The Company has successfully transitioned from a technological investment phase into a period of sustainable and accelerated growth. During the reporting period, its endogenous growth momentum continued to strengthen, alongside improved capital efficiency and resilient cash flow generation. Trade receivables collection rose to a record RMB 4.87 billion. Operating cash flow recorded a positive net inflow in the second half of 2025, marking the first such achievement since listing.

Looking ahead, the Company plans to launch a new foundational model based on its second-generation NEO architecture in the second quarter of this year. The model is expected to deliver a further step in efficiency and cost-effectiveness, enabling the broad deployment of agentic AI applications.

Dr. Xu Li, Chairman of the Board and Chief Executive Officer of SenseTime, said: “We firmly believe that the deep integration of language and vision represents the most effective path to pushing the boundaries of artificial intelligence. Built on our innovative NEO architecture, we have achieved unified understanding and generation, while exploring a new ‘Scaling Law’ for multimodal intelligence. These technological breakthroughs, together with the deep integration of agentic AI, will unlock new application possibilities and enable a new generation of vertical use cases. While continuing to drive innovation in core technologies, SenseTime has delivered dual growth in both revenue and EBITDA, demonstrating strong growth resilience and operational efficiency across the industry, and is steadily advancing towards high-quality development.”

Continuously rolling out native multimodal large models, “efficiency enhancement + cost reduction” drives accelerated commercial growth

In 2025, SenseTime continued to invest in cutting-edge R&D, delivering breakthrough progress across large-model architecture innovation, training paradigms, inference efficiency, and spatial understanding, while maintaining its leading position in China.

The performance of the SenseNova multimodal large model continues to improve steadily, maintaining long-term leadership across multiple authoritative benchmarks. The Company has successively released and open-sourced the SenseNova-SI Spatial Intelligence Model. In a comprehensive evaluation incorporating several internationally recognized spatial intelligence benchmarks, SenseNova-SI delivered outstanding results, ranking first globally among peer models. In parallel, the open-source Kairos-SenseNova became the first embodied native world model to achieve integrated multimodal understanding, generation, and prediction. In December 2025, SenseTime also unveiled and open-sourced NEO, an fundamentally new native multimodal model architecture. NEO achieves state-of-the-art performance comparable to industry peers of a similar scale while requiring only one-tenth of the training data and computing power. Together, these breakthroughs redefined model training paradigms and inference efficiency, marking SenseTime’s entry into a new phase in its pursuit of deep multimodal integration

Leveraging the leading capabilities of its SenseNova multimodal large model, SenseTime has established a robust, scalable and replicable B2B business model. By applying a comprehensive framework that assesses task complexity and fault tolerance, the Company is able to unlock the closed-loop value of AI agents across strategic use cases, including office productivity, finance, marketing, and content generation. Its customer base spans a broad range of industries, encompassing automotive, smart devices, consumer goods, Internet services, embodied intelligence, financial services, education and healthcare.

Meanwhile, SenseTime has continued to make progress in developing a new generation of AI-native consumer applications. The Kapi product series has successfully built a user base numbering in the tens of millions, underscoring the strong growth potential of AI-native applications and demonstrating the Company’s ability to embed advanced AI technologies into everyday life.

SenseCore: Deeper coordination between computing infrastructure and model R&D unlocks international commercial potential

As the core technological pillar of its “Infrastructure – Model – Application” strategy, SenseCore achieved a significant transition in 2025, converting technological strengths to a fully developed industrial closed loop. During the year, it supported nearly one million model R&D tasks, effectively unlocking the full chain from underlying hardware to top-level applications, and from software stacks to model adaptation. Notably, the LightX2V World Model Inference System delivered breakthrough performance on domestic hardware, outperforming leading overseas chips.

SenseCore accelerated collaboration across China’s domestic technology ecosystem, positioning itself as a key chain master within industry. It partnered with more than a dozen chip makers, including Huawei Ascend, Hygon and Cambricon, to launch the SenseCore Computing Power Mall. SenseCore has since become a core partner of leading research institutes, Internet giants, pan-entertainment groups, embodied-intelligence robotics firms and large-model unicorns, and has also launched China’s first overseas Chinese-led computing cluster in Saudi Arabia.

As of the release of this result announcement, the total operational computing scale of SenseCore had reached 40,400 PetaFLOPS (FP16).

CV 2.0 records first‑ever net profit and positive cash flow; “X” Businesses gains strong recognition from external investors

SenseTime’s Computer Vision (CV; Visual AI) business is transitioning from a technology investment phase into a period of large-scale commercial returns. During the reporting period, CV 2.0 achieved profitability for the first time and generated positive cash flow for a second consecutive year, emerging as a key driver of the Company’s revenue growth and cash-flow improvement. SenseTime has maintained the leading position in China’s CV market for nine consecutive years. Internationally, clients across Southeast Asia, Northeast Asia and the Middle East continued to repurchase CV products and services, while interest increased from clients in South America, Europe and other regions, forming a replicable, scalable “Chinnovation” model for global expansion.

In 2025, SenseTime firmly advanced its “1+X” strategy, establishing a highly efficient collaborative system in which the core platform (the Company) provides foundational capabilities, while ecosystem partners compete and scale within targeted verticals. This marked a strategic shift from standalone business expansion towards the multiplication of ecosystem-wide value. Ecosystem enterprises incubated by SenseTime made steady progress in financing during the year, attracting strong interest from external investors including Internet giants, leading venture capital firms and industrial funds. Notably, the edge AI chip and intelligent driving business successfully completed financing rounds and were subsequently spun off to operate independently.

Looking ahead to 2026, the Company believes it is well positioned to compete in the critical phase of the global AI industry. It will continue to deepen its native multimodal architecture, reinforcing its global leadership in the integration of native multimodal large models and spatial intelligence. Meanwhile, the Company plans to capitalize on opportunities in the merging agentic AI market, targeting rapid growth in both user scale and commercial value. The Company will also accelerate the adaptation of domestic Chinese chips, further reduce large-model inference costs, and enhance product competitiveness through superior cost-performance. Within its Visual AI business, it will leverage both domestic and international growth engines to drive large-scale expansion and establish a global benchmark for intelligent industries. In parallel, the Company will continue to advance its “1+X” strategy to capture incremental returns arising from the accelerating intelligent transformation of vertical industries.

Hashtag: #SenseTimeGroup

The issuer is solely responsible for the content of this announcement.

About SenseTime

SenseTime is a leading AI software company focused on creating a better AI-empowered future through innovation. We are committed to advancing the state of the art in AI research, developing scalable and affordable AI software platforms that benefit businesses, people and society as a whole, while attracting and nurturing top talents to shape the future together.

With our roots in the academic world, we invest in our original and cutting-edge research that allows us to offer and continuously improve industry-leading AI capabilities in universal multimodal and multi-task models, covering key fields across perception intelligence, natural language processing, decision intelligence, AI-enabled content generation, as well as key capabilities in AI chips, sensors and computing infrastructure.

SenseTime’s business spans generative AI, vision AI, and innovative businesses. Through high-efficiency, low-cost, and scalable AI innovation and deployment, we complete the commercial value loop and lead the industrialization of artificial intelligence. Built with strategic foresight, our next-generation AI infrastructure, SenseCore, integrates computing power, algorithms, and platforms. Based on this foundation, SenseTime’s ‘SenseNova’ large model and R&D system can enable general AI tasks at a low cost. In addition, SenseTime continues to lead the vision AI market. Leveraging years of accumulated expertise in vision AI and empowered by cutting-edge multimodal large models, our SenseFoundry platform provides more stable and efficient vision AI support across diverse industries globally.

SenseTime advocates for a philosophy of “balanced development” in AI ethics and actively participates in the formulation of industry, national, and international standards related to data security, privacy protection, AI ethics, and sustainable AI. It has established close collaboration with multiple institutions in the Chinese Mainland and multilateral organizations to promote the sustainable and ethical development of AI. SenseTime’s Code of Ethics for Sustainable AI Development was included in the United Nations’ “Resource Guide on AI Strategies”, and its related research and practices have been repeatedly recognized by international partners.

SenseTime Group Inc. has successfully listed on the Main Board of the Stock Exchange of Hong Kong Limited (HKEX). We have offices in markets including Hong Kong, Shanghai, Beijing, Shenzhen, Chengdu, Hangzhou, Xi’an, Singapore, Bangkok, Kuala Lumpur, Riyadh, Abu Dhabi, Dubai, and Seoul, among others, as well as a presence in Germany, Thailand, Indonesia and the Philippines. For more information, please visit SenseTime’s official website or LinkedIn, X, and Facebook pages.