[Photo: Shutterstock]

Investor sentiment toward AI appears to be steadily freezing ahead of Anthropic's initial public offering, which is proceeding on a record scale.

The fallout from Jensen Huang's resignation as Nvidia CEO amid an extramarital relationship and spying scandal is continuing. A report has also lowered its outlook for AI infrastructure investment.

Companies that have relied on Nvidia's financial backing are now facing an urgent situation after Huang's resignation.

The Wall Street Journal (WSJ) reported that Huang had also been laying groundwork in his final months as Nvidia CEO to scale back funding for the AI industry.

Colette Kress, the chief financial officer (CFO) who took the helm at Nvidia on an interim basis after Huang, is a figure who had been sceptical about some of the deals Huang had pushed. As a result, startups that had expected Nvidia support worry that backing will be cut even more, the WSJ reported.

UBS analyst Tim Arcuri estimated in a recent report that if Nvidia scales back financial support, more than $50 billion could leave the market. He also analysed that, with interest rates rising, it could become even harder for companies to secure funding.

Axios reported that Joe Brusuelas, chief economist at RSM, said Huang's exit could burst the AI bubble as a whole.

Most of the money Nvidia invested in AI-related companies is then used to buy Nvidia GPUs again.

So-called circular financing, in which a company that receives investment uses that money to buy products from the investor, has been cited as one of the main factors stoking concerns about an AI bubble. Nvidia was a representative company that led circular-financing deals in the AI arena.

ㆍWin-win or bubble...Big Tech divided over 'AI circular financing'

The International Monetary Fund (IMF) has also warned about circular financing spreading among AI companies.

CNBC reported on Nov. 16 local time that IMF Managing Director Kristalina Georgieva estimated the scale of circular financing was a little over $1 trillion, and that two-thirds of it was funded by debt. She said, "The global economy is leaning on one leg called AI, and if this axis shakes it could spread into a financial stability issue."

She said, "If investment confidence in the AI sector suddenly breaks and money flows out, economic growth could also weaken." She added that investment enthusiasm around AI is running ahead of results actually proven by the technology, and that as valuations rise, some adjustment is inevitable.

U.S. Federal Reserve Chair Kevin Warsh also warned that if the AI boom falls short of expectations, stock prices could fall and unemployment could rise. In opening remarks at a Fed conference on Nov. 16 on AI's impact on productivity and the labour market, Warsh said, "After massive capital investment and stock price gains, expected returns could falter and stock prices could fall," adding, "The economy could slow and unemployment could rise noticeably."

So far, infrastructure investment has driven growth in the AI market. Astronomical sums poured into infrastructure investment on expectations that demand for AI models and services would surge, and this led to rapid growth for companies selling AI infrastructure such as Nvidia. Even when AI companies' revenue models were questioned, the infrastructure investment boom itself did not subside, but the recent mood is different.

With concerns about overinvestment, some big tech companies including Microsoft are moving into a more cautious mode on AI infrastructure investment. There is also analysis that AI infrastructure investment may not be as large as expected.

Semiconductor market research firm SemiAnalysis has also sharply lowered its AI infrastructure investment outlook.

In a report released on Nov. 13, SemiAnalysis lowered its 2028 forecast for global capital spending on AI accelerators to $758 billion from $1.054 trillion, a 28 percent cut. It reduced its 2030 forecast to $923 billion from $1.531 trillion, a cut of about 40 percent. It lowered its forecast for cumulative silicon wafer shipment area for 2026 to 2030 by 17 percent, its 2028 forecast for high-bandwidth memory (HBM) shipments by 30 percent, and its 2030 forecast for power demand by 34 percent.

SemiAnalysis cited three reasons for the downgrade: new annual recurring revenue (ARR) growth at frontier AI research and development companies slowed in the third quarter, a worsening fundraising environment, and the possibility that data centre construction projects under way could be delayed due to related regulations after U.S. midterm elections.

Keyword

#Anthropic #Nvidia #IMF #SemiAnalysis #AI accelerators
Copyright © DigitalToday. All rights reserved. Unauthorized reproduction and redistribution are prohibited.