For most of the internet era, there was a fairly clear division inside the tech world.
Software companies built applications, and infrastructure companies built the systems that those applications ran on.
Today, artificial intelligence is blurring those lines.
When we talk about AI, like we did all last week, we mostly focus on the software side. New, more powerful models and increasingly capable AI assistants are easy to assess because they show up directly in the products we use.
But underneath all that progress, a larger shift is taking place in tech.
Building and running AI systems requires enormous physical capacity. AI needs computing clusters, networking hardware, power contracts and facilities designed to operate at industrial scale.
I’ve written before about how this growth is creating infrastructure constraints.
But those constraints don’t just shape engineering decisions. They also shape capital allocation.
And right now, capital spending is one of the clearest ways to see how aggressively the largest tech companies are competing to stay at the forefront of AI.
Spending Big on the Future
Today’s chart tracks capital spending by Microsoft (Nasdaq: MSFT), Alphabet (Nasdaq: GOOG), Amazon (Nasdaq: AMZN) and Meta (Nasdaq: META) from 2022 through 2025.
What stands out isn’t simply that annual expenditures have gone up.
It’s how fast they’ve gone up, plus the fact that all four companies are making a similar leap at the same time.
Put simply, AI is no longer just a software story. The global AI boom has pushed the largest technology platforms into what is effectively a shared infrastructure buildout.
Microsoft has been investing heavily to expand Azure, its cloud platform, and to support AI services for business customers.
Meta, once criticized for infrastructure overspending, is now expanding data center and compute capacity as AI becomes core to its product strategy.
Amazon continues to channel significant cash flow through AWS into physical capacity, with analysts noting its outsized contribution to industrywide capital expansion.
Alphabet (Google) has treated custom infrastructure as a competitive lever for years, funding everything from large server footprints to in-house chips designed to support AI workloads.
In other words, infrastructure is starting to be a main driver in the competition between these tech giants.
The spending patterns in today’s chart reflect that shift. And recent guidance and estimates suggest that the pace of this buildout is accelerating.
These four hyperscalers are now on track to spend around $665 billion in 2026. This represents a significant increase from earlier in the decade, when comparable capital spending across this group totaled closer to $100 billion per year.
But that was before AI accelerated the demand for computing capacity.
In fact, quarterly infrastructure investment from these four companies has already jumped about 77% year-over-year, which gives you a sense of how quickly these buildouts are moving.
That money is going into data centers, networking gear, custom chips, land, power agreements and cooling systems, the operational backbone required to run AI workloads every day.
Projects like these are planned years in advance. Which means these companies are spending a lot of money on AI before it’s making them much money in return.
But they don’t seem to have a choice.
Because analysts estimate that global data-center expansion tied to AI could require $7 trillion in investment by 2030.
Today’s chart reflects the early stages of this trajectory.
Naturally, investors are watching all this spending closely. Companies talk about infrastructure spending on nearly every earnings call now, and their stocks often react when those plans change.
And if the recent tech sell-off is any indication, I expect AI will be a contentious topic until it starts to become a real driver of revenue.
Here’s My Take
The point of this chart isn’t which company spends the most in a given year.
It’s what this spending tells us about where the industry is heading.
When companies with very different business models begin investing in similar ways, it reflects a shared expectation about future demand. In this case, demand for AI computing power is pushing technology back toward infrastructure.
Building this infrastructure is expensive. It requires the kind of money and long-term planning that only a handful of companies can sustain at scale.
And it doesn’t guarantee success. AI is still early, and it’s not making much money yet.
But that’s not the point.
The companies building capacity now are putting themselves in position to move faster if-and-when adoption accelerates tomorrow. Because as AI continues advancing toward more autonomous and capable systems — and eventually to artificial superintelligence — the limiting factor won’t just be software.
It’ll be access to computing power.
That’s exactly what all this spending is buying.
Regards,
Ian King
Chief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to dailydisruptor@banyanhill.com.
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!