(Updates to show Anthropic declined to comment in 16th paragraph.)
The world's largest artificial intelligence firms could earn $200 billion in revenue this year, more than five times the $37 billion they brought in last year, according to estimates from Menlo Ventures.
The venture capital firm has tracked AI revenue growth from Microsoft Corp. (MSFT), Amazon (AMZN), Alphabet's Google (GOOG, GOOGL), OpenAI, Anthropic, Cursor and other companies for the past three years.
Revenue, which in 2024 was $11.5 billion, is on track to blow previous years' numbers out of the water, Derek Xiao, a principal at Menlo and co-author of the firm's annual AI research report, said in an interview with.
"We've always actually had a trend of underestimating how these things grow because it's hard to predict an exponential, but I would put it at $200 billion," he said.
While that number is an estimate, growth in AI revenue since January "hockey sticked" and is based on new models of AI now being used, he said.
"Instead of just this call and response chat pattern that we've seen traditionally, you have background agents that can run for minutes or hours at a time, and that unlocks an order of magnitude more of both things that it can do, but also spend on some of these AI tools," Xiao said.
The revenue growth is at the heart of an ongoing debate as company spending has raised fears of an AI bubble. Capital expenditures by Google, Amazon, Microsoft and Meta Platforms Inc. (META) -- collectively known as hyperscalers because they offer massive cloud computing services and global data center infrastructure -- is expected to be around $800 billion this year, with another $1 trillion in 2027, according to the companies and analyst estimates.
The investment in new data centers, software and equipment was so large in the first quarter that it accounted for about two-thirds of the growth in US gross domestic product, according to data from the US Bureau of Economic Analysis.
While revenue growth has been robust, the capital outlays are still sparking fears that spending has gotten too far ahead of future earnings potential.
"Most enterprises are yet to generate any returns from their AI spending," James Covello, head of Global Equity Research at Goldman Sachs, said in a note to clients this week.
"The companies making the models and the hyperscalers building the AI infrastructure are burning through cash and boosting their borrowing. While semiconductor companies are seeing record revenue and profits, the overall dynamic is 'unprecedented and unsustainable,'" Goldman Sachs said in a summary of Covello's report.
Covello cited a Massachusetts Institute of Technology study last year that said 95% of AI investment has had no effect on company earnings. While AI tools helped improve individual worker efficiency, "the core barrier to scaling is not infrastructure, regulation, or talent.
"It is learning," the MIT study said. "Most GenAI systems do not retain feedback, adapt to context, or improve over time."
Still, the revenue growth in the AI sector is unlike anything most analysts have seen in other technologies including the advent of the personal computer or the Internet. On the consumer side, AI adoption "has been spectacular," Covello said in his note. He cited a Stanford Institute of Human-Centered AI study that found that 53% of consumers have adopted generative AI tools within three years of the release of ChatGPT.
"The bull-bear gap on AI is wider than almost anything I've written about," said Philipp Dubach, a strategy consultant and independent researcher who has written widely about AI.
He cited reports that showed OpenAI went from $2 billion in annualized run-rate revenue to $24 billion in 24 months and Anthropic growth from $1 billion to $30 billion in 15 months as "unprecedented growth rates." Anthrophic declined to comment and OpenAI didn't immediately respond to a request for comment Friday.
Yet "the math on capex still doesn't close at any plausible revenue figure I can build up to," he said.
Dubach estimates that year-to-date 2026 revenue for AI is about $100 billion shared between Microsoft, Anthropic, OpenAI, Amazon and Google. That figure accounts for double-counting that's common in the industry, he said. Many AI systems are integrated into each other and their services can overlap.
The capex spending, however, "is a staggering amount of money chasing returns that haven't fully landed," Dubach said.
In the three months through March, Microsoft reported 30% revenue growth in its Intelligent Cloud segment, with sales for its cloud-based computing Azure service up 40%. Google Cloud revenue rose 63% from a year earlier; Amazon Web Services was up 28%.
Microsoft and Amazon declined to comment to, while Google didn't respond to a request for comment.
"The thing is, can these companies grow their revenue fast enough to fund the AI build that's required?" said Bruce Murray, CEO and chief investment officer at the Murray Wealth Group.
Murray, who owns all four stocks, said the companies have other business lines to generate the capital to spend on building out their AI infrastructure.
"It's going to be really difficult to tell who gets over their skis a little too far," he said. As a long-term investor Murray said he has confidence that the AI play will pan out.
"Meta's maybe the one we'd be the most concerned about, but still, on a relative basis we're still sticking with it," he said. Meta's first-quarter ad revenue rose 33%. The company didn't respond to a request for comment.
The Facebook parent "is growing nicely in their advertising business, but it seems to be a bit behind on getting something achievable with AI that's actually going to generate the money," Murray said.
One area of AI revenue that has yet to emerge is retail users. A separate Menlo Ventures report from June 2025 found that while 61% of consumers it surveyed had used AI in the past six months, only 3% of users were paying for it. Still, companies including Google and Amazon can earn advertising revenue from those users who aren't yet paying. "My wife uses ChatGPT for everything before she sends it out," but doesn't pay for it, Murray said.
As a venture capital firm, Menlo Ventures invests in some of the companies it analyzes such as Anthropic, Wispr Flow, OpenRouter, Numeric and others.
Menlo's Xiao said the firm saw very different activity in AI revenue compared with what the MIT report concluded last year.
"Part of our report was sort of standing in opposition to the MIT report, pointing to the real use cases and the real enterprise dollars that are flowing into the ecosystem and being spent and actually transforming how work is done," Xiao said. "AI looks a lot different from previous waves that we've seen where there has been maybe irrational exuberance."
The demand for the tech that's driving the spending may appear "scary," he said.
"If you contrast that to the build out of the telco boom in the early 2000s, they were laying thousands of miles of fiber that would not be used for years," Xiao said. "There's a difference between this time as opposed to last time, that I think does make it quite exciting, at least from our view. It feels like this time is much more sustainable."
Matthew Leising
Price: $426.79, Change: $+17.36, Percent Change: +4.24%