Tired of ads? Enjoy an ad-free experience by signing up.
👩‍🍳 How we use AI at Tech in Asia, thoughtfully and responsibly.
🧔‍♂️ A friendly human may check it before it goes live. More news here

Chinese AI lab DeepSeek to run v4 on Huawei chips

DeepSeek plans to run its upcoming V4 model on the latest chips designed by Huawei Technologies, The Information reported, citing five people with direct knowledge of the matter.

The report said Chinese tech giants, including Alibaba Group, ByteDance, and Tencent Holdings, have placed bulk orders for Huawei’s upcoming chip in preparation for the model’s launch.

Those purchases total hundreds of thousands of units, The Information said, citing five people with direct knowledge of the orders.

🔗 Source: Reuters

🧠 Food for thought

Implications, context, and why it matters.

Sanctions push Chinese tech giants to bet on a less powerful, but available, alternative

  • The Information did not name Huawei’s “upcoming chip.” The sources also do not confirm it comes from Huawei’s Ascend 950 series. Ascend 950 is widely described as launching in 2026, with availability in Q4 2026 1.
  • The draft ties large orders by Alibaba and Tencent to U.S. export controls that limit access to NVIDIA’s most advanced chips in China. The cited source does not back the specific wording that controls “block” shipments of “top-tier Blackwell” into China. It says the rules cap NVIDIA’s top AI chips sold to China and require downgraded China variants 2.
  • The restrictions set up a locked-in market. Huawei is projected to take about 50% of China’s AI chip share by 2026 2.
  • Huawei can make up for weaker per-chip performance by scaling into bigger, power-hungry systems. One comparison pairs Huawei’s CloudMatrix 384 (a large AI server cluster) with Nvidia’s GB200 NVL72 (a high-end AI server system based on NVIDIA’s Blackwell platform) 3.

A parallel AI ecosystem is emerging beyond China’s borders

  • Pressure to use Ascend chips speeds work on software alternatives to NVIDIA’s CUDA, built around Huawei’s CANN framework (Huawei’s AI software platform for programming and running models on Ascend chips) 2.
  • Huawei is developing software to smooth over hardware differences and pull together idle compute. Flex:AI (software intended to pool and better schedule AI compute across hardware) targets a jump in AI chip utilization from about 35% to 70% 4.
  • The shift is pushing a “two-track” global AI infrastructure. NVIDIA leads outside China, while Huawei’s stack grows inside China 5.
  • Huawei is already shipping an integrated bundle into emerging markets. That includes chips, cloud services, and machine learning frameworks, with activity in parts of Africa, the Middle East, and some Latin American markets 6.

Stay ahead in Asia’s tech landscape

You've reached your 2 free content limit for the month. Sign up for free to read the full story.

🏄 For casual readers / 👶 Free

Basic

US$0

Free forever

Get instant access to this article and more every month

0 premium content

Unlimited news briefs

5

5 articles

Ad-free reading experience

Just US$0 per day

⌛Sign up in 20s. No payment details needed.

📖 For learners / 👍 Starter

Lite

US$4.92/month

Billed annually at US$59/year

Get instant access to this article and more every month

4

4 premium content

Unlimited news briefs & articles

Ad-free reading experience

Just US$0.17 per day

Cancel anytime

Our subscriber community includes professionals from these companies:

Stay updated on the go with our mobile app.

Get latest insights with smoother, more personalized experience through TIA mobile app.