Huawei says it cannot produce enough AI computing equipment to meet demand in China and is limiting sales abroad as a result. Rotating chairman Eric Xu detailed a faster chip roadmap and new computing architecture aimed at closing the gap with Nvidia.
Huawei cannot produce enough artificial-intelligence computing equipment to meet demand in China and is limiting overseas sales, rotating chairman Eric Xu said Thursday. The shortage comes as the Chinese tech giant steps up its challenge to Nvidia in AI chips.
Xu said testing of Huawei's Ascend 950DT chip had produced good results, and the company was in extensive discussions with Chinese AI model developers. He expected many of them to begin training models on systems using the chip next year.
Huawei prioritizes China over exports
Xu told reporters at the Huawei Connect conference in Shanghai that the company does not have enough capacity to satisfy Chinese demand, so it has no plan to expand into international markets in a fully-fledged way. Huawei supplies some countries where demand is particularly strong, he said, but volumes remain limited.
Xu also said reliable data on Nvidia's share of China's AI chip market was hard to obtain. According to Reuters: "I think Ascend market share should be bigger than Nvidia's." Huawei did not provide data to support the estimate.
Faster chip rollout planned
Huawei detailed its next generation of AI processors on Thursday. Rotating chairman David Wang said the Ascend 960DT chip will be ready in the first quarter of 2027, three quarters earlier than previously planned. The 960PR is due in the third quarter, one quarter ahead of schedule. Huawei intends to release a new generation of Ascend chips every year, with the 970 and 980 to follow in 2028 and 2029.
Linking chips into larger systems
Huawei is trying to compensate for limits on individual Chinese chips by connecting large numbers of processors so they work together as one system. The company said clusters of about 100,000 chips have become standard for training some of the largest AI models, while communication between machines can consume more than 40% of training time in conventional server systems.
The company unveiled a new computing architecture called Peerium, designed eventually to allow as many as 1 million processors to operate together as a single system. Its new Ascend 960 supernode would connect as many as 4,096 AI processors, with multiple supernodes then linked into clusters containing hundreds of thousands of processors.
The company has already deployed more than 1,000 systems using its earlier Ascend 910C processors, while its Ascend 950 systems have entered commercial use. More than 5,200 developers were active each month on software for its Ascend chips, and more than 40 AI models had been trained directly on its computing platform.
Nvidia retains a major advantage through its widely used CUDA platform, which lets developers build and run AI applications on its chips. The US chipmaker did not immediately respond to a request for comment.
Source: All News
Trading involves risk.
Risk Disclaimer
