Alibaba chair warns of AI data centre bubble amid massive buildouts: report

Alibaba chair warns of AI data centre bubble amid massive buildouts: report

AI-generated image of generic data centre servers, with bubbles emerging from the rack - representing an emerging investment bubble

As tech giants pour hundreds of billions into global data center expansions, Alibaba's chair Joe Tsai warns of an emerging investment bubble.

Bloomberg reports the Alibaba Group chair telling the HSBC Global Investment Summit in Hong Kong this week that he was “astounded” by the numbers being “thrown around” in terms of AI data centre investments.

“I start to see the beginning of some kind of bubble,” Tsai said. “I start to get worried when people are building data centres on spec. There are a number of people coming up, funds coming out, to raise billions or millions of capital.”

Data centre buildout has exponentially increased in recent months, with hyperscalers, real estate developers, and data centre operators all looking to cash in on the increasing demand for AI and digital services.

That demand, according to Tsai, isn’t there yet: “People are talking, literally talking about $500 billion, several 100 billion dollars. I don’t think that’s entirely necessary.

“I think in a way, people are investing ahead of the demand that they’re seeing today, but they are projecting much bigger demand,” Tsai said.

In the week leading up to Tsai’s comments in Hong Kong, Oracle pledged $5 billion to expand cloud infrastructure in the UK, while Nvidia and Elon Musk’s xAI joined a Microsoft, SoftBank and MGX fund which plans to invest $100 billion in AI data centres.

Other projects with sizable financial commitments beyond Stargate include France’s rival €109 billion (USD$ 117 billion) buildout, Meta’s reported $200 billion data centre expansion plan, and the $35 billion effort to build the largest data centre in the world in Korea.

There have, however, been some signs of uncertainty, notably with Microsoft reportedly cancelling multiple data centre leases in the US over fears it's in an “oversupply position” — just weeks after it revealed plans to spend half of its $80 billion AI data centre war chest in the country.

Tsai’s comments came as Alibaba too is looking to double down on AI, focusing on developing open source AI that enterprises can use to power their applications, such as its Tongyi Qianwen, or Qwen, family of models.

Alibaba has also developed a platform similar to Amazon’s Bedrock, which lets users access a variety of generative AI models, including more than 5,000 models created by the Chinese conglomerate.

Tsai reportedly told attendees Alibaba was undertaking a “reboot” as it looks to push further into the AI market after a period of significant regulatory scrutiny, both at home and abroad.

RELATED STORIES

Are we scaling fast enough for the AI data centre boom?

Data: AI set to outpace cloud, big data in data centre infrastructure needs

Apple, Alibaba Cloud join board of group developing GPU interconnect standards

Gift this article