appConfigEntityId

bodyType

customContents

entityId

entityUuid

isVideo

articlePageQuery

article

articleSchemaContent

content

hooksArticlePageAdvertisementConfigArticle

hooksArticlePageFloatingButtonConfigArticle

appConfig

filter

json

liveArticle

Iris Deng

Ann Cao

China is ramping up efforts in the AI race, as tech giants Alibaba Group Holding and Huawei Technologies deploy massive computing clusters in the push to develop home-grown infrastructure.
E-commerce giant Alibaba has announced the deployment of a 10,000-card intelligent computing cluster powered by the Zhenwu AI chips developed by its T-Head semiconductor design arm.
Launched in collaboration with China Telecom in the Shaoguan data centre in Guangdong province, the “fully domestic” cluster was the first Zhenwu-powered project of such scale in the Greater Bay Area, Alibaba’s cloud unit said in a statement issued on Tuesday.
The Zhenwu-powered cluster is the latest evidence that China is doubling down on home-grown infrastructure to supercharge rapidly evolving AI development and meet surging demand as the artificial intelligence race with US rivals including Meta, Microsoft and Elon Musk’s xAI intensifies.
The announcement of the Alibaba-backed cluster followed the activation of the country’s first 10,000-card intelligent computing cluster, built with Huawei’s Ascend 910C AI chips, in Shenzhen, Guangdong, late last month.
The new cluster showed China’s advanced computing power was “moving from high-end performance breakthroughs to large-scale industrial implementation”, Alibaba Cloud said in its statement on Tuesday.
The [government] sector’s rigid demand for data sovereignty and security has driven the fastest deployment
Charlie Zheng, Samoyed Cloud Technology
Charlie Zheng, chief economist at Samoyed Cloud Technology Group Holdings, said the recent launches of domestic computing clusters came as China’s AI industry was shifting from “hardware replacement” to “software collaboration”.
He said such domestic technology-based computing clusters were now seeing rapid deployment in the government service and city governance sector, which had the highest requirements for home-grown computing power.
“The sector’s rigid demand for data sovereignty and security has driven the fastest deployment,” Zheng said.
Alibaba said the Zhenwu cluster promised ultra-low latency of 4 microseconds thanks to its next-generation high-performance networking architecture, which allowed the 10,000 chips to work as a single supercomputer and train models with hundreds of billions of parameters.
It said the cluster could deliver 30 per cent higher training and inference efficiency, with single-card throughput increasing by nearly 10 times.
The cluster has already been deployed in healthcare and advanced manufacturing industries, the company said. Small and medium-sized enterprises could now access its computing power through China Telecom’s platform and pay by the card or hour, it said.
Alibaba also revealed plans to expand the scale of the cluster to 100,000 cards, to further lower costs and improve the efficiency of computing resources. Alibaba owns the South China Morning Post.
While the individual chips developed by Chinese tech giants still trail behind those of global leaders such as Nvidia, Beijing is looking to catch up with large-scale computing clusters, banking on innovative and efficient network architecture to improve computing performance.
Beijing included intelligent computing infrastructure in the country’s 15th five-year plan last month, pledging to boost the supply of high-performance computing resources and push the construction of ultra-large-scale intelligent computing clusters.
A State Council AI action plan released in August emphasised the buildout and optimised distribution of computing resources across China.
Local governments have quickly devised plans to boost intelligent computing facilities. In its computing road map for the next three years, the southern Chinese tech hub of Shenzhen made “autonomous controllability” a top priority and vowed to build a full-stack ecosystem by engaging the supply chain, hyperscalers and AI model operators.
By the end of June last year, the country’s total computing power reached 962,000 petaflops, commanding 21 per cent of the world’s total capacity on the back of a 73 per cent year-on-year increase, according to the China Academy of Information and Communications Technology.
The Ascend-powered cluster in Shenzhen has a computing capacity of 11,000 petaflops, and has been combined with a 3,000-petaflop cluster activated last year.
In eastern China, Shanghai’s AI infrastructure efforts are being anchored by Shanghai Intelligent Computing Technology, a subsidiary of state-owned INESA, with a 10,000-card cluster compatible with a range of domestic chips, the Shanghai media outlet ThePaper.cn reported on Monday.


B-2 Spirit, AI cooperation, no US-China conflict: key Xi-Trump takeaways so far

Singapore’s finance firms pledge to train 80,000 local staff in AI skills

Don’t let AI take over our human creative expression

Alibaba unveils ‘pragmatic’ AI road map to drive monetisation, infrastructure efficiency

Artificial intelligence (AI)

As AI race with US intensifies, China’s Alibaba launches 10,000-card computing cluster

In collaboration with China Telecom, Alibaba has deployed a 10,000-card intelligent computing cluster powered by its Zhenwu AI chips in the Shaoguan data centre in Guangdong province.  Photo: Sina

China turns to offshore wind farms, subsea data centres to ease AI computing bottleneck

An underwater data centre being developed by Highlander is seen under construction at a shipyard in Nantong,  Jiangsu province on September 11, 2025. Photo: AFP

A visitor observes a miniature model of undersea data cabins for an undersea data centre, in Lingshui, south China’s Hainan province, Sept. 25, 2025. Photo: Xinhua

Chinese AI rivals clash over Anthropic’s OpenClaw exit amid global token crunch

A lobster-shaped cutout representing OpenClaw stands outside the Baidu offices in Beijing, March 17, 2026. Photo: Reuters

Several AI applications can be seen on a smartphone screen, including ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot, Meta AI, Grok and DeepSeek. Photo: dpa

A setup session to install OpenClaw on laptops, outside the Baidu offices in Beijing, March 17, 2026. Photo: Reuters

Semiconductor, AI talent race heats up as Taiwan tightens crackdown on alleged poaching

Technicians work on chip processing equipment at a semiconductor manufacturing plant in Suqian, 
Jiangsu province, in October. Photo: AFP

SMIC’s facilities in the Pudong district of Shanghai. Photo: AFP

A chip processing plant in Suqian, in eastern China’s Jiangsu province. Photo: AFP

Chinese firms trail US peers in AI adoption due to corporate culture: ex-OpenAI executive

China has a very “techno-centric consumer”, while the US has a very “techno-centric enterprise”, according to a former OpenAI executive. Photo: Shutterstock

 Chinese consumers are more tech-savvy than their American counterparts, says Zack Kass. Photo: Getty Images

How can Beijing attract top-tier Chinese AI professionals based abroad?

China’s strength in training large numbers of AI talent has not been matched by equally robust development environments. Photo: Xinhua

Tech

Economy

China Economy

News

China Future Tech

Semiconductors

Cluster in Shaoguan powered by domestically developed Zhenwu chips latest evidence that China is doubling down on home-grown infrastructure


Cluster in Shaoguan powered by domestically developed Zhenwu chips latest evidence that China is doubling down on home-grown infrastructure.


As AI race with US intensifies, China’s Alibaba launches 10,000-card computing cluster

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Cluster in Shaoguan powered by domestically developed Zhenwu chips latest evidence that China is doubling down on home-grown infrastructure

Cluster in Shaoguan powered by domestically developed Zhenwu chips latest evidence that China is doubling down on home-grown infrastructure