The artificial intelligence gold rush might be moving too fast for its own good. As Big Tech giants pour hundreds of billions into AI, some of the world’s most powerful CEOs are starting to sound the alarm.
At the Global Financial Leaders’ Investment Summit in Hong Kong, HSBC CEO Georges Elhedery and General Atlantic CEO William Ford warned that AI investment has become dangerously disconnected from reality. Their message was clear: the spending boom is massive, the profits are not.
Elhedery said that while building computing power for AI is essential, the revenue simply isn’t there to justify the record-breaking spending. Companies, he warned, are betting heavily on a future that may take years to pay off.
Tech giants are leading the charge. Alphabet, Meta, Microsoft, and Amazon together are expected to spend more than 380 billion dollars this year, most of it on data centers, chips, and cloud infrastructure to fuel AI demand. OpenAI, the company that kicked off the AI frenzy with ChatGPT, has announced an eye-popping one trillion dollars in infrastructure deals with partners like Nvidia, Oracle, and Broadcom.
The problem, according to Elhedery, is that consumers aren’t ready to pay for AI at the scale companies expect, and businesses won’t see major productivity gains right away. “These are five-year trends,” he said. “Real revenue benefits and readiness to pay for it will come later than investors expect.”
In other words, companies are spending like it’s 2030, but the returns may not show up until then.
William Ford, chairman and CEO of General Atlantic, agreed that the industry is still in the early innings. “In the long term, you’re going to create a whole new set of industries and applications, and there will be a productivity payoff,” he said. “But that’s a 10- or 20-year play.”
Ford compared today’s AI craze to the early years of railroads and electricity—revolutionary technologies that eventually reshaped the economy, but only after decades of uneven returns and market shakeouts. He warned that the early stages of these transformations often come with misallocated capital, overvaluation, and what he called “irrational exuberance.”
“You’re really betting on this being a broad-based technology,” Ford said. “It’s going to reshape the economy, but it’s very hard to predict how in the first few years.”
The scale of investment is staggering. Morgan Stanley estimates that global data center capacity will grow sixfold in the next five years, with the total cost hitting three trillion dollars by 2028. McKinsey predicts that data centers built for AI workloads will need more than five trillion dollars in capital by 2030—over three times what traditional IT systems will require.
The spending spree reflects Big Tech’s race to dominate the next generation of computing. Companies are scrambling to secure access to the most powerful AI chips and build data centers that can handle the exploding demand for machine learning models. But as costs climb, many analysts are starting to question whether the economics add up.
AI models are expensive to train and even more expensive to run. Every query to an AI chatbot consumes significantly more power and computing resources than a typical Google search or social media post. That’s why companies are spending billions to expand infrastructure—but without clear paths to monetize the technology, those investments could strain profits for years.
Elhedery believes the market has overestimated how quickly consumers and corporations will pay for AI-powered products. “We will start seeing real revenue benefits later than investors expect,” he said. “Consumers are not ready to pay for it, and productivity benefits won’t appear in a year or two.”
Ford added that while AI will eventually deliver massive returns, the early phase will be marked by volatility and overhype. “This sector will be capital-intensive initially,” he said. “You need to pay up front for the opportunity that’s going to come down the road.”
The warning comes as some investors are already starting to draw parallels between today’s AI boom and past tech bubbles. Just like the dot-com era, there’s real innovation happening—but also an undeniable element of speculative mania.
OpenAI’s trillion-dollar infrastructure push, combined with massive cloud and chip investments from Amazon, Google, and Microsoft, has created what some analysts are calling an “AI arms race.” Everyone wants to be first, even if the payback could be a decade away.
For now, the hype continues to fuel spending. Nvidia, which makes the high-performance chips that power most AI systems, has seen its market value skyrocket over the past two years. Meanwhile, startups in AI software, robotics, and automation are raising billions on the promise of future breakthroughs, even when many have little or no revenue.
Still, the reality check from HSBC and General Atlantic stands out. Both executives agree that AI will be transformative—but they’re urging investors and companies to take a breath.
“This is going to reshape industries,” Ford said, “but it’s not a one-year story. It’s a generation-long evolution.”
In other words, AI might indeed change everything—but it won’t happen overnight. Between now and that future lies a decade of trial, error, and enormous spending that may not immediately pay off.
So while Silicon Valley and Wall Street chase the next great AI opportunity, some of the world’s most seasoned business leaders are reminding everyone of a hard truth: revolutions take time. And this one, they say, could get very expensive before it gets profitable.
Suggested headline optio
