About that ‘AI bubble’ — Microsoft doesn’t see one.

The alarms and flashing red lights are warning we’re in an AI bubble. Hundreds of billions of dollars are being invested in the technology by Microsoft, Google, Amazon, Meta, Nvidia, Intel, and generative AI (genAI) startups like Anthropic, OpenAI, and others.

More and more, economists, analysts, and even tech companies worry: What if the genAI bubble will be the biggest tech bust of all time?

A McKinsey report warns: “Nearly eight in 10 companies report using gen AI — yet just as many report no significant bottom-line impact. Think of it as the ‘genAI paradox.’” An MIT report, The GenAI Divide: State of AI in Business 2025, adds that 95% of genAI pilots in businesses are failing. And the data and analytics firm S&P Global is warning that 42% of companies abandoned most of their AI pilots by the end of last year, up 17% from 2023.

Gary Marcus, a founder of two AI companies, has been sounding the call loudest of all, most recently in a column for The New York Times, warning that the countless billions spent on genAI will be wasted because there will be little payback. Instead, he argues, companies like Microsoft should instead put their money in more targeted ways for specific purposes like medical research.

Microsoft begs to differ. It’s essentially been saying: “AI bubble? What AI bubble?” and doubled down on its AI spending. The company’s not alone. Amazon, Google, Meta, OpenAI and all the other AI believers are doing the same thing.

Follow the money

To understand why Microsoft continues to bet big on AI, take a look at its spending plans and the customer demand for its AI services. Microsoft execs laid them out during the company’s most recent earnings call, which covered the first quarter of the 2026 financial year.

Microsoft reported in the call that its capital spending for the quarter jumped to $34.9 billion, driven primarily by AI spending. That’s about a 50% increase over the $24.2 billion in capital spending in the previous quarter.

And that’s just the beginning. It says it’s going to spend more than $80 billion in capital spending, and possibly much more. All that spending, the company says, is because of surging demand for AI.

That doesn’t even include the deal it made to invest $15.2 billion in the United Arab Emirates for AI infrastructure. So in all likelihood, spending in the 2026 fiscal year will be far beyond $80 billion.

Amy Hood, Microsoft’s finance chief, said the company has to keep spending that and more because it can’t keep up with customers’ demands for its AI services.

“I thought we were going to catch up,” she said during the call. “We are not. Demand is increasing. It is not increasing in just one place. It is increasing across many places.”

She added that Microsoft already has $400 billion under contract for future sales. “That’s for booked business,” she said. “Today.” 

The $400 billion is an understatement. It doesn’t include $250 billion in computing power for AI that OpenAI has agreed to buy from Microsoft.

Bubble or no bubble?

So is there an AI bubble? And if so, what will be the effects on Microsoft if it pops?

Most people agree there is a bubble. Even those with the most to gain from AI think we’re in one. OpenAI CEO Sam Altman believes it will eventually burst. And when it does, he says, “Someone is going to lose a phenomenal amount of money. We don’t know who, and a lot of people are going to make a phenomenal amount of money.” 

No surprise: He believes that he and OpenAI will be among those making “a phenomenal amount of money.”

He’s right that there’s a bubble. Bubbles are an inevitable part of the way technology makes its way into the economy and people’s lives. Venture capitalists invest in many companies during bubbles, knowing that most will fail. But even if only a handful of their investments hit it big, VCs more than make up for their bad bets. That’s the phase we’re in right now.

The last time this happened in a big way was 25 years ago, at the height of the dot-com bust. Companies that died during it deserved to die — nothing supported their sky-high valuations. Pets.com, the poster child for the bust, left investors with $369 million in losses. By the fourth quarter of 1999, it had spent $61.8 million and had only $5.8 million in sales. The company went from its IPO to liquidation in a mere 268 days.

The dot-com bubble bursting, though, didn’t mean that there was no money to be made on the internet, or that it wouldn’t change the way people live and do business. Our world has changed dramatically because of the internet. Trillion-dollar fortunes have been made on it and continue to be made.

The same thing will happen with the AI bubble. Companies with no real value will fail; those that offer value will thrive.

How will the bubble affect Microsoft?

Expect Microsoft to be one of those that flourish. Unlike many AI companies, it’s got lots of interested customers and significant AI revenue — as the company noted in its earnings report, it can’t keep up with demand.

The company is also aware that genAI might not pay out in a big way, but that other kinds of AI will. So it’s betting big on helping companies build AI agents that do things such as advanced data analysis, task automation and decision making. Microsoft is also targeting specific tasks, such as for medicine. More will follow.

So, bubble or not, Microsoft will likely thrive no matter what comes next.