Skip to main content

Biren Technology

Category: AI Infrastructure

A leading Chinese fabless semiconductor company specializing in high-end general-purpose GPUs (GPGPUs) designed for large-scale AI training, inference, and high-performance computing. Biren Technology was founded in 2019. The company is led by Zhang Wen. Based in Shanghai, China. Team size: 500-1000. Total funding raised: $1.621B. Latest round: IPO. Key investors include Shanghai State-owned Capital Investment (SSCI); Guangdong Province State Fund; Shanghai Government Fund.

AMW Analysis

Biren Technology is a Shanghai-based fabless semiconductor company, founded in 2019, that designs high-end general-purpose GPUs for large-scale AI training, inference, and high-performance computing. Its positioning centers on providing domestic alternatives to international benchmark GPUs, aimed at ensuring supply chain security for the Chinese market amid trade restrictions, and it has drawn backing from state-linked investors including Shanghai State-owned Capital Investment, Guangdong Province State Fund, and Shanghai Government Fund.

Recent news flow shows a rapid scaling of both capital and technical ambition. The company completed a Hong Kong IPO raising $717 million, with shares surging on debut and heavy retail oversubscription, alongside a reported order backlog alongside a first-half loss; proceeds were largely earmarked for R&D on its BR20X and BR30X chip lines. This was followed by an additional share sale raising roughly $892.5 million to expand GPU production capacity, again met with a sharp stock rally. Most recently, the company unveiled an optical "supernode" cluster architecture using near-packaged optics to link large numbers of processor cards, indicating continued investment in scaling infrastructure design beyond individual chip performance. Total funding raised stands at $1.621 billion.

AMW analysis, generated from 5 tracked news signals.

Founded
2019
Headquarters
Shanghai, China
Team size
500-1000
Total funding
$1.621B

Value proposition

Delivers high-throughput, energy-efficient domestic GPU alternatives that rival international benchmarks (such as NVIDIA's A100/H100) while ensuring supply chain security for the Chinese market amid global trade restrictions.

Products and solutions

BR100: Flagship GPGPU featuring advanced chiplet architecture and 77 billion transistors., BR104: Monolithic, single-die variant designed for standard PCIe form factors., Bili Series: A family of OAM-compatible cards and PCIe boards for diverse compute workloads., BIRENSUPA: A proprietary integrated software platform and programming stack for AI development., Haixuan: High-density OAM server modules for data center deployment.

Unique value

First Chinese GPU developer to break the 1 PFLOPS (FP16) performance barrier and a pioneer in utilizing 7nm chiplet technology for domestic high-end GPGPUs.

Target customer

Hyperscale data centers, cloud service providers (CSPs), AI model developers, and state-backed infrastructure projects requiring domestic high-performance computing solutions.

Industries served

Cloud Computing, Artificial Intelligence (LLM Training & Inference), Finance & Fintech, Healthcare & Bioinformatics, Autonomous Driving, Scientific Research

Technology advantage

Proprietary 'BiLiren' architecture optimized for high memory bandwidth (HBM2e) and massive parallel processing; seamless compatibility with mainstream AI frameworks like PyTorch and TensorFlow; strong strategic positioning as a 'national champion' for semiconductor self-reliance.

How they differentiate

Utilizes advanced 7nm chiplet architecture (BR100) to achieve high-performance GPGPU capabilities that rival international benchmarks while maintaining domestic supply chain security. Complemented by the proprietary BIRENSUPA software stack for seamless AI framework integration.

Main competitors

NVIDIA, Moore Threads, Huawei (HiSilicon), Enflame Technology

Key partnerships

Tencent (Infinigence AI): Collaboration to optimize LLM training capacity., Alibaba: Support for Tongyi QWQ-32B and other large-scale models., DeepSeek: Full-stack training and inference support for DeepSeek-V3., China Mobile & Shanghai AI Lab: Strategic infrastructure and ecosystem development., Baidu: Integration with the PaddlePaddle deep learning framework.

Notable customers

Tencent (Infinigence AI), Alibaba, China Mobile, China Telecom, SenseTime, DeepSeek

Major milestones

Launched the BR100 GPU in 2022, the first Chinese chip to break the 1 PFLOPS performance barrier., Navigated inclusion on the US Entity List in 2023 by securing $280M in state-backed bridge funding., Completed a $717M Hong Kong IPO, listing on the Hong Kong Stock Exchange on January 2, 2026 (Stock Code: 6082.HK), becoming HK's first new listing of 2026 and the largest 18C IPO to date.

Growth metrics

Unicorn valued at $2.2B (2024 Hurun); $207M fresh funding (Jun 2025); Hong Kong IPO targeting $624M; State-backed investors from Shanghai and Guangdong; Competing with Nvidia in AI chips

Market positioning

A 'national champion' and leading Chinese challenger in the high-end AI processor market, specifically targeting large-scale LLM training and data center acceleration.

Geographic focus

Mainland China (primary), Hong Kong, and Southeast Asia.

Patents and IP

Extensive IP portfolio with hundreds of patents filed in GPU architecture, chiplet interconnects, memory management, and high-speed data transmission.

About Zhang Wen

Zhang Wen (Michael Zhang) is a Harvard-educated lawyer and seasoned tech executive. Before founding Biren Technology in 2019, he served as the President of SenseTime, where he led the establishment of its Shanghai headquarters. His career includes serving as CEO of Enraytek Optoelectronics and working as a senior investor and lawyer on Wall Street (Simpson Thacher & Bartlett). He holds a Juris Doctor from Harvard University and degrees from Columbia University and the University of Science and Technology of China.

Latest news about Biren Technology

More AI Infrastructure companies

Official website: