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

Alibaba expands AI push with 10,000-chip cluster

Alibaba said it has deployed a 10,000-chip AI computing cluster in southern China, as the country accelerates efforts to build domestic AI infrastructure.

The system uses chips developed by Alibaba’s T-Head semiconductor unit and is located at a China Telecom data centre in Guangdong.

The company said the project marks its first large-scale deployment of the Zhenwu chips in the region.

The move comes as Chinese tech firms ramp up investment in homegrown AI systems, following similar large-scale deployments by Huawei.

🔗 Source: South China Morning Post

🧠 Food for thought

Implications, context, and why it matters.

Alibaba’s chip ties together its end-to-end AI ecosystem

  • Alibaba calls the Zhenwu 810E chip part of its “golden triangle” plan, which links its in-house AI models, cloud services, and custom-designed chips into one stack 1.
  • The chip aims to serve as a China-made option for AI work. A report citing The Information said some Alibaba employees say Zhenwu competes with Nvidia’s export-restricted H20 chip 2.
  • Before the 10,000-card cluster announcement, Alibaba Cloud had already put Zhenwu chips into several large clusters. Those systems supported more than 400 customers, including State Grid of China, the state-owned power grid operator, and Xpeng Motors, the Chinese electric-vehicle maker 1.

China’s domestic AI buildout keeps growing despite slower chips

  • Alibaba and Huawei released back-to-back updates that add to China’s homegrown AI infrastructure. Huawei’s 10,000-card cluster in Shenzhen reached 92% bookings across two phases after launch 3.
  • That drive for self-reliance brings a performance tradeoff. An analysis by the Council on Foreign Relations, a US think tank focused on foreign policy, says Huawei’s AI chips still trail global leaders 4.
  • Huawei’s Ascend 910C delivers about 60% of Nvidia H100 inference performance under comparable conditions, according to the same analysis 4.
  • The same analysis expects the gap to grow due to US export controls and supply-chain limits. It projects top US AI chips could exceed Huawei’s best offerings by more than 17 times by 2027 4.

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.