2 Sources
[1]
Z.AI built a giant AI data centre on Chinese-made chips
A Chinese AI lab has finished a giant data centre that runs on Chinese-made chips alone, with no Nvidia inside. It is one of the clearest signs yet that Beijing's push to build around US export controls is working. For a year, one question has hung over China's fast-improving AI models: what are they actually running on? Z.AI has just given part of the answer. It has built a huge data centre that uses only Chinese-made chips. The company, formerly known as Zhipu, has started partially operating the site, Bloomberg reported, citing a person familiar with the matter. It is a 1-gigawatt hub built to train Z.AI's GLM models. Z.AI did not respond to a request for comment. What was built The scale is the headline. A gigawatt is roughly the power draw of 750,000 homes at any one moment. That puts the site among the largest server hubs any Chinese AI lab has built. The chip count is just as striking. Z.AI now runs several computing clusters, each holding more than 10,000 chips, the person said. Crucially, none of them are Nvidia's. Bloomberg did not name the exact chips. China's leading designer of AI accelerators is Huawei, which competes with local firms such as Cambricon and Alibaba to close the gap on Nvidia. Why it matters US export controls have blocked China's labs from buying Nvidia's most powerful chips. The open question was whether homegrown parts could carry the load for training frontier models, not just running them. A 1GW cluster built entirely on domestic silicon is a real answer to that question. It suggests Chinese labs can keep scaling even while cut off from the chips the rest of the industry treats as default. The timing sharpens the point. The buildout lands days after Beijing-based Moonshot's Kimi K3 matched top US models, then ran short of compute and paused new sign-ups. Z.AI is racing the same rivals, and it is betting on owning the hardware. The bigger buildout Z.AI is not doing this alone. China is preparing to spend around 2 trillion yuan, about $295bn, over five years on data centres across the country. Cloud giants Alibaba and China Telecom remain the biggest builders so far. The company also has the money to compete. Fresh from a Hong Kong listing and a follow-on share sale, Z.AI is on track for $1bn in annual recurring revenue, which would make it the first Chinese AI firm to reach that mark. Its shares jumped almost 20% on the day. It is positioning itself as an enterprise AI supplier, a role that invites comparison with Anthropic. The catch A few caveats are worth keeping. The chip details come from an anonymous source, and Z.AI has not confirmed them. Building the cluster is not the same as proving it can train a frontier model as efficiently as an Nvidia one. Domestic chips can still lag on raw performance, and stitching thousands of them into a stable system is hard. The real test is the next GLM model, and how it compares with what the US labs ship. Still, the direction is clear. The effort to build a Chinese alternative to Nvidia has moved from slideware to a working, gigawatt-scale data centre. That alone changes the debate about how far export controls can hold.
[2]
China AI chip stocks rally after Z.AI's domestic-chip data center milestone By Investing.com
Investing.com -- Chinese artificial intelligence and semiconductor stocks rallied on Tuesday after Bloomberg reported that AI startup Z.AI had completed a large data center powered entirely by domestically produced chips, reinforcing investor optimism that China's AI ambitions can advance despite U.S. curbs on Nvidia's most advanced processors. Track China's AI and semiconductor leaders with InvestingPro - now 60% off Chipmakers led gains across mainland markets, with NAURA Technology Group Co Ltd (SZ:002371) jumping the daily 10% limit, Cambricon Technologies Corp Ltd (SS:688256) climbing 8.5%, Foxconn Industrial Internet Co Ltd (SS:601138) gaining 4.8%, BOE Technology Group Co Ltd (SZ:000725) rising 6.1% and Luxshare Precision Industry Co Ltd (SZ:002475) adding 6.1%. AI software developer Iflytek Co Ltd (SZ:002230) rose 0.5%, while smartphone maker Xiaomi gained 0.4%. In Hong Kong, Semiconductor Manufacturing International Corp (HK:0981) climbed 8.8%, Alibaba Group Holding Ltd (HK:9988) rose 1.8% and JD.com (HK:9618) gained 1.3%. The AI rally also helped fuel gains in China's small-cap technology universe, lifting the CSI 1000 Index 3.5%, while the Shanghai Composite rose nearly 1% and Hong Kong's Hang Seng Index was little changed. The rally reflects growing confidence that China's AI ecosystem is becoming less dependent on imported computing hardware. Investors have increasingly focused on whether domestic chipmakers can provide the computing power needed to train advanced AI models as U.S. export restrictions tighten access to Nvidia's leading AI processors. Bloomberg reported that Z.AI, formerly known as Zhipu, has completed construction of a 1-gigawatt data center built entirely with Chinese-made chips and has begun partially operating the facility to train its GLM family of AI models. The company has also built or operates several computing clusters containing more than 10,000 chips each, according to a person familiar with the matter. Evidence that a major Chinese AI developer can deploy a large-scale computing cluster using only locally designed chips is being viewed as a positive signal for the country's semiconductor supply chain, benefiting equipment makers, chip designers and manufacturers alike. The report comes as competition among China's leading AI developers intensifies following Moonshot AI's launch of its Kimi K3 model last week. Bloomberg also reported that China is preparing to invest around 2 trillion yuan ($295 billion) over the next five years to expand data center infrastructure, highlighting Beijing's push to build a self-sufficient AI ecosystem.
Share
Copy Link
Chinese AI lab Z.AI has completed a 1-gigawatt data centre running exclusively on domestically produced chips, marking a major milestone in China's push for AI self-sufficiency. The facility operates multiple clusters with over 10,000 chips each to train GLM models, demonstrating that Chinese labs can scale frontier AI development despite US export controls blocking access to Nvidia's most advanced processors.
Chinese AI lab Z.AI has finished construction of a massive AI data centre that runs entirely on Chinese-made chips, providing one of the clearest answers yet to a question that has dominated discussions about China's AI capabilities: can domestic silicon truly power frontier model training? The facility, which has begun partial operations, represents a 1-gigawatt hub designed to train Z.AI's GLM models, according to Bloomberg
1
. The scale is striking—a gigawatt roughly equals the power draw of 750,000 homes at any moment, placing this site among the largest server hubs any Chinese AI lab has constructed.
Source: The Next Web
The company, formerly known as Zhipu, now operates several computing clusters, each containing more than 10,000 chips, with crucially none of them being Nvidia's products
1
. While Bloomberg did not specify the exact chip models, China's leading designer of AI accelerators is Huawei, which competes with local firms such as Cambricon Technologies and Alibaba to close the performance gap on Nvidia. This buildout demonstrates that domestically produced chips can handle the computational demands of training advanced AI models, not just running inference workloads.The news triggered a significant rally in China AI chip stocks on Tuesday, reflecting growing investor confidence that the country's AI ecosystem is becoming less dependent on imported AI hardware. NAURA Technology Group jumped the daily 10% limit, while Cambricon Technologies climbed 8.5% and Semiconductor Manufacturing International Corp (SMIC) surged 8.8% in Hong Kong trading
2
. Other semiconductor stocks also posted gains, with BOE Technology Group and Luxshare Precision Industry each rising 6.1%, while Foxconn Industrial Internet gained 4.8%.The rally extended beyond chip manufacturers to the broader technology sector, lifting the CSI 1000 Index by 3.5% and pushing the Shanghai Composite nearly 1% higher
2
. Evidence that a major Chinese AI developer can deploy large-scale AI infrastructure using only locally designed chips is being viewed as validation of the country's semiconductor supply chain, benefiting equipment makers, chip designers, and manufacturers alike.US export controls have blocked China's AI labs from purchasing Nvidia's most powerful chips, forcing a fundamental question about whether homegrown parts could carry the load for training frontier models. A 1-gigawatt cluster built entirely on domestic silicon provides a concrete answer, suggesting Chinese labs can continue scaling even while cut off from the chips the rest of the industry treats as default
1
.The timing sharpens the strategic implications. The buildout arrives days after Beijing-based Moonshot AI's Kimi K3 matched top US models, then ran short of compute capacity and paused new sign-ups. Z.AI is racing the same rivals and betting on owning its AI hardware rather than relying on external suppliers. The company has positioned itself as an enterprise AI market supplier, drawing comparisons to Anthropic in its business model approach.
Related Stories
Z.AI's achievement fits within a broader national strategy. China is preparing to invest around 2 trillion yuan, approximately $295 billion, over five years on data centers across the country, with cloud giants Alibaba and China Telecom remaining the biggest builders so far
1
. This massive capital deployment underscores Beijing's commitment to building a self-sufficient AI ecosystem that can compete globally despite geopolitical strategies aimed at limiting its access to advanced technology.Z.AI has the financial resources to compete in this landscape. Fresh from a Hong Kong listing and a follow-on share sale, the company is on track for $1 billion in annual recurring revenue, which would make it the first Chinese AI firm to reach that milestone
1
. Its shares jumped almost 20% following the data centre news, reflecting market enthusiasm about its prospects.While the achievement is significant, several caveats warrant attention. The chip details come from an anonymous source, and Z.AI has not officially confirmed them or responded to requests for comment. Building the cluster is not the same as proving it can train a frontier model as efficiently as an Nvidia-powered system. Domestic chips can still lag on raw performance metrics, and stitching thousands of them into a stable, reliable system presents substantial engineering challenges
1
.The real test will be the next GLM models release and how it compares with what US labs ship in terms of capabilities, training efficiency, and cost-effectiveness. Investors and industry observers should watch whether Z.AI can demonstrate that its domestically powered infrastructure can produce models competitive with those trained on Nvidia's H100 or upcoming B200 chips. The company's ability to scale its enterprise AI market presence while maintaining model quality will determine whether this infrastructure investment translates into sustainable competitive advantage.
Still, the direction is clear. The effort to build a Chinese alternative to Nvidia has moved from concept to a working, gigawatt-scale facility. That alone shifts the debate about how effectively export controls can constrain China's AI ambitions, suggesting that restrictions may accelerate domestic innovation rather than halt progress.
Summarized by
Navi
[1]
10 Jul 2025•Technology

28 Aug 2025•Technology

01 Apr 2026•Business and Economy

1
Technology

2
Policy and Regulation

3
Science and Research
