For the past two years, Big Tech has asked investors to accept a pretty extraordinary bargain:
Let us spend hundreds of billions of dollars building AI infrastructure today… and trust that the profits will eventually follow.
For a while, Wall Street was happy to oblige.
AI demand was exploding. Nvidia couldn’t make chips fast enough. Data center capacity became scarce. And companies like Microsoft, Amazon, Google, and Meta kept increasing their spending plans.
But lately, investors have started asking a harder question:
Where is the return on all this money?
It’s a fair question.
The biggest technology companies are now expected to spend more than $700 billion this year as the AI infrastructure arms race continues. And the spending is becoming so aggressive that even some of the world’s most profitable companies are seeing pressure on their free cash flow.
Alphabet (GOOG), for example, spent nearly $45 billion on capital expenditures in the second quarter alone—and reported negative free cash flow for the first time.
Investors have been waiting for proof that these enormous investments can actually produce an equally enormous financial payoff.
This earnings season, we finally got some.
And it showed up somewhere a lot of investors weren’t looking…
AI’s $370 billion side effect
Most discussions about AI monetization focus on the obvious stuff.
How many people are paying for ChatGPT?
How much can Microsoft charge for Copilot?
Those businesses matter.
But they’re only one part of the AI economy.
Behind practically every AI application is an enormous amount of computing infrastructure.
Models need to be trained. Data needs to be stored and processed. Applications need computing power every time someone asks a chatbot a question, creates an image, runs an AI agent, or analyzes a database.
And businesses generally aren’t building all that infrastructure themselves.
They’re renting it from companies like Amazon, Microsoft, and Google.
The result: The latest cloud numbers are incredible.
Google Cloud revenue surged 82% year over year last quarter, reaching $24.8 billion. CEO Sundar Pichai specifically credited strong demand for AI infrastructure and AI solutions.
Microsoft’s Azure cloud business grew 43% in its latest quarter. Azure generated more than $100 billion in revenue for Microsoft’s full fiscal year for the first time.
And Amazon Web Services (AWS) grew 37%—its fastest growth rate in 18 quarters—bringing the business to a $169 billion annualized revenue run rate. Its AI business within AWS alone has now surpassed a $25 billion annual revenue run rate and is growing at a triple-digit percentage.
Put the three businesses together, and you’re looking at roughly $370 billion in annualized cloud revenue.
And demand is outrunning supply.
Google Cloud’s backlog—the amount of contracted business that has yet to be recognized as revenue—has exploded to $514 billion.
And earlier this year, Microsoft said customer demand for Azure continued to exceed available supply, even as the company raced to bring more computing capacity online. Management said it expected those capacity constraints to continue through at least 2026.
In other words, AI adoption isn’t replacing the cloud…
It’s making the cloud more valuable.
And that helps explain something investors have spent the past year struggling to understand.
Why Big Tech keeps spending
Every time one of these companies announces another huge jump in AI spending, investors get nervous.
Again, that’s understandable.
But the most important question isn’t simply how much Big Tech is spending. It’s whether they’re able to monetize that spending—and cloud growth is starting to provide the answer.
Look at Amazon.
AWS produced $42.2 billion in revenue last quarter—and $16.6 billion in operating income.
That’s an operating margin of roughly 39%.
Google Cloud generated $24.8 billion in quarterly revenue and $8.8 billion in operating income.
Those are extraordinary economics.
And they’re getting better even as these companies pour money into AI infrastructure.
Recent analysis found Microsoft, Amazon, and Google collectively added roughly $56 billion in annualized cloud revenue over the past year—and more than half of the incremental revenue translated into operating profit.
That’s a much different picture than a company spending billions of dollars hoping customers eventually arrive.
Customers are already there. The bigger problem is finding enough capacity to serve them.
Big Tech owns the toll roads
Everyone is obsessed with figuring out who will “win” AI.
Will OpenAI dominate?
Will Anthropic take the lead?
Will some model we haven’t heard of yet leapfrog all of them?
But Amazon, Microsoft, and Google have a major advantage regardless of which individual AI product wins.
They own the infrastructure the winners need. Think of them as the toll roads of the AI economy.
Thousands of companies can compete to build the best AI applications. But those applications need enormous amounts of cloud computing. Whether a company builds with OpenAI, Anthropic, Gemini, or a model of its own, the underlying workload still needs somewhere to run.
The more AI applications people use, the more traffic moves across that infrastructure.
The cloud providers can collect revenue from the entire ecosystem.
The bottom line
Big Tech’s capital commitments are staggering. The five largest hyperscalers are spending hundreds of billions in combined investment, and several are seeing free cash flow squeezed as spending accelerates.
Investors should absolutely keep watching whether those returns justify the spending.
But there’s an important distinction between companies pouring money into an unproven idea… and companies increasing capacity because customers are already lining up for more.
Cloud growth suggests the hyperscalers are increasingly in the second category.
So as earnings continue to roll in, don’t just watch capex. Watch cloud growth, margins, backlog, and whether demand continues to exceed the infrastructure these companies can bring online.
For two years, Big Tech has asked investors to trust that the profits would eventually follow the spending.
Cloud growth is becoming the strongest evidence yet that they are.
Want to hear more about this earnings season… and what it reveals about the AI trade?
Frank and Daniel break down the biggest market stories—and what they actually mean for your money—on Wall Street Unplugged.
