Biren Technology is a fabless Chinese semiconductor company focused on GPGPU chips, AI accelerator cards, computing software and integrated intelligent-computing solutions. Its business model is product-led: the company sells bundled hardware and software platforms for AI and high-performance computing, with revenue typically recognized after customer adaptation is completed and the customer accepts the solution.
The company has one reportable operating segment because management reviews the business on a consolidated basis. In practical terms, Biren operates as an AI compute platform provider rather than a diversified semiconductor conglomerate.
- Intelligent computing solutions: This is the core revenue stream. In 2025, it generated RMB1,027.7 million of Biren’s RMB1,034.6 million total revenue, making the business overwhelmingly dependent on sales of GPGPU-based solutions and ready-to-use applications.
- Support and extended warranty: This remains small, with RMB0.9 million of revenue in 2025. It supports the installed base but is not yet a meaningful recurring revenue stream.
- Entrusted R and D services: Biren generated RMB4.6 million from entrusted R and D service in 2025, reflecting selective customer or partner development work.
- Intelligent computing cluster rental: Rental income was RMB1.4 million in 2025, showing early use of compute-cluster monetization outside direct product sales.
Biren’s market position is tied to China’s push for domestic AI infrastructure and semiconductor self-sufficiency. Substantially all revenue comes from the PRC, and substantially all non-current assets are located in the PRC. The company presents itself as a homegrown compute supplier with chips designed in China, manufactured in China and deployed in China.
Commercialization is still early, but scaling quickly. Revenue rose from RMB0.5 million in 2022 to RMB62.0 million in 2023, RMB336.8 million in 2024 and RMB1,034.6 million in 2025. Gross profit reached RMB557.0 million in 2025, with a gross margin of 53.8%. The company remains loss-making, with an operating loss of RMB1,041.9 million and an adjusted loss of RMB873.8 million in 2025, mainly as heavy R and D investment continued.
Biren’s competitive advantages are its domestic positioning, integrated hardware-software approach and access to public-market funding after its January 2026 Hong Kong listing. Net proceeds from the global offering were approximately HK$6.20 billion, equal to RMB5.63 billion, and 85% was allocated to intelligent-computing R and D. This gives the company a larger funding base to develop next-generation AI compute products through 2029.
The main competitive pressure comes from Nvidia globally and from Chinese AI chip peers such as Moore Threads, Iluvatar CoreX, MetaX and Enflame. Nvidia remains the benchmark for AI accelerator performance, software ecosystem depth and global data-center adoption. Biren’s opportunity is different: it competes as a China-focused alternative for customers that need domestic AI accelerators under export-control and supply-chain constraints.
Customer concentration is a key feature of the current model. Five customers each contributed more than 10% of 2025 revenue, with the largest disclosed customer contributing RMB197.6 million. This supports near-term scale, but it also makes revenue more dependent on a small group of large projects and on acceptance timing for bundled solutions.
Biren is best viewed as an early-stage public-market AI accelerator platform company with strong China exposure, rapid revenue growth and an unproven profitability profile. Its position is strategically important in China’s domestic GPU market, but its ability to close the gap with global leaders and scale production profitably remains central to the investment case.