Chief Executive Eddie Wu introduced the chip at Alibaba’s annual Apsara Conference in Hangzhou, placing proprietary semiconductors at the centre of the company’s expanding artificial intelligence strategy. Alibaba said the V900, designed by its T-Head semiconductor unit, can be linked in clusters of as many as 500,000 chips to support the training of frontier AI models.
Wu described the V900 as the most powerful AI chip currently available in China. That assessment is Alibaba’s own performance claim and has not been independently benchmarked against every competing domestic accelerator. The company said mass production is expected to begin by early 2027.
The announcement comes as China’s largest technology groups accelerate development of domestic processors amid continuing United States restrictions on exports of advanced AI chips and semiconductor manufacturing technology. Alibaba, Huawei and other companies have been building alternatives intended to reduce dependence on Nvidia hardware for training and running large AI models.
Alibaba’s chip push is closely tied to a broader expansion of its cloud infrastructure. Wu said Alibaba Cloud aims to increase its global data-centre capacity to more than 20 gigawatts by 2032, a scale that would require substantial investment in computing hardware, power and networking systems.
The company has already committed 380 billion yuan, equivalent to about $53 billion, over three years to cloud computing and artificial intelligence infrastructure. It raised additional capital through a Hong Kong share offering in August as spending on data centres, models and semiconductor development increased.
Alongside the V900, Wu said Alibaba plans to train a new foundation model containing between 5 trillion and 10 trillion parameters. That would make it roughly two to four times larger by parameter count than the company’s current flagship Qwen 3.8 Max model, which has about 2.4 trillion parameters.
Alibaba has not disclosed a timetable for releasing the larger model, and parameter count alone does not determine model quality or capability. Wu said the company is pursuing improvements across model architecture, chips and computing infrastructure as it seeks to handle more complex tasks over longer periods.
The V900 follows the Zhenwu M890, which Alibaba introduced in May for both AI training and inference. The M890 has 144 gigabytes of high-bandwidth memory and inter-chip bandwidth of up to 800 gigabytes per second, according to company specifications. Alibaba said that processor delivered three times the performance of the earlier Zhenwu 810E.
T-Head has become an increasingly important part of Alibaba’s effort to control more of the technology stack underpinning its cloud and AI businesses. The company has said its in-house chips can be deployed in Alibaba Cloud data centres and supplied through servers to external customers, giving it greater flexibility as access to some foreign processors remains constrained.
Alibaba said earlier this year that more than 560,000 Zhenwu chips had been shipped to over 400 customers across sectors including finance and automotive. The figures were disclosed by the company and provide an indication of the scale at which its domestic accelerator business has grown.
Shares in Alibaba rose more than 5 per cent in Hong Kong trading on Tuesday, reaching a one-month high as investors responded to the company’s latest AI announcements. The stock has become increasingly sensitive to expectations surrounding Alibaba’s cloud growth, capital spending and ability to monetise artificial intelligence services.
The V900 also enters a domestic market where Huawei is advancing its Ascend processor roadmap and large-scale computing clusters. Huawei said this month that it plans to introduce new Ascend chips from 2027 and is developing systems capable of linking large numbers of processors for demanding AI workloads.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
