Company profiles›Biren Technology

Biren Technology壁仞科技

6082.HKFounded 2019Shanghaibirentech.comReviewed 2026-08-30

Biren is one of China's most visible attempts to build GPU-class chips for AI training and inference.

Scale

~$143.9Mrevenue FY2025 · CNY 1.0B
~$11.9Bmarket cap · 2026-10-04

2 tracked chips with sourced pricing/specs

Gap vs frontier

Substitutesector position

Sector-level position; no priceable accelerator silicon to compare chip-by-chip.

Biren Technology: substitute tier in AI accelerators / GPUs, ~$143.9M FY2025 revenue.

The chip stackEach layer depends on the layers below it
1 · ChipsWhat the stack produces
2 · Design toolsSoftware and IP to design them
3 · ManufacturingPlants that turn designs into silicon
4 · EquipmentMachines inside the plants
5 · MaterialsWhat the plants consume

Company at a glance

Scale

Revenue up ≈17× since FY2023

Growth prospects

Expanding customer base and product roadmap

Private-company disclosure remains limited; revenue, customer, and deployment evidence are still relatively opaque.

China competition

Competes with Huawei Ascend, Moore Threads, Cambricon, Enflame in AI accelerators — peers listed at sector level.

Substitute

Global comparables

Global reference set: Nvidia, AMD, Google TPU.

Nvidia and AMD for datacenter GPU comparison.

Technical gap

Sector rated substitute vs the frontier.

Obstacles

Sector bottleneck applies

CUDA-like software ecosystem, HBM supply, advanced packaging, leading-node access, and large-scale deployment proof.

Official filings: www1.hkexnews.hk ↗ · www1.hkexnews.hk ↗

All sourced facts (6)
  • Founded: 2019
  • Headquarters: Shanghai.
  • Market cap / valuation: reportedly seeking about US$900 million in a funding round
  • Company type: Private datacenter GPU / GPGPU startup founded in 2019.
  • Known product family: BR-series processors and related accelerator products.
  • Disclosure: Private-company disclosure remains limited; revenue, customer, and deployment evidence are still relatively opaque.

Tracked chips: Biren BR100 · Biren BR104

About

Develops general-purpose GPU and AI accelerator silicon for datacenter workloads.

Shanghai · Founded 2019

Why it matters

Biren Technology is China's most prominent data-center GPU startup and an explicit Nvidia challenger, but its entire arc turns on whether it can sustain an advanced-AI-accelerator roadmap after US export controls severed access to leading-edge TSMC logic and CoWoS packaging. The test that confirms the thesis is volume deployment of a sanctions-compliant successor at named Chinese cloud and AI-lab customers; the test that breaks it is a stalled node roadmap and continued dependence on cut-off foreign foundry capacity.

Where it sits in the stack

Key semiconductor products

  • General-purpose GPU / GPGPU processors for AI training, inference, and datacenter compute.
  • Accelerator cards and systems built around BR-series silicon.
  • BR100 data-center GPU
  • BR104 GPU
  • BIRENSUPA software stack

Chip pricing

Financials

6082.HK · HKEX 6082 (Ch.18C, IPO Jan 2026) · via Yahoo Finance ↗

Financial trajectory via annual filings via Yahoo Finance

RMB billions
RMB 0.06BRMB -1.74B
2023
RMB 0.34BRMB -1.54B
2024
RMB 1.03BRMB -16.49B
2025

Market data refreshed daily; annual figures from exchange filings. Figures in the reporting currency shown on the chart.

Ecosystem

Competitors

  • NVIDIA
  • Huawei (Ascend)
  • AMD

Customers / end markets

  • Biren Technology
  • Target markets are datacenter AI training/inference, cloud infrastructure, and large enterprise or state-backed compute deployments
  • Customer signals is more opaque than for listed companies; confirmed deployment and procurement signals matter more than benchmark slides

Global peer set

  • Nvidia and AMD for datacenter GPU comparison
  • Domestic comparables include Huawei Ascend, Cambricon, Enflame, and Moore Threads; each has different software and customer-access strengths

Competitive position

  • Biren is a high-signal startup because its product ambition is close to the Nvidia/AMD datacenter-GPU lane, but the signals bar should be high.
  • Biren is best judged on shipped deployments, software maturity, memory/packaging access, and manufacturability, not launch claims.
  • Designs Nvidia-class data-center GPUs but lost leading-edge manufacturing access; capability gap is now process/packaging-bound rather than design-bound.
  • Leading-edge foundry, CoWoS packaging, and HBM supply are the binding constraints, all tightened by US export controls.
  • Global comparables: Nvidia and AMD for datacenter GPU comparison.
  • Global comparables: Domestic comparables include Huawei Ascend, Cambricon, Enflame, and Moore Threads; each has different software and customer-access strengths.
  • Advanced foundry access, CoWoS-like advanced packaging, high-bandwidth memory, board/system integration, and export-control exposure are central.
  • The biggest product risk is software ecosystem depth: compilers, framework support, kernels, model coverage, debugging, and customer migration costs.

Milestones and signals

  • 2019—Biren Technology was founded.
  • 2026-10-04—Market cap: HKD 93.0B
  • FY2025—Revenue: RMB 1.0B
  • FY2025—Net Profit: RMB -16.5B
  • reported (date not in snippet)—Funding / Valuation: USD about US$900 million (funding target, reported)

What to watch

  • Confirmed production deployments and repeat customers.
  • Product roadmap signals for BR-series accelerators and systems.
  • Manufacturing and packaging partners.
  • Software framework compatibility and model benchmark transparency.

Open questions

Last reviewed 2026-06-28
  • Which Biren products are actually deployed in production AI workloads?
  • Can Biren secure manufacturing, memory, and packaging capacity under export-control pressure?
  • Does its software stack lower migration friction enough for non-captive customers?