Alibaba has unveiled "China's most powerful AI chip." Is this a threat to Nvidia?

Alibaba Group, a Chinese IT giant, has unveiled what it claims is the “most powerful” Chinese artificial intelligence chip. It is expected to serve as an accelerator capable of competing with Nvidia’s products and to enable a large-scale expansion of data center capacity in the coming years, according to Bloomberg.

On the back of this news, Alibaba's shares jumped nearly 2% in Hong Kong and 0.79% in U.S. after-hours trading, while Nvidia's shares fell 0.5%.

Alibaba unveiled its new chip at the annual Apsara Conference, the flagship event of the company’s cloud division, held in Hangzhou, according to CNBC. The chip in question is the next-generation AI chip, the Zhenwu V900, which, according to the company, delivers three times the performance of its predecessor, the — the Zhenwu M890 — and can be clustered into groups of up to 500,000 units to train the most powerful models.

“The new accelerator is equipped with 216 GB of memory <...> and inter-chip bandwidth of 1,200 GB per second, and is designed for powerful AI tasks,” the company’s press release states. According to Alibaba, the Zhenwu V900 is capable of “significantly reducing the cost of inference (generating responses for end users—Oninvest note) while increasing computational density.”

Mass production and the commercial launch of the chip are scheduled for the first quarter of 2027. Zhenwu’s existing chips are already being used by more than 650 customers across various industries, including finance, energy, the automotive industry, and others.

The company also announced plans to expand the global capacity of its data centers to more than 20 gigawatts by 2032. Although this is a massive project for the Chinese company, it is still relatively smaller than some other similar projects being implemented around the world, Bloomberg notes. For example, SoftBank plans to build a 10-gigawatt data center in Ohio alone, the agency points out.

In addition, the Chinese tech giant announced plans to develop an AI model with 5–10 trillion parameters, which is significantly more than that of its current competitors.

Overall, Alibaba plans to spend more than $53 billion over three years to expand its capabilities in the field of AI, and also expects to increase its annual revenue from cloud technologies and AI fivefold—to $100 billion—over the next five years. “Last year, we stated that the more powerful AI becomes, the stronger humanity will be. Today, I still stand by that idea,” said the company’s CEO, Eddie Wu, at the conference.

Alibaba's presentation took place ahead of the summit in Washington between U.S. President Donald Trump and Chinese President Xi Jinping, at which AI issues will be one of the key topics, according to Bloomberg. Nvidia CEO Jensen Huang and Microsoft CEO Satya Nadella are among those expected to attend the state dinner at the White House. China’s recent advances in AI give Beijing more leverage in the upcoming talks between Xi and Trump, The Independent reports, citing the views of several unnamed analysts.

Amid ongoing U.S. export controls that prevent Chinese companies from accessing Nvidia’s most advanced processors, Alibaba—like some other Chinese IT companies, including Huawei— is focusing on developing its own chips. As a result, external customers already account for a significant portion of Alibaba’s cloud division’s output, which may indicate a growing domestic ecosystem of Chinese companies willing to rely on local AI processors.

Before the export restrictions were imposed, China was one of Nvidia’s major markets outside the U.S.—the country accounted for about 13% of its revenue, but in May 2026, Huang stated that the company’s share of the AI market in China had “dropped to zero.”

This article was AI-translated and verified by a human editor