“AI Gold Rush Could End in a Crash”: HSBC and General Atlantic Warn of Billion-Dollar Bubble
Top CEOs say companies are overspending on AI with no clear path to profits — “It’s starting to look like irrational exuberance.”
The artificial intelligence boom that’s reshaping Silicon Valley — and the global economy — may be getting out of hand.
That’s the stark warning from HSBC’s CEO Georges Elhedery and General Atlantic’s Chairman and CEO William Ford, who told global investors this week that AI investments are running far ahead of actual returns.
Speaking at the Global Financial Leaders’ Investment Summit in Hong Kong, both executives cautioned that the hype around artificial intelligence — and the hundreds of billions of dollars being poured into it — may not yet be justified by real-world revenues or productivity gains.
“Revenues don’t justify the massive AI spending”
Elhedery didn’t mince words. “While computing power for AI is essential, current revenue profiles may not justify such massive spending,” he said.
The numbers tell the story. Tech giants Alphabet, Meta, Microsoft, and Amazon now expect to spend a combined $380 billion in capital expenditures (capex) this year — much of it to build out data centers and AI infrastructure.
And OpenAI, the company behind ChatGPT, has announced a staggering $1 trillion in infrastructure partnerships with Nvidia, Oracle, and Broadcom to keep up with exploding AI demand.
But Elhedery warned that consumers aren’t yet ready to pay for it, and that businesses will take years to see measurable productivity benefits.
“These are five-year trends,” he said. “We’ll start seeing real revenue benefits and readiness to pay for it later than investors expect.”
In other words: the payoff might come — just not anytime soon.
The trillion-dollar question: When will AI actually pay off?
Morgan Stanley estimated in July that global data center capacity will grow sixfold in the next five years, requiring about $3 trillion in new spending by 2028.
And according to a McKinsey report, data centers capable of handling AI workloads could demand $5.2 trillion in capital investment by 2030 — dwarfing the $1.5 trillion needed for traditional IT systems.
That’s an enormous bet on a technology still finding its footing in real-world business use.
“This is a 10- to 20-year play”
William Ford of General Atlantic echoed the warning.
“In the long term, AI will create entirely new industries and applications,” he said. “There will be a productivity payoff — but that’s a 10-, maybe 20-year play.”
Ford compared today’s AI boom to the early days of railroads or electricity — groundbreaking technologies that transformed economies but took decades to deliver consistent profits.
“You’re really betting on this being a broad-based technology, like railroads or electricity,” Ford said. “It’s going to reshape the economy, but predicting how in the first few years is nearly impossible.”
The risk of “irrational exuberance”
Ford also warned that the rush into AI could spark capital misallocation, overvaluation, and destruction of value — a phrase that recalls the dot-com bubble of the late 1990s.
“There’s definitely a risk of irrational exuberance in the early stages,” he said. “You need to pay up front for opportunities that may only materialize far down the road.”
In short: while AI is the future, not every AI bet will pay off.
Why the hype keeps growing anyway
Despite the warnings, the AI spending spree shows no sign of slowing.
The explosion of generative AI tools like ChatGPT, Gemini, and Claude has pushed companies to scale up computing capacity at breakneck speed. Cloud providers are racing to meet demand for high-performance chips from Nvidia, which now dominates the AI hardware market.
For Big Tech, the logic is clear: dominate AI now, and you own the future.
But for investors, that strategy comes with enormous upfront costs and uncertain payback timelines — a dangerous mix in an economy still navigating inflation and high interest rates.
“Consumers aren’t ready to pay for it”
Elhedery highlighted another challenge: AI’s economic model still doesn’t add up.
Right now, most users interact with AI tools for free — from chatbots to image generators — while the companies behind them burn through billions in cloud costs and energy bills.
Until people or businesses are willing to pay meaningful amounts for AI services, the massive capital spending could become unsustainable.
“The ramp-up is long,” Elhedery said. “We’ll see real readiness to pay probably later than investors expect.”
Investors are getting nervous
While AI stocks remain hot, Wall Street is starting to question just how profitable the technology really is in the short term.
Some analysts warn that the AI trade could be entering bubble territory, especially as companies spend heavily without clear monetization paths.
As one Hong Kong-based hedge fund manager put it: “Everyone wants in on AI, but no one can yet show the math.”
A reality check for the AI gold rush
For now, both Ford and Elhedery agree: the AI boom isn’t a mirage, but it’s far from a guaranteed windfall.
Yes, AI will reshape industries — from healthcare to finance to logistics — but not overnight. The next decade will likely bring a mix of breakthroughs and disappointments before the true winners emerge.
Until then, the message from two of the world’s top financial leaders is simple: don’t believe all the hype — and don’t bet your balance sheet on it.
