A tangled web of deals stokes AI bubble fears in Silicon Valley

At DevDay OpenAI This week, OpenAI chief Sam Altman did something American tech bosses rarely do these days: he actually answered questions from reporters.

“I know it's very tempting to write a story about a bubble,” Mr. Altman told me, sitting surrounded by his top aides. “There are actually a lot of parts of AI that I think are currently in development.”

In Silicon Valley, the debate over whether artificial intelligence companies are overvalued has taken on new urgency.

Skeptics privately, and some now publicly, wonder whether the rapid rise in the value of artificial intelligence technology companies could be, at least in part, the result of what they call “financial engineering.”

In other words, there are concerns that these companies are overvalued.

Mr. Altman said he expects investors will make some bad decisions and dumb startups will walk away with crazy amounts of money.

But with OpenAI, he told me, “there is something real going on here.”

Not everyone is convinced of this.

In recent days, warnings about an AI bubble have come from the Bank of England, the International Monetary Fund, and JP Morgan CEO Jamie Dimon told the BBC “the level of uncertainty must be higher in the minds of most people.”

And here, in a country often called the technology capital of the world, concerns are growing.

At a panel discussion at the Silicon Valley Computer History Museum this week, early AI entrepreneur Jerry Kaplan told a packed audience that he had lived through four bubbles.

He's especially concerned now, given the size of the money on the table compared to the dot-com boom. We have something to lose.

“When [the bubble] breaks, it will be very bad, and not just for people working in artificial intelligence,” he said.

“It will drag down the rest of the economy.”

But Professor Anat Admati of Stanford Graduate School of Business, which has produced its fair share of tech entrepreneurs, says that while there have been many attempts to simulate a situation where we're in a bubble, it may be a futile exercise.

“It’s very difficult to time the bubble,” Professor Admati told me. “And you can't really say you were in it until the bubble burst.”

But this data worries many.

Artificial Intelligence Businesses amounted to 80% staggering growth of the US stock market this year, according to Gartner global spending on AI is likely to reach a whopping $1.5 trillion (£1.1 trillion) before the end of 2025.

OpenAI, which brought AI to the consumer market with ChatGPT in 2022, is at the center of a tangled web of deals that are drawing scrutiny.

Last month, for example, it struck a $100 billion deal with chipmaker Nvidia, itself the world's most valuable publicly traded company.

It expands on existing investments Nvidia has already made in Mr. Altman's company, with OpenAI expected to build data centers based on Nvidia's cutting-edge chips.

Then on Monday, OpenAI announced plans to buy billions of dollars worth of artificial intelligence development equipment from Nvidia rival AMD in a deal that could make it one of AMD's largest shareholders.

Remember that this is a private company, albeit one recently valued at half a trillion dollars.

There's also tech giant Microsoft, which has major investments, and cloud computing giant Oracle has a $300 billion deal with OpenAI.

OpenAI Stargate Project in Abilene, Texas, funded with help from Oracle and Japanese conglomerate SoftBank and announced at the White House during President Donald Trump's first week in office, is growing larger every few months.

And as for Nvidia, it has a stake in artificial intelligence startup CoreWeave, which provides OpenAI with some of its huge infrastructure needs.

And as these increasingly complex funding mechanisms become more common, experts here in Silicon Valley say they could cloud perceptions of demand for AI.

Some people are also free to talk about it, calling the deals “circular finance” or even “vendor finance” – where a company invests or lends to its own customers so they can continue making purchases.

“Yes, investment loans are unprecedented,” Mr. Altman told me on Monday.

But, he added, “it's also unprecedented for companies that are growing revenues so quickly.”

OpenAI's revenue is growing quickly, but it has never been profitable.

And it's hardly a good sign that people I've talked to keep mentioning Nortel, a Canadian telecom equipment maker that borrowed heavily to finance deals for its customers (and thereby artificially inflated demand for its products).

For his part, Nvidia's Jensen Huang defended his deal with OpenAI on CNBC on Monday, saying the firm is not obligated to buy his company's technology with the money he invests.

“They can use it to do whatever they want,” Huang said.

“No exclusivity. Our main goal is simply to support them and help them grow – and develop the ecosystem.”

Mr. Kaplan says he sees There are several clear signs that the AI ​​sector – and therefore the economy as a whole – may be in trouble.

In tough times, he said, companies announce big initiatives and plans to produce products for which they don't yet have the capital.

Meanwhile, retail investors are eager to get involved in the startup project.

The rise in AMD stock this week could indicate that investors are trying to get a piece of the ChatGPT wealth machine – and while all this is happening, the actual physical infrastructure is being built to feed the seemingly insatiable hunger for further AI development.

“We are creating a new man-made environmental disaster: huge data centers in remote places like deserts that will rust and release harmful substances into the environment, and there will be no one to hold accountable because the builders and investors will be long gone,” Mr. Kaplan said.

But even if we're in a bubble, Silicon Valley is hopeful that the investments being made now won't necessarily go to waste.

“I take solace in the fact that the Internet was built on the ashes of yesterday's overinvestment in telecommunications infrastructure,” said Jeff Boudier, who builds products at the AI ​​community hub Hugging Face.

“If there is overinvestment in infrastructure for AI workloads, there may be financial risks associated with it,” he said.

“But it will create a lot of great new products and experiences, including ones we don’t think about today.”

Many believe in the potential of AI to change society.

The question is money to finance the ambitions of the sector's leading companies may be drying up.

“Nvidia looks to be the last lender or investor,” said Richard Jark, founder of the newsletter UncoverAlpha.

“Who else has the ability to invest $100 billion in another company right now?”