally produced alternatives. These clients represent a high-quality base with stable demand. At the same time, the Company continues to expand its client portfolio, securing new business across key industries such as energy, manufacturing, finance, and retail.

II. API Business (Pay-as-you-go, the Fastest-growing Driver)

Core Products: With Phanthy as the core platform, complemented by PhanRouter and PhanClaw, the Company has established a comprehensive Token-based payment ecosystem. Each of the three products plays a distinct role, while operating in synergy. Leveraging the technological capabilities of the Sage Platform, they enable seamless integration between Tokens and core products, facilitating scenario-based monetisation. These products are detailed as follows:

Phanthy: Serving as the core platform of the Token-based payment ecosystem, Phanthy has evolved from a paradigm model development platform. While primarily focused on B-end Token-based payments, it also accommodates C-end operations, acting as the central hub linking the Company's AI infrastructure, model resources, and customer requirements. The platform is dedicated to scenario-based deployment and monetisation of enterprise-level large model applications, with services gradually being extended to C-end users. The platform is closely integrated with PhanthyCloud's cloud services, over 30,000 adapted models available through ModelHub XC, and a wide range of industry-specific vertical models accumulated over years of development on the paradigm platform, forming a comprehensive ecosystem of model resources. It also provides a full suite of API capabilities, supporting enterprise clients, consumer electronics devices, and developers to quickly integrate model and computing resources, thereby lowering the barrier to AI application development. At present, the platform reaches over 100 million consumer electronics devices. Supported by the on-device AI Agent module technology from Phancy, the Company's paradigm consumer electronics business brand, Phanthy's API capabilities have been widely applied across the terminal products of partner manufacturers, including Lenovo, Acer, and Konka. These applications span a range of consumer electronics categories, such as smartwatches and AI-enabled smart glasses, achieving the large-scale deployment of AI capabilities. The platform's core functions include model selection, computing power scheduling, Token top-up and consumption management, as well as usage data statistics. In addition, Phanthy continues to broaden its range of application scenarios and, in conjunction with the industry solutions offered by the Agentic AI business, extends the Token-based payment mode to AI use scenarios across various industries, thereby promoting the large-scale expansion of the Token-based payment ecosystem.

– 6 –

PhanRouter: As a unified API gateway for large models, PhanRouter is closely integrated with PhanthyCloud and serves as the primary channel for model invocation in the Token ecosystem. It is designed to address key industry challenges, including fragmented access to multiple models, complex operations and maintenance, and difficulties in cost control for enterprises. The product is fully compatible with the OpenAI interface standard and has achieved coverage of dozens of mainstream large model providers. It is also interoperable with products like the Phanthy world-level model and PhanClaw, creating a full-chain synergy from model invocation to computing power scheduling. In respect of adaptation to locally developed products, PhanRouter is closely integrated with PhanthyCloud's computing power scheduling capabilities and is fully compatible with ModelHub XC's domestic computing power adaptation system. It supports all mainstream domestic chips such as Huawei Ascend, Moore Threads, Iluvatar CoreX, Hygon, and Kunlunxin, and provides access to over 30,000 adapted models, ensuring alignment with domestic ITAI policy directives. In addition, the product supports private and independent deployment and can run independently within an enterprise's intranet environment to safeguard the security and compliance of core business data. It also provides refined cost control functions, including transparent usage tracking, intelligent scheduling, and quota alerts. Through the Token-based payment mode, clients can invoke services on demand, further reducing model invocation costs and operational complexity. The product is now officially available on PhanthyCloud, where clients can activate related services as required using Tokens.

PhanClaw: As the core product for security and permission control within the Token ecosystem, PhanClaw works in close coordination with Phanthy and PhanRouter. It undertakes critical functions, such as Token security verification, permission management, invocation auditing, and risk control, ensuring the stable and compliant operation of the Token-based payment ecosystem. The product enables lifecycle management of Tokens, encompassing generation, distribution, utilisation, and expiry or recycling. It provides precise verification of the legitimacy and permission scope of each Token invocation, preventing security risks such as misuse and unauthorised access. At the same time, PhanClaw supports customised permission allocation, enabling differentiated Token usage rights and quotas to be assigned across departments and roles in line with the client's organisational structure and business needs, thereby achieving refined permission management. It also provides a call log auditing function, recording Token invocation details in real time. This enables clients to trace usage records and conduct efficient troubleshooting, meeting the stringent security and compliance requirements of industries, such as finance and government affairs. Furthermore, PhanClaw is fully integrated with PhanRouter, enabling security verification for every model invocation routed through the gateway. In conjunction with Phanthy, it supports real-time synchronisation of Token quotas, ensuring compliant and controllable Token usage. This further enhances the full-chain Token service framework, combining platform and gateway control.

Monetisation Mode and Business Progress: PhanClaw adopts a pay-as-you-go, Token-based payment mode, focusing on key application scenarios such as AI model invocation, model adaptation, data interaction, and intelligent decision-making to drive scenario-based monetisation. This approach closely aligns with clients' on-demand usage requirements. Driven by the rapid rollout of OpenClaw, PhanClaw, serving as the security core of the Company's Token ecosystem, has further strengthened the security control framework for intelligent agent invocations. The large-scale adoption of AI agents is, in turn, fuelling exponential growth in Token consumption. Unlike the bursty, one-off Token consumption patterns associated with traditional single-turn AI query interactions, intelligent agents are capable of autonomously executing multi-step, closed-loop tasks. As a result, Token usage per task has increased markedly from the hundreds to the millions, significantly accelerating overall Token demand. On this basis, the Company remains confident that the intelligent agent-driven Token business will experience explosive growth in the future. It will continue to deepen the synergy between intelligent agents and the Token ecosystem, leveraging core products such as Phanthy and PhanRouter to further expand application scenarios of intelligent agents. This approach will support the wider adoption of the Token-based payment mode in more industries, drive large-scale expansion of the Token business, and reinforce its position as the fastest-growing segment, providing sustained momentum for the Company's overall performance. In 2025, this business segment recorded revenue of RMB80 million, representing a year-on-year increase of $129.2\%$ and accounting for $1.1\%$ of the Group's total revenue.

Core Value: By leveraging a flexible payment mode, the Token business reduces clients' upfront investment costs, enhancing their willingness to adopt and their payment engagement. At the same time, it enables cross-product reuse of Tokens, fostering synergy across multiple business segments. This approach brings sustained and stable revenue streams for the Company and supports its continued performance growth.

Core Strengths:

Combined with tailored Token packages, the flexible pay-as-you-go mode meets enterprise clients' needs for on-demand investment and effective cost control, thereby enhancing their willingness to adopt and increasing payment frequency.
Tokens can be used across the Company's core products, such as the Sage Platform and ModelHub XC, enabling multi-product synergy and enhancing Token reusability. This, in turn, promotes collaborative development across various business segments, and maximises customer value.
The flexible payment mode and cross-product versatility effectively enhance clients' payment engagement. Coupled with the ongoing incremental demand from the Company's key clients, this drives sustained consumption and provides a stable foundation for the continued growth of the Token business.

III. Agentic AI Business (Result-as-a-Service)

Business Positioning and Progress: Focusing on high-value industries and closely following the “AI+” national policy for high-quality energy development, the Company has prioritised promoting a result-as-a-service (RaaS) cooperation mode. This mode is being actively scaled across multiple industries and demonstrates a high level of replicability. Once an industry-specific use case, like electricity trading within the energy sector, has been successfully established, the corresponding AI solution, service pattern, and technical architecture can be quickly replicated across other similar scenarios and clients in the same industry, without the need for substantial additional R&D investment. Subsequent replications generate significant economies of scale, greatly enhancing the efficiency of business expansion, reducing operational costs, and further extending the Company’s business reach. At the same time, leveraging the Company’s full-stack AI capabilities, the RaaS mode can be quickly tailored to the personalised needs of different industries. This enables rapid replication from a single use case to multiple scenarios and from one industry to many, thereby driving the large-scale growth of the Agentic AI business. In 2025, this business segment recorded revenue of RMB503 million, representing a year-on-year increase of 93.2% and accounting for 7.1% of the Group’s total revenue.

Key Developments in Electricity Trading: Focusing on the two core scenarios of spot electricity trading and medium- to long-term electricity trading, the Company has developed a full-chain AI solution for electricity trading encompassing forecasting, decision-making, risk control, and post-trading review. This solution enables energy enterprises to improve trading efficiency, reduce operating costs, and mitigate market risks. It has been implemented in several pilot provinces for spot electricity trading and has gained recognition from key industry clients. Leveraging this solution, the Company has achieved significant business revenue growth across three sub-sectors: wind power, photovoltaic power, and energy storage. In the wind and photovoltaic power segments, accurate output forecasting and optimised trading strategies have enabled clients to increase electricity generation revenue and reduce deviation assessment costs. In the energy storage segment, AI-driven optimisation of charging and discharging strategies, integrated with electricity trading, has significantly enhanced the operational efficiency and profitability of energy storage stations. Overall, the Company has realised synergistic efficiency gains and sustained revenue growth across these three sub-sectors. At the same time, the Company has entered into an agreement with HyperStrong to jointly establish a joint venture. Relying on HyperStrong’s technological practice in R&D, production, and full-lifecycle management of energy storage systems, together with the Company’s leading capabilities in AI agents, machine learning platforms, and smart energy, the partnership will advance the application of AI-enabled electricity trading in energy storage projects, drive the comprehensive deployment of AI technologies in the energy storage field, and facilitate the intelligent upgrade of energy storage systems throughout their entire lifecycle.

Core Value: As the foundational contributor to the Company’s revenue base, this business segment operates stably. Through the widespread expansion of the RaaS mode across industries such as energy, manufacturing, and finance, it provides sustained and long-term support for the Company’s revenue growth, ensuring the overall stability of the Company’s performance. At the same time, through its deep industry presence and tailored services across multiple industries, the business segment supports the Company in consolidating its industry position, further tapping into the value of clients from various sectors, and fostering synergies with the AI Platform and API businesses. This, in turn, promotes the high-quality development of the Company’s overall operations.

Core Strengths and Business Performance:

Drawing on the Company’s extensive experience accumulated over more than a decade in the AI field across over 20 industries, this business segment is well positioned to accurately identify the challenges faced by different sectors in their digital transformation journeys. It continues to explore and expand the opportunities presented by the RaaS mode. At the same time, this mode demonstrates a high level of replicability. Once an industry-specific use case, like electricity trading within the energy sector, has been successfully established, the corresponding technology, solution, and service pattern can be quickly replicated across other similar scenarios and clients in the same industry, without the need for substantial additional R&D investment. Subsequent replications generate significant economies of scale, further reducing service costs, boosting profitability, and enabling rapid business expansion. Furthermore, following implementation, the business achieves a high degree of client integration and strong customer retention. Through the consistent delivery of long-term services and demonstrable value, the Company establishes trusted, enduring partnerships with its clients. This deep level of engagement strengthens client relationships, enhances retention, and drives higher rates of repeat business.

Strategic Outlook:

Looking ahead to 2026, the Company’s business will continue to be driven by the “three engines” of AI platform business, API business, and Agentic AI business. The AI platform business serves as the cornerstone and driver for the Company’s growth. Building on this foundation, the API and Agentic AI businesses are poised for even more robust development.

The API business is an explosively growing traffic hub — our youngest yet most resilient segment. With surging demand for computing power and models driven by phenomenal applications such as OpenClaw, the API business is experiencing exponential growth. An illustrative data point is that Token revenue for the first quarter of 2026 alone has already surpassed the full-year total for 2025.

The Agentic AI business acts as the Company’s “value multiplier” for empowering a wide range of industries. If the AI Platform is the foundation and the API is the reach, then Agentic AI is our “spearhead” for deeply addressing industry pain points. By building AI Agents with autonomous capabilities, we deliver business outcomes directly to clients, rather than merely providing tools. This year, this business will further penetrate clients’ core business areas, helping the Company evolve into an indispensable “digital productivity partner” for its enterprise clients.

A closed loop featuring “stable growth + resilient explosive growth + deep empowerment” will be further established and enhanced. With this new framework, the Company will continue to create long-term value for its Shareholders.

10 -

Corporate Social Responsibility

The Board places great importance on corporate sustainable development capabilities. It firmly believes that a robust ESG management system serves as the important foundation for effectively fulfilling environmental and social responsibilities. The Group has integrated sustainable development principles into its corporate strategy. The Group continuously reviews the progress of ESG initiatives and evaluates their effectiveness, while continuously strengthening sustainability risk management to address business demand adjustments arising from climate change, technological advancements, and changes in operating costs. The Group also continuously enhances its employees' awareness of ESG practices in strategy and operations through various means. To better manage the Group's ESG performance, identify potential risks, and support the Board in deepening ESG governance, the Group has established a top-down governance structure comprising the Board, an ESG Working Group, and business departments, thereby continuously advancing the implementation of ESG-related initiatives.

Phancy is keenly aware of the significant role of AI technology in promoting a low-carbon social transformation and sustainable development. Leveraging our leading AI technologies, products, and services, we are committed to assisting carbon-intensive enterprises in achieving their green transformation within sectors, such as energy, manufacturing, and transportation. Building on this foundation, the Group actively fulfills its social responsibilities by continuously investing financial, human, and material resources in areas such as public safety, culture and education, and environmental protection, thereby serving the needs of society and the public. At the same time, the Group remains focused on its core business. By leveraging technology to support community development, promote fairness and inclusiveness, it strives to advance the application of AI technologies in a wider range of public scenarios, giving back to society through concrete actions.

During the Reporting Period, facing the challenges brought by the summer flood season, Phancy collaborated with water conservancy agencies to develop a “Smart Flood Control Large Model”. Powered by AI to predict the evolution paths of flood disasters, the model enables precise dispatching of flood control resources. This project promotes the digital upgrading of public safety governance, while ensuring regional socio-economic stability and ecological security, highlighting the inclusive value of technology. Meanwhile, in response to the extreme heavy rainfall from severe convective weather and the resulting floods and landslide disasters in parts of Beijing, the Group took swift action. It donated RMB2 million to the disaster-stricken areas to support emergency rescue efforts, the provision of essential supplies for affected residents, and the restoration of post-disaster infrastructure—addressing urgent needs and helping local communities overcome hardships and rebuild their homes.

In addition, Phancy received the “ESG Sustainable Development Excellence Enterprise Award” at the 2025 Jinge Award for its innovative AI applications in ecological protection and disaster prevention. In the future, the Group will coordinate supply chain management and strengthen industry technology exchanges and collaboration. Through these efforts, it aims to advance responsible AI and jointly promote the high-quality development of the AI industry. The Group will continue to integrate ESG principles into the entire process of technology R&D and business implementation. It will explore pathways for AI for Good in more sectors, leverage AI to address social pain points, and truly achieve the vision of AI for everyone.

11 -

CONSOLIDATED STATEMENT OF COMPREHENSIVE INCOME

	Year ended December 31,
	Notes	2025 RMB'000	2024 RMB'000
Revenue	4	7,135,286	5,260,650
Cost of sales	6	(4,651,850)	(3,015,821)
Gross profit		2,483,436	2,244,829
Selling and marketing expenses	6	(205,183)	(268,699)
Administrative expenses	6	(154,374)	(193,310)
Research and development expenses	6	(2,337,401)	(2,169,767)
Provision for credit loss allowance, net		(17,398)	(199,961)
Other income	5	96,241	120,145
Other gains or losses, net		1,210	111,681
Operating loss		(133,469)	(355,082)
Share of results of investments accounted for using the equity method		17,410	19,265
Finance income		77,320	51,866
Finance costs		(1,031)	(6,137)
Loss before tax		(39,770)	(290,088)
Income tax credit/(expenses)	7	1,992	(6,179)
Loss for the year		(37,778)	(296,267)
Other comprehensive (loss)/income:
Item that may be reclassified to profit or loss
Currency translation differences		(13,231)	3,258
Item that will not be reclassified to profit or loss
Share of other comprehensive loss of investments accounted for using the equity method		-	(10,961)
Other comprehensive loss for the year, net of tax		(13,231)	(7,703)
Total comprehensive loss for the year		(51,009)	(303,970)

12 -

	Year ended December 31,
	2025	2024
	Notes	RMB’000
Loss attributable to:
Owners of the Company	(26,266)	(268,788)
Non-controlling interests	(11,512)	(27,479)
	(37,778)	(296,267)
Total comprehensive loss attributable to:
Owners of the Company	(39,497)	(276,491)
Non-controlling interests	(11,512)	(27,479)
	(51,009)	(303,970)
Loss per share for loss attributable to owners of the Company (expressed in RMB per share)	8
Basic	(0.05)	(0.58)
Diluted	(0.05)	(0.58)

CONSOLIDATED BALANCE SHEET

	Notes	As at December 31,
		2025 RMB'000	2024 RMB'000
Assets
Non-current assets
Right-of-use assets		14,554	24,299
Property and equipment		291,037	34,685
Intangible assets		476	189,747
Investments accounted for using the equity method		718,071	554,509
Financial assets at fair value through profit or loss	9	711,131	459,968
Contract assets		-	1,236
Term bank deposits		566,459	405,009
Prepayments and other receivables		22,303	227,927
		2,324,031	1,897,380
Current assets
Inventories		328,512	171,617
Contract assets		1,266	1,026
Trade receivables	10	2,340,882	3,085,640
Prepayments and other receivables		1,090,751	535,966
Financial assets at fair value through profit or loss	9	832,353	475,234
Term bank deposits		377,559	559,653
Restricted cash		465	2,511
Cash and cash equivalents		1,997,010	858,618
		6,968,798	5,690,265
Total assets		9,292,829	7,587,645

14 -

	As at December 31,
	2025	2024
	RMB'000	RMB'000
Equity and liabilities
Equity attributable to owners of the Company
Share capital	519,679	465,859
Treasury stock	(28,947)	(18,107)
Shares held for share award schemes	(255)	-
Reserves	12,416,509	9,969,530
Accumulated losses	(5,381,429)	(5,355,163)
	7,525,557	5,062,119
Non-controlling interests	(21,854)	8,769
Total equity	7,503,703	5,070,888
Liabilities
Non-current liabilities
Lease liabilities	1,023	11,470
Deferred tax liabilities	6,337	7,782
Borrowings	12	178,500
Other liabilities	17,743	-
	203,603	19,252
Current liabilities
Trade payables	11	1,172,241
Other payables and accruals	45,883	94,789
Contract liabilities	305,031	173,055
Lease liabilities	10,532	11,009
Income tax payables	-	1,336
Borrowings	12	23,383
Other liabilities	28,453	28,170
	1,585,523	2,497,505
Total liabilities	1,789,126	2,516,757
Total equity and liabilities	9,292,829	7,587,645

NOTES

1 GENERAL INFORMATION

Phancy Group Co., Ltd. (the “Company”) was incorporated in Shenzhen, the People’s Republic of China (the “PRC”) on September 17, 2014 as a limited liability company, and relocated to Beijing, the PRC on April 21, 2021. On July 9, 2021, the Company was converted into a joint stock company with limited liability under the Company Law of the PRC. The address of the Company’s registered office is Room 303, 3/F, Block A, Heying Centre, Building 1, No. 10 Xiaoying West Road, Haidian District, Beijing, the PRC.

The Company has changed its English name from “Beijing Fourth Paradigm Technology Co., Ltd” to “Phancy Group Co., Ltd.” with effect from January 7, 2026 and the name in Chinese from “北京第四範式智能技術股份有限公司” to “範式智能技術集團股份有限公司” with effect from February 4, 2026 in Hong Kong.

The Company is an investment holding company. The Company and its subsidiaries (collectively, the “Group”) are primarily engaged in operation of AI platform, and provision of API and Agentic AI Services in the PRC and certain overseas countries and regions.

Dr. Dai Wenyuan is the ultimate controlling shareholder of the Group as at the date of approval of these consolidated financial statements.

The consolidated financial statements of the Group are presented in Renminbi (“RMB”), which is also the functional currency of the Company.

2 APPLICATION OF NEW AND AMENDMENTS TO IFRS ACCOUNTING STANDARDS

Amendments to an IFRS Accounting Standard that are mandatorily effective for the current year

In the current year, the Group has applied the following amendments to an IFRS Accounting Standard as issued by the International Accounting Standards Board (“IASB”) for the first time, which are mandatorily effective for the Group’s annual period beginning on January 1, 2025 for the preparation of the consolidated financial statements:

Amendments to IAS 21	Lack of Exchangeability
The application of the amendments to an IFRS Accounting Standard in the current year has had no material impact on the Group’s financial positions and performance for the current and prior years and/or on the disclosures set out in these consolidated financial statements.

New and amendments to IFRS Accounting Standards in issue but not yet effective

The Group has not early applied the following new and amendments to IFRS Accounting Standards that have been issued but are not yet effective:

Amendments to IFRS 9 and IFRS 7	Amendments to the Classification and Measurement of Financial Instruments²
Amendments to IFRS 9 and IFRS 7	Contracts Referencing Nature-dependent Electricity²
Amendments to IFRS 10 and IAS 28	Sale or Contribution of Assets between an Investor and its Associate or Joint Venture¹
Amendments to IFRS Accounting Standards	Annual Improvements to IFRS Accounting Standards–Volume 11²
IFRS 18	Presentation and Disclosure in Financial Statements³
Amendments to IAS 21	Translation to a Hyperinflationary Presentation Currency³

16 -

Effective for annual periods beginning on or after a date to be determined.
Effective for annual periods beginning on or after January 1, 2026.
Effective for annual periods beginning on or after January 1, 2027.

Except for the new and amendments to IFRS Accounting Standards mentioned below, the directors of the Company anticipate that the application of all other new and amendments to IFRS Accounting Standards will have no material impact on the consolidated financial statements in the foreseeable future.

IFRS 18 Presentation and Disclosure in Financial Statements

IFRS 18 Presentation and Disclosure in Financial Statements, which sets out requirements on presentation and disclosures in financial statements, will replace IAS 1 Presentation of Financial Statements. This new IFRS Accounting Standard, while carrying forward many of the requirements in IAS 1, introduces new requirements to present specified categories and defined subtotals in the statement of profit or loss; provide disclosures on management-defined performance measures (MPMs) in the notes to the financial statements and improve aggregation and disaggregation information to be disclosed in the financial statements. In addition, some IAS 1 paragraphs have been moved to IAS 8 Accounting Policies, Changes in Accounting Estimates and Errors (the title of which will be changed to Basis of Preparation of Financial Statements upon effective of IFRS 18) and IFRS 17. Minor amendments to IAS 7 Statement of Cash Flows and IAS 33 Earnings per Share are also made.

IFRS 18, and amendments to other standards, will be effective for annual periods beginning on or after January 1, 2027, with early application permitted. IFRS 18 requires retrospective application with specific transition provisions. The application of new standard is not expected to have significant impact on the financial performance and positions of the Group in terms of recognition and measurement. However, it is expected to affect the structure and presentation of the consolidated financial statements.

3 BASIS OF PREPARATION OF CONSOLIDATED FINANCIAL STATEMENTS

The consolidated financial statements have been prepared in accordance with IFRS Accounting Standards as issued by the International Accounting Standards Board (“IASB”). For the purpose of preparation of the consolidated financial statements, information is considered material if such information is reasonably expected to influence decisions made by primary users. In addition, the consolidated financial statements include applicable disclosures required by the Rules Governing the Listing of Securities on The Stock Exchange of Hong Kong Limited (“Listing Rules”) and by the Hong Kong Companies Ordinance.

4 SEGMENT INFORMATION AND REVENUE

The Group’s business activities primarily involve the operation of AI platform, and provision of API and Agentic AI services in the PRC.

The Group’s chief operating decision maker has been identified as the Chief Executive Officer, who reviews consolidated results when making strategic decisions about allocating resources and assessing performance of the Group as a whole and hence, the Group has only one reportable segment. The Group does not distinguish between markets or segments for the purpose of internal reports. As substantially all of the Group’s non-current assets are located in the PRC and substantially all of the Group’s revenue are derived from the PRC based on the delivery location of the goods and services, no geographical information is presented.

17 -

The Group derives revenue from the transfer of goods and services at a point in time and over time are analysed as follows:

	Year ended December 31,
	2025	2024
	RMB'000	RMB'000
Point in time	7,121,797	5,237,020
Over time	13,489	23,630
	7,135,286	5,260,650

5 OTHER INCOME

	Year ended December 31,
	2025	2024
	RMB'000	RMB'000
Government grants (note)	43,040	24,485
Value-added tax and other tax refunds	52,869	95,517
Others	332	143
	96,241	120,145

Note: Government grants primarily relate to grants in connection with the Group’s contributions to technology development and investments in local business districts. There are no unfulfilled conditions or contingencies relating to these incomes.

6 EXPENSES BY NATURE

	Year ended December 31,
	2025	2024
	RMB'000	RMB'000
Employee benefit expenses	231,598	443,363
Share-based payment expenses	44,108	3,797
Cloud service and other technical service fees	723,816	233,264
Technology service fees	2,045,698	2,595,768
Cost of finished goods sold	4,046,529	2,064,221
Advertising and marketing expenses	155,233	148,179
Depreciation and amortisation
- property and equipment	12,075	22,400
- right-of-use assets	9,846	22,488
- intangible assets	4,317	16,890
Auditor’s remuneration
- audit services	2,200	5,100
- non-audit services	1,000	100
Other professional fees	16,082	20,185
Business travel expenses	6,591	18,767
Impairment provision for inventories	3,277	6,192
Others	46,438	46,883
	7,348,808	5,647,597

19 -

7 INCOME TAX CREDIT/(EXPENSES)

The income tax credit/(expenses) of the Group for the years ended December 31, 2025 and 2024 are analysed as follows:

	Year ended December 31,
	2025
RMB’000	2024
RMB’000
Current tax	–	(604)
Deferred tax	1,992	(5,575)
Income tax credit/(expenses)	1,992	(6,179)

8 LOSS PER SHARE

The basic loss per share is calculated by dividing the loss attributable to owners of the Company by the weighted average number of ordinary shares in issue during the respective years. In determining the weighted average number of ordinary shares in issue, treasury stock and shares held for share award schemes are excluded from the calculation.

Diluted loss per share is calculated by adjusting the weighted average number of ordinary shares outstanding to assume conversion of all dilutive potential ordinary shares. As the Group incurred losses for both years, the potential ordinary shares were not included in the calculation of diluted loss per share as their inclusion would be anti-dilutive. Hence, diluted loss per share for the years ended December 31, 2025 and 2024 are the same as basic loss per share for the respective years.

	Year ended December 31,
	2025	2024
Loss attributable to owners of the Company (RMB’000)	(26,266)	(268,788)
Weighted average number of ordinary shares in issue excluding treasury stock and shares held for share award schemes (thousand shares)	500,025	465,649
Basic and diluted loss per share for loss attributable to owners of the Company (expressed in RMB per share)	(0.05)	(0.58)

9 FINANCIAL ASSETS AT FAIR VALUE THROUGH PROFIT OR LOSS

	As at December 31,
	2025
RMB'000	2024
RMB'000
Non-current assets
Long-term investments measured at fair value through profit or loss
- Unlisted equity securities	333,684	113,695
- Preferred shares investments	-	3,634
- Fund investments	377,447	342,639
	711,131	459,968
Current assets
Short-term investments measured at fair value through profit or loss
- Listed equity securities	3,011	-
- Wealth management products	310,032	-
- Fund investments	519,310	475,234
	832,353	475,234

10 TRADE RECEIVABLES

	As at December 31,
	2025
RMB'000	2024
RMB'000
Third parties	2,633,458	3,366,568
Less: credit loss allowance	(292,576)	(280,928)
	2,340,882	3,085,640

The Group generally allows a credit period within 90 days to its customers. Aging analysis of trade receivables based on invoice date is as follows:

	As at December 31,
	2025
RMB'000	2024
RMB'000
Trade receivables
Up to 3 months	970,446	1,708,710
3 to 6 months	511,539	875,386
6 months to 1 year	237,756	346,588
Over 1 year	913,717	435,884
	2,633,458	3,366,568

20 -

21 -

11 TRADE PAYABLES

Trade payables primarily include payables for inventories and outsourcing service fees.

Trade payables and their aging analysis based on invoice date are as follows:

	As at December 31,
	2025	2024
	RMB'000	RMB'000
Up to 3 months	640,124	1,463,293
3 to 6 months	142,634	328,772
Over 6 months	389,483	391,198
	1,172,241	2,183,263

12 BORROWINGS

	As at December 31,
	2025	2024
	RMB'000	RMB'000
Included in current liabilities
Unsecured and guaranteed borrowing	-	4,000
Unsecured and unguaranteed borrowings	23,383	1,883
	23,383	5,883
Included in non-current liabilities
Unsecured and unguaranteed borrowings	178,500	-

13 DIVIDENDS

No dividends have been paid or declared by the Company during the years ended December 31, 2025, and 2024.

14 EVENT AFTER THE REPORTING PERIOD

On January 20, 2026, the Company acquired addition 39.58% of equity interest in Wake Up from other shareholders of Wake Up at a total consideration of approximately RMB278,363,000 and the consideration was paid as at the date of this report. Upon completion of the acquisition on January 20, 2026, Wake Up will be owned as to 84.58% by the Company and become a direct non-wholly owned subsidiary of the Company.

MANAGEMENT DISCUSSION AND ANALYSIS

FINANCIAL REVIEW

In 2025, the Company's total revenue amounted to RMB7,135.3 million, representing a year-on-year increase of $35.6\%$ . Gross profit was RMB2,483.4 million, with a gross profit margin of $34.8\%$ . With clear profitability goals, the loss for the year attributable to owners of the parent was RMB26.3 million, with a loss reduction of RMB242.5 million compared to the full year of 2024, narrowing by $90.2\%$ year-on-year. The adjusted net profit for the year was RMB6.3 million, compared with an adjusted net loss of RMB292.5 million for 2024, achieving a turnaround from loss to profit.

According to the latest report from the internationally authoritative research firm IDC, the Company has maintained its position as the leader in China's machine learning platform market share for seven consecutive years.

Revenue

As a steadily growing and developing AI software company, Phancy has further expanded its industry influence and business presence through our robust technological expertise and domain application experience. The Company has currently established three core business segments: AI Platform, API business and Agentic AI business. During the Reporting Period, our total revenue amounted to RMB7,135.3 million (same period in 2024: RMB5,260.7 million), representing a year-on-year increase of $35.6\%$ , mainly due to substantial revenue growth from AI platform. For the year ended December 31, 2025, revenue from AI platform amounted to RMB6,552.2 million, representing a year-on-year increase of $32.0\%$ and accounting for $91.8\%$ of our total revenue. Revenue from API business amounted to RMB79.9 million, representing a year-on-year increase of $129.2\%$ and accounting for $1.1\%$ of our total revenue. Revenue from Agentic AI business amounted to RMB503.2 million, representing a year-on-year increase of $93.2\%$ and accounting for $7.1\%$ of our total revenue. The following table sets forth a breakdown of our revenue by business segment for the periods indicated.

	Year ended December 31,
	2025 RMB in millions	2024 RMB in millions	Change
AI Platform	6,552.2	4,965.4	32.0%
API business	79.9	34.9	129.2%
Agentic AI business	503.2	260.4	93.2%
Total	7,135.3	5,260.7	35.6%

AI Platform

During the Reporting Period, revenue from AI Platform amounted to RMB6,552.2 million (same period in 2024: RMB4,965.4 million), representing a year-on-year increase of 32.0%, mainly due to the increasing demand in the overall AI market and the empowerment of our products through the Company’s large model and generative AI capabilities. The proportion of revenue from AI platform to total revenue decreased from 94.4% for the year ended December 31, 2024 to 91.8% for the year ended December 31, 2025.

API Business

During the Reporting Period, revenue from API business amounted to RMB79.9 million (same period in 2024: RMB34.9 million), representing a year-on-year increase of 129.2%. The proportion of revenue from API business to total revenue increased from 0.7% for the year ended December 31, 2024 to 1.1% for the year ended December 31, 2025, mainly due to our rising demand in the overall AI market.

Agentic AI Business

During the Reporting Period, revenue from Agentic AI business amounted to RMB503.2 million (same period in 2024: RMB260.4 million), representing a year-on-year increase of 93.2%. The proportion of revenue from Agentic AI business to total revenue increased from 5.0% for the year ended December 31, 2024 to 7.1% for the year ended December 31, 2025. This growth was primarily attributed to our strategic expansion pace for the high added value solutions and services.

Cost of Sales

Our cost of sales primarily includes: (1) cost of finished goods sold, primarily representing procurement cost of hardware components from third-party vendors, (2) technology service fees, primarily representing technology implementation costs paid to third-party service providers for delivery, deployment and installation of customized AI applications that we develop at users’ request, (3) employee benefit expenses, representing wages and benefits of our implementation and maintenance personnel for our enterprise AI solutions, and (4) others.

During the Reporting Period, our cost of sales amounted to RMB4,651.9 million (same period in 2024: RMB3,015.8 million), representing a year-on-year increase of 54.2%. Our cost of sales increased in line with the growth in our total revenue, mainly driven by our hardware procurement costs.

Gross Profit and Gross Profit Margin

Our gross profit increased by 10.6% from RMB2,244.8 million for the year ended December 31, 2024 to RMB2,483.4 million for the year ended December 31, 2025. Our gross profit margin decreased from 42.7% for the year ended December 31, 2024 to 34.8% for the year ended December 31, 2025, mainly due to the changes in the Company’s product portfolio compared to the previous year.

23 -

24 -

Selling and Marketing Expenses

During the Reporting Period, our selling and marketing expenses amounted to RMB205.2 million (same period in 2024: RMB268.7 million), representing a year-on-year decrease of 23.6%. This reduction was primarily due to the optimization of our brand promotion strategy, which allowed us to meet brand promotion needs while appropriately controlling marketing and brand promotion activity costs.

Administrative Expenses

During the Reporting Period, our administrative expenses amounted to RMB154.4 million (same period in 2024: RMB193.3 million), representing a year-on-year decrease of 20.2%, mainly due to the decrease of employee benefit expenses and professional services.

Research and Development Expenses

We continuously invest in the development and enhancement of our solutions and technologies as in the past. During the Reporting Period, the Group’s total research and development expenses further increased, reaching RMB2,337.4 million (same period in 2024: RMB2,169.8 million), representing a year-on-year increase of 7.7%, mainly due to increase of cloud service fees and decrease of technology service fees.

Provision for credit loss allowance, net

Our provision for credit loss allowance, net primarily includes impairment losses on trade receivables, contract assets, and other receivables.

During the Reporting Period, our provision for credit loss allowance, net amounted to RMB17.4 million (same period in 2024: RMB200.0 million), mainly due to the decrease of trade receivable balance and the enhancement of our collection management. In accordance with our accounting policy and relevant internal management system, we have made prudent provision for bad debts and will continue to closely monitor the recovery of receivables.

Other Income

Our other income primarily includes (i) government grants; and (ii) value-added tax refunds and other tax rebates.

During the Reporting Period, our other income amounted to RMB96.2 million (same period in 2024: RMB120.1 million).

Other gains or losses, net

Other gains or losses, net primarily include (i) fair value changes of financial assets measured at fair value through profit or loss; (ii) net foreign exchange gains/(losses); (iii) net gains on disposal/dilution of investments accounted for using the equity method; and (iv) net gains on partial disposal of subsidiaries.

During the Reporting Period, our other gains, net amounted to RMB1.2 million (same period in 2024: other gains, net of RMB111.7 million), mainly due to increase of foreign exchange losses and the decline of gains on partial disposal of subsidiaries.

Operating Loss

Due to the above factors, during the Reporting Period, our operating loss amounted to RMB133.5 million, representing a year-on-year decrease of 62.4% (same period in 2024: operating loss of RMB355.1 million).

Finance Income

During the Reporting Period, our finance income amounted to RMB77.3 million, representing a year-on-year increase of 49.1% (same period in 2024: RMB51.9 million), which was mainly due to the Company having conducted two placings during the Reporting Period, resulting in an increase in cash on hand.

Finance Costs

During the Reporting Period, our finance costs amounted to RMB1.0 million, representing a year-on-year decrease of 83.2% (same period in 2024: RMB6.1 million), mainly due to decrease of interest expense on borrowings and interest expense on payable for acquisition of subsidiaries recognized last year.

Loss for the Year

Due to the above factors, during the Reporting Period, our loss for the year amounted to approximately RMB37.8 million (same period in 2024: loss for the year of RMB296.3 million).

Non-IFRS Measures

To supplement our consolidated financial statements prepared in accordance with International Financial Reporting Standards (IFRS), we use adjusted net loss (a non-IFRS measure) as an additional financial metric. We believe that these non-IFRS measures facilitate comparisons of operating performance from period to period and company to company by eliminating potential impacts of certain items. We believe that these measures provide useful information to investors in understanding and evaluating our consolidated results of operations in the same manner as they help our management. However, presentation of adjusted net loss (a non-IFRS measure) may not be comparable to similarly titled measures presented by other companies. The use of these non-IFRS measures has limitations as an analytical tool, and investors should not consider them in isolation from, or as a substitute for analysis of, our results of operations or financial conditions as reported under IFRS.

25 -

26 -

Adjusted Net Profit/(Loss)

We define adjusted net profit/(loss) as the annual loss for the year adding back share-based compensation.

The following table reconciles our adjusted net profit/(loss) for the periods presented to the most directly comparable financial measures calculated and presented in accordance with IFRS, which is profit/(loss) for the year.

	Year ended December 31,		Change
	2025
RMB’000	2024
RMB’000
Reconciliation of loss for the year and adjusted net profit/(loss)
Loss for the year	(37,778)	(296,267)	–87.2%
Add: Share-based compensation	44,108	3,797	1,061.7%
Adjusted net profit/(Loss)	6,330	(292,470)	Turnaround to profit
Loss for the year attributable to the owners of the Company	(26,266)	(268,788)	–90.2%
Add: Share-based compensation expenses	44,108	3,797	1,061.7%
Adjusted net profit/(loss) attributable to the owners of the Company	17,842	(264,991)	Turnaround to profit

Liquidity and Financial Resources

In the past, we have primarily relied on shareholder contributions to meet our cash needs. Following the Global Offering, we plan to fund our future capital requirements through a balanced approach of equity financing and debt financing activities. We do not anticipate any changes in the financing required for our future operations.

As of December 31, 2025, all of the Group’s borrowings were denominated in RMB. The Group’s borrowings bear fixed interest rates between $2.4\% - 4.0\%$. All of the Group’s borrowings are due within three years.

Liquid Cash Resources

The Group’s liquid cash resources include cash and cash equivalents, short-term and long-term bank deposits, short-term investments measured at fair value through profit or loss, and restricted cash. As of December 31, 2025, the Group’s liquid cash resources amounted to approximately RMB3,773.8 million (December 31, 2024: RMB2,301.0 million).

27 -

Net Current Assets

Our net current assets increased from RMB3,192.8 million as of December 31, 2024 to RMB5,383.3 million as of December 31, 2025, mainly due to increase of current cash resources and decrease of trade payables.

Capital Management and Gearing Ratio

As of December 31, 2025, our gearing ratio (calculated as total borrowings divided by total equity attributable to equity holders of the Company) was approximately 2.7% (December 31, 2024: 0.1%). Our gearing ratio increased mainly due to the increase of borrowings. After taking into account the Group’s future business plans and the macroeconomic environment, we may seek debt or equity financing as necessary to continuously optimize our capital structure.

Capital Commitments

As of December 31, 2025, our capital commitments related to equity investments amounted to RMB4.0 million (December 31, 2024: RMB4.0 million).

Contingent Liabilities

As at December 31, 2025, we did not have any material contingent liabilities.

Pledge of Assets

As of December 31, 2025, except for restricted cash of RMB0.5 million as security deposits mainly for bidding, issuance of letter of guarantee or bank acceptance bills, the Group had no material pledge of assets.

Foreign Exchange Risk Management

The functional currency of the Company and its major subsidiaries is RMB. The majority of the Group’s revenue is derived from operations in the PRC. Foreign exchange risk primarily arises from recognized assets and liabilities denominated in currencies other than the functional currency of the Group’s entities. Fluctuations in exchange rates between RMB and other currencies during the Group’s operations may impact its financial position and operating results. The foreign exchange risk facing the Group mainly comes from the US dollars/RMB and Hong Kong dollars/RMB exchange rates. As of December 31, 2025, the Group did not hold any financial instruments for hedging purposes, nor did it hold any foreign currency investments hedged through currency borrowings or other hedging instruments.

28 -

Credit Risk

We are exposed to credit risk in relation to our cash and cash equivalents, restricted cash, term bank deposits, investments in debt instruments measured at fair value through profit or loss, trade receivables, other receivables and contract assets. The carrying amounts of each class of the above financial assets and contract assets represent our maximum exposure to credit risk in relation to financial assets and contract assets.

To manage risk arising from cash and cash equivalents, restricted cash, term bank deposits and investments in debt instruments measured at fair value through profit or loss, we only transact with state-owned banks and reputable or licensed financial institutions. There has been no recent history of default in relation to these financial institutions.

To manage risk arising from trade receivables and contract assets, we have policies in place to ensure that sales with credit terms are made to counterparties with an appropriate credit history and the management performs ongoing credit evaluations of its counterparties. We usually grant a credit term no more than 90 days and the credit quality of these customers is assessed by taking into account their financial position, past experience and other factors.

For other receivables, management makes periodic collective assessments as well as individual assessment on the recoverability of other receivables based on historical settlement records and past experiences. In view of the history of cooperation with debtors and the sound collection history of receivables due from them, the management believes that the credit risk inherent in the Group's outstanding other receivables balances due from them is low.

Fund and Working Capital Management

Our funds and liquidity management are centrally carried out by our finance department. Our finance department is generally responsible for overall management and implementation of funds, including formulating the capital management policy for the Group, guiding, coordinating and standardizing the fund management of regional companies, making annual funding plans, reviewing and summarizing annual capital budget, overseeing and assessing fund management of each regional company. We have also adopted sophisticated fund management policies and implemented a set of rules and guidelines on fund management to enhance the effectiveness and efficiency of fund management, thereby ensuring our financial security and reducing cost of capital.

To manage our idle cash on hand, we primarily purchase and redeem wealth management products using them as our "cash pool" from which we could readily access cash as needed and generate higher yield than bank deposits. The underlying financial assets of the wealth management products in which we invested primarily consist of the low-risk wealth management products issued by state-owned banks or other high-quality reputable banks in the PRC. The amount of the purchase will be determined based on our surplus funds. We consistently comply with our treasury policy during the procedures of purchasing the wealth management products and managing the relevant departments, as well as in conducting business, accounting and filing.

We are committed to safeguarding overall financial security and maintaining strong cash position and a healthy debt profile with strong repayment ability. By adopting a full, reasonable and professional assessment mechanism, preparing annual and monthly funding plans, we have established prudent fund management principle, which allows us to efficiently manage market risks.

MATERIAL ACQUISITIONS AND DISPOSALS OF SUBSIDIARIES, ASSOCIATES AND JOINT VENTURES

During the Reporting Period, we did not make any material acquisitions or disposals of subsidiaries, associates, or joint ventures.

FUTURE PLANS FOR SIGNIFICANT INVESTMENTS AND CAPITAL ASSETS

Except as disclosed in this announcement and the Prospectus, as at the date of this announcement, we did not have any plans for significant investments or capital assets.

EMPLOYEES

As at December 31, 2025, we had a total of 619 employees (as at December 31, 2024: 967 employees). Our success depends on our ability to attract, retain and motivate qualified personnel. During the Reporting Period, the total staff costs of the Group amounted to approximately RMB275.7 million, including wages, salaries, bonuses, share-based compensation expenses, pension costs, other social security costs, housing benefits and other employee benefits. We provide various incentives and benefits to our employees. We offer competitive salaries, bonuses and share-based compensation to our employees, especially key employees.

As required under PRC regulations, we participate in various employee social security plans that are organized by applicable local municipal and provincial governments, including housing, pension, medical, work-related injury, maternity, and unemployment benefit plans.

To maintain the quality, knowledge and skill levels of our workforce, we provide continuing education and training programs, including internal and external training, for our employees to improve their technical, professional or management skills. We also provide training programs to our employees from time to time to ensure their awareness and compliance with our policies and procedures in various aspects.

Prior to the Listing, the Board of Directors of the Company approved the adoption of the Employee Incentive Scheme on April 25, 2021. After the Listing, the Company adopted an equity incentive scheme (the "Equity Incentive Scheme") on September 19, 2024. The Equity Incentive Scheme consists of the H-Share Restricted Share Unit Scheme and the Partnership Employee Stock Ownership Plan of the Company. In addition, the Company adopted the Share Option Scheme on September 19, 2024. We have granted and may grant options or share-based incentive awards to our employees in the future to incentivize their contributions to our growth and development.

29 -

30 -

CHANGE OF THE ENGLISH NAME OF THE COMPANY, THE CHINESE NAME OF THE COMPANY, THE STOCK SHORT NAME OF THE COMPANY AND COMPANY LOGO

The English name of the Company has been changed from “Beijing Fourth Paradigm Technology Co., Ltd.” to “Phancy Group Co., Ltd.”, with effect from January 7, 2026. The Chinese name of the Company has been changed from “北京第四範式智能技術股份有限公司” to “範式智能技術集團股份有限公司”, with the Certificate of Registration of Alteration of Name of Registered Non-Hong Kong Company issued by the Registrar of Companies in Hong Kong on March 13, 2026.

The stock short name of H Shares for trading on the Stock Exchange changed from “FOURTH PARADIGM” to “PHANCY” in English and from “第四範式” to “範式智能” in Chinese, with effect from 9:00 a.m. on January 20, 2026. The Company has changed to a new company logo with effect from March 20, 2026.

Please refer to the Company’s announcements dated December 1, 2025, December 19, 2025, January 14, 2026, March 17, 2026 and March 20, 2026 and the circular dated December 3, 2025 and January 15, 2026 for details.

EVENTS AFTER THE END OF THE REPORTING PERIOD

Except for those as disclosed in note 14 to the consolidated financial information as included in this announcement, there were no significant events for the Group after the Reporting Period and up to the date of this announcement.

FINAL DIVIDEND

The Board does not recommend the distribution of a final dividend for the year ended December 31, 2025.

CORPORATE GOVERNANCE

The Board is committed to achieving good standards of corporate governance.

The Board believes that good corporate governance standards are essential for providing the Company with a framework to safeguard shareholders’ interests, enhance corporate value, formulate business strategies and policies, and improve transparency and accountability.

The Company has adopted the principles and code provisions under Part 2 of the Corporate Governance Code (the “CG Code”) contained in Appendix C1 to the Listing Rules as the basis of the Company’s corporate governance practices.

In the opinion of the Directors, during the Reporting Period, the Company has complied with all code provisions set out in the CG Code save as the code provision C.2.1 of the CG Code as set out below:

Pursuant to code provision C.2.1 of the CG Code, companies listed on the Stock Exchange are expected to comply with, but may choose to deviate from the requirement that the responsibilities between Chairman and Chief Executive Officer should be segregated and should not be performed by the same individual. The Company has not separated the roles of the Chairman and the Chief Executive Officer, and currently, Dr. Dai holds both positions. Dr. Dai has assumed the role of Chief Executive Officer of our Company since 2015. He has extensive experience in the business operations and management of our Group and in the AI industry. Given his experience, personal qualifications, and role in the Company, the Board believes that Dr. Dai is the most suitable Director to identify strategic opportunities and serve as the core of the Board due to his comprehensive understanding of our business as the Chief Executive Officer. The Board also believes that having the same individual serve as both Chairman and Chief Executive Officer can (i) ensure unified leadership within the Group, (ii) make the Board's overall strategic planning and execution of strategic measures more effective and efficient, and (iii) facilitate the flow of information between management and the Board. The Board believes that the current arrangement will not impair the balance of power and authority, and this arrangement will enable the Company to make and implement decisions swiftly and effectively. The Board will continue to review and, taking into account the overall situation of the Group, consider separating the roles of the Chairman and the Chief Executive Officer at an appropriate time.

The Company will continue to regularly review and monitor its corporate governance practices to ensure compliance with the CG Code and maintain high standards of corporate governance.

Save as disclosed herein, the Group has complied with the code provisions set out in the CG Code throughout the Reporting Period and up to the date of this announcement.

MODEL CODE FOR SECURITIES TRANSACTIONS

The Company has adopted the Model Code for Securities Transactions by Directors of Listed Issuers (the "Model Code") set out in Appendix C3 to the Listing Rules as the code of conduct for securities transactions by Directors. Upon specific inquiries made by the Company to the Directors, all Directors confirmed that they have complied with the required standards set out in the Model Code during the Reporting Period.

Employees who may possess unpublished inside information of the Company are also required to comply with the Model Code. During the Reporting Period and up to the date of this announcement, the Company has not been aware of any incidents of non-compliance with the Model Code by employees.

31 -

PURCHASE, SALE OR REDEMPTION OF LISTED SECURITIES OF THE COMPANY

During the Reporting Period, the Company repurchased a total of 308,900 Shares (the "Repurchased Shares") on the Stock Exchange for a total consideration (before expenses) of HK$11,711,150. Details of the Repurchased Shares are set out below:

Month	Number of Repurchased Shares	Price paid per Share		Total consideration (before expenses)
		Highest (HK$)	Lowest (HK$)
April 2025	308,900	40.20	35.95	11,711,150
Total	308,900			11,711,150

During the Reporting Period, the trustee of the Equity Incentive Scheme purchased a total of 235,700 H Shares on the Stock Exchange for an aggregate consideration (before expenses) of HK$10,872,950 in accordance with the rules of the Equity Incentive Scheme and the terms of the trust deed.

Save as disclosed above, neither the Company nor any of its subsidiaries has purchased, sold or redeemed any of the Company's listed securities (including transfer of treasury shares) during the Reporting Period.

As at the end of the Reporting Period, the Company held a total of 186,100 H Shares which have been repurchased but not yet cancelled and 502,200 H Shares which have been repurchased and retained as treasury shares. The treasury shares are intended to be used for employee incentives, sale or transfer to obtain liquidity and other purposes, as determined by the Board of Directors.

AUDIT COMMITTEE

The Board has established the Audit Committee which comprises two independent non-executive Directors and one non-executive Director, namely Mr. Li Jianbin (李建濱), Mr. Liu Chijin (劉持金) and Dr. Yang Qiang (楊強). Mr. Li Jianbin, being the chairman of the Audit Committee, is appropriately qualified as required under Rules 3.10(2) and 3.21 of the Listing Rules. The Audit Committee has also adopted written terms of reference which deal clearly with its duties and responsibilities (the terms of reference are available on the websites of the Company and Stock Exchange).

The Audit Committee has jointly reviewed with the management of the Company the accounting principles and policies adopted by the Company and discussed internal control and financial reporting matters (including the review of the audited financial statements for the year ended December 31, 2025) of the Group. The Audit Committee is of the opinion that the annual financial information for the year ended December 31, 2025 complies with applicable accounting standards, laws and regulations and that appropriate disclosures have been made.

33 -

SCOPE OF WORK OF THE AUDITOR

The figures in respect of the Group’s consolidated statement of comprehensive income, consolidated balance sheet, and the related notes thereto for the year ended December 31, 2025 as set out in this results announcement have been agreed to the amounts set out in the Group’s audited consolidated financial statements for the year ended December 31, 2025 by Rongcheng (Hong Kong) CPA Limited, the Company’s auditor. The work performed by the Company’s auditor in this respect did not constitute an assurance engagement and consequently no opinion or assurance conclusion has been expressed by the Company’s auditor on this results announcement.

PUBLICATION OF ANNUAL RESULTS ANNOUNCEMENT AND ANNUAL REPORT 2025

This annual results announcement is published on the websites of the Stock Exchange (www.hkexnews.hk) and the Company (www.4paradigm.com). The annual report of the Company for the Reporting Period, which sets out all the information required under the Listing Rules, will be published on the respective websites of the Stock Exchange and the Company in accordance with the requirements under the Listing Rules in due course.

ACKNOWLEDGEMENT

The Board would like to express its sincere gratitude to the Shareholders, the Group’s management team, employees, business partners and customers for their support and contribution to the Group.

DEFINITIONS

In this announcement, unless the context otherwise requires, the following expressions have the following meanings. These expressions and their definitions may not correspond to any industry standard definitions, and may not be directly comparable to similarly titled terms adopted by other companies operating in the same industries as the Company.

“associate(s)” has the meaning ascribed thereto under the Listing Rules

“Audit Committee” the audit committee of the Board

“Board” or “Board of Directors” Board of Directors

“China” or “PRC” the People’s Republic of China, for the purpose of this announcement and for geographical reference only, excluding Hong Kong and Macao Special Administrative Regions and Taiwan

34 -

“Company” or “Phancy”
Phancy Group Co., Ltd. (範式智能技術集團股份有限公司), the H Shares of which are listed on the Main Board of the Stock Exchange (stock code: 6682)

“Director(s)”
the director(s) of the Company

“Dr. Dai”
Dr. Dai Wenyuan (戴文淵), the chairman of the Board, an executive Director, chief executive officer and one of our Controlling Shareholders

“Global Offering”
the global offering of the H Shares in Hong Kong as described in the Prospectus

“Group” “our Group” “we” or “us”
our Company and our subsidiaries (or our Company and any one or more of our subsidiaries, as the context may require)

“H Share(s)”
overseas listed foreign share(s) in the share capital of our Company with a nominal value of RMB1.00 each, which is/are subscribed for and traded in HK dollars and listed on the Hong Kong Stock Exchange

“HK$”
Hong Kong dollars, the lawful currency of Hong Kong

“Hong Kong” or “HK”
the Hong Kong Special Administrative Region of the PRC

“Hong Kong Stock Exchange” or “Stock Exchange”
The Stock Exchange of Hong Kong Limited, a wholly-owned subsidiary of Hong Kong Exchanges and Clearing Limited

“IFRS”
the International Financial Reporting Standards, which include standards, amendments and interpretations promulgated by International Accounting Standards Board and the International Accounting Standards and interpretations issued by the International Accounting Standards Committee

“Listing”
listing of the H Shares on the Main Board of the Hong Kong Stock Exchange

“Listing Date”
September 28, 2023, on which our H Shares are listed and from which dealings therein are permitted to take place on the Hong Kong Stock Exchange

“Listing Rules” or “Hong Kong Listing Rules”
the Rules Governing the Listing of Securities on the Hong Kong Stock Exchange, as amended, supplemented or otherwise modified from time to time

35 -

"Main Board"
the stock exchange (excluding the option market) operated by the Hong Kong Stock Exchange which is independent from and operated in parallel with Growth Enterprise Market of the Hong Kong Stock Exchange

"Prospectus"
the prospectus dated September 18, 2023 issued by the Company in connection with Hong Kong public offering under the Global Offering

"Reporting Period"
for the year ended December 31, 2025

"RMB" or "Renminbi"
Renminbi, the lawful currency of the PRC

"Share(s)"
ordinary share(s) in the capital of our Company with a nominal value of RMB1.00 each

"Shareholder(s)"
holder(s) of Share(s)

"subsidiary(ies)"
has the meaning ascribed thereto under the Listing Rules

"year-on-year"
year-on-year change

"%"
Per cent.

For ease of reference, the names of Chinese laws and regulations, governmental authorities, institutions, natural persons or other entities (including our subsidiary) have been included in this announcement in both the Chinese and English languages and in the event of any inconsistency, the Chinese versions shall prevail.

Any difference between the arithmetical sum of individual figures and the aggregated figures in the tables in the announcement are due to rounding.

GLOSSARY OF TECHNICAL TERMS

“AI” artificial intelligence, simulation of human intelligence by machines

“AIGS” AI-Generated Software

“algorithm” a procedure of formula for solving a problem, based on conducting a sequence of specified actions

“cloud” the computers and connections that support cloud computing

“enterprise AI” AI technologies and software applied by enterprises to address their business needs and drive their digital and automation transformation

“IDC” International Data Corporation

“machine learning” the scientific study of algorithms and statistical models that computer systems use to effectively perform specific tasks without being explicitly programmed to do so

“R&D” research and development

By order of the Board

Phancy Group Co., Ltd.

範式智能技術集團股份有限公司

Dr. Dai Wenyuan

Chairman and Executive Director

Hong Kong, March 30, 2026

As at the date of this announcement, the executive Directors are Dr. Dai Wenyuan, Mr. Chen Yuqiang and Mr. Yu Zhonghao; the non-executive Directors are Dr. Yang Qiang, Mr. Dou Shuai and Mr. Zhang Jing; the independent non-executive Directors are Mr. Li Jianbin, Mr. Liu Chijin, Ms. Ke Yele and Mr. Pan Jialin; and the employee representative Director is Mr. Chai Yifei.

36 -

AI assistant

Phancy Group Co., Ltd. — Earnings Release 2025

51031_rns_2026-03-30_66c2ea04-8eac-4292-9c69-bd62fd8ff1d8.pdf

范式

PHANCY

Phancy Group Co., Ltd.

範式智能技術集團股份有限公司

ANNOUNCEMENT OF ANNUAL RESULTS FOR THE YEAR ENDED DECEMBER 31, 2025

BUSINESS REVIEW AND OUTLOOK

Overall Performance

II. API Business (Pay-as-you-go, the Fastest-growing Driver)

Core Strengths:

Core Strengths and Business Performance:

Strategic Outlook:

Corporate Social Responsibility

CONSOLIDATED BALANCE SHEET

1 GENERAL INFORMATION

2 APPLICATION OF NEW AND AMENDMENTS TO IFRS ACCOUNTING STANDARDS

Amendments to an IFRS Accounting Standard that are mandatorily effective for the current year

New and amendments to IFRS Accounting Standards in issue but not yet effective

IFRS 18 Presentation and Disclosure in Financial Statements

3 BASIS OF PREPARATION OF CONSOLIDATED FINANCIAL STATEMENTS

4 SEGMENT INFORMATION AND REVENUE

5 OTHER INCOME

6 EXPENSES BY NATURE

7 INCOME TAX CREDIT/(EXPENSES)

8 LOSS PER SHARE

11 TRADE PAYABLES

12 BORROWINGS

13 DIVIDENDS

14 EVENT AFTER THE REPORTING PERIOD

MANAGEMENT DISCUSSION AND ANALYSIS

FINANCIAL REVIEW

Revenue

Selling and Marketing Expenses

Administrative Expenses

Research and Development Expenses

Provision for credit loss allowance, net

Other Income

Other gains or losses, net

Operating Loss

Finance Income

Finance Costs

Loss for the Year

Non-IFRS Measures

Adjusted Net Profit/(Loss)

Liquidity and Financial Resources

Liquid Cash Resources

Net Current Assets

Capital Management and Gearing Ratio

Capital Commitments

Contingent Liabilities

Pledge of Assets

Foreign Exchange Risk Management

Credit Risk

Fund and Working Capital Management

MATERIAL ACQUISITIONS AND DISPOSALS OF SUBSIDIARIES, ASSOCIATES AND JOINT VENTURES

FUTURE PLANS FOR SIGNIFICANT INVESTMENTS AND CAPITAL ASSETS

EMPLOYEES

CHANGE OF THE ENGLISH NAME OF THE COMPANY, THE CHINESE NAME OF THE COMPANY, THE STOCK SHORT NAME OF THE COMPANY AND COMPANY LOGO

EVENTS AFTER THE END OF THE REPORTING PERIOD

FINAL DIVIDEND

CORPORATE GOVERNANCE

MODEL CODE FOR SECURITIES TRANSACTIONS

PURCHASE, SALE OR REDEMPTION OF LISTED SECURITIES OF THE COMPANY

AUDIT COMMITTEE

SCOPE OF WORK OF THE AUDITOR

PUBLICATION OF ANNUAL RESULTS ANNOUNCEMENT AND ANNUAL REPORT 2025

ACKNOWLEDGEMENT

DEFINITIONS

More from Phancy Group Co., Ltd.

Phancy Group Co., Ltd. Earnings Release 2025