If you needed proof that the AI infrastructure boom is for real, the first quarter 2026 earnings season delivered it in black and white. Alphabet posted $109.9 billion in revenue, up 22% year-over-year. Google Cloud hit $20.03 billion, up 63%. Microsoft posted $82.9 billion in revenue, up 18%, with its AI business now running at a $37 billion annual rate. The numbers are not gradually improving. They are compounding.

And behind the revenue headlines, a capital expenditure story is unfolding that has no modern precedent. Alphabet guided its 2026 capex to $180 billion to $190 billion. Microsoft said it expects to spend $190 billion this year alone. Meta raised its range to $125 billion to $145 billion. Combined, the four largest cloud and AI players are on track to spend more than $600 billion on AI infrastructure in 2026. That is a lot of GPUs, data centres, and networking gear.

(For context, the entire Australian economy is expected to generate around $2 trillion in GDP this year. Four American technology companies are spending more than a quarter of a trillion dollars on servers and chips in a single year.)

What the money is buying

Microsoft's CFO Amy Hood gave the most specific breakdown. About two-thirds of its capex is going to GPUs and CPUs, driven by Azure customer demand and internal AI tools like M365 Copilot. Even with $40 billion-plus in quarterly capex, Microsoft says it expects to remain capacity constrained through the end of 2026. That is not a company worried about overspending. That is a company that cannot build fast enough.

Alphabet's CFO Anat Ashkenazi was equally direct. The company is seeing, in her words, "unprecedented internal and external demand for AI compute resources." Google's Cloud backlog hit $462 billion, nearly doubling in a single quarter. She expects just over half that backlog to convert into revenue over the next 24 months.

These are not projection slides. These are signed contracts with enterprise customers.

The market rewarded one, punished another

Here is the part that makes the picture more complicated. Both Alphabet and Microsoft crushed expectations. Yet their market reactions could not have been more different.

Alphabet shares rose nearly 7% after hours. Microsoft was essentially flat. Meta dropped more than 6% after hours.

The difference came down to clarity of return. Alphabet's Google Cloud delivered a "meaningful beat" in the words of analysts, growing 63% year-over-year and indicating the company is gaining market share from AWS and Microsoft Azure. When analysts asked Mark Zuckerberg to explain how Meta measures return on its $125 billion to $145 billion capex programme, his answer was technically precise but did not reassure investors the way Alphabet's numbers did.

"The formula for our company has always been to build experiences that can get to billions of people and focus on monetizing them once you get to scale," Zuckerberg said. Investors wanted to know when scale would translate into cloud-style recurring revenue. The answer was not satisfying to everyone.

What the infrastructure buildout means for businesses

The most underappreciated angle of this earnings season might be the backlog number. When a company signs a $1 billion-plus cloud contract, that does not show up as revenue immediately. It sits in backlog and converts over quarters and years. Alphabet's backlog nearly doubled to $462 billion in a single quarter. Microsoft Cloud has a massive backlog of its own.

For businesses, this means AI infrastructure partnerships are becoming long-term commitments. Enterprises are locking in multi-year compute agreements with the major cloud providers, betting that AI capabilities will become as foundational as electricity or internet connectivity. That is a significant shift in how corporate IT strategy is being built.

The supply chain is also tightening. Memory and component costs are rising, which Microsoft attributed roughly $25 billion of its capex increase to. This is not unique to one company. Every major buyer is competing for the same silicon, and that competition is pushing prices up across the board.

Meanwhile, OpenAI is having a different conversation

It is worth stepping back to notice the contrast. While Google and Microsoft were delivering blowout results, OpenAI was reportedly missing its own internal revenue and user targets. CFO Sarah Friar reportedly warned that the company may struggle to fund future compute contracts if revenue does not keep pace with its obligations.

The market reaction was swift. Oracle dropped around 4%, CoreWeave sank more than 5%, and SoftBank, one of OpenAI's largest investors, fell roughly 10% during Tokyo trading. These are not small, obscure names. They are central players in the AI compute chain.

What this suggests is that the AI boom is not uniform. Building foundation models is expensive and the path to profitable scale is far from clear. Infrastructure, on the other hand, has clear demand, long-term contracts, and compounding revenue. That is the bet the big tech incumbents are making with $600 billion in 2026 capex.

What comes next

The AI infrastructure cycle is still in its early innings. Microsoft expects to stay capacity constrained through 2026. Alphabet is guiding capex to "significantly increase" in 2027. Meta is still building out its AI compute layer. Amazon has not shown its full hand yet.

For business readers and investors, the signal is clear. The companies best positioned to benefit from AI are not necessarily the companies building the flashiest models. They are the companies with the capital to build the underlying infrastructure that every other AI company depends on.

The next quarter of earnings will tell us whether the backlog converts as cleanly as the companies expect. Until then, the buildout continues. And it is not slowing down.