To The Moon

Every earnings season has a dominant theme.

This quarter, it wasn't AI demand.

It was whether the companies spending hundreds of billions on AI infrastructure could continue doing so without breaking their financial model.

After listening to the latest earnings calls from $Microsoft(MSFT)$ $Alphabet(GOOG)$ $Meta Platforms, Inc.(META)$ my takeaway is simple:

The AI buildout isn't slowing. If anything, the conviction behind it is getting stronger.

It starts with nearly $2 trillion of backlog.

The combined commercial backlog across the major hyperscalers now stands at roughly $2 trillion.

A significant portion of those contracts is expected to convert into revenue over the next 12–24 months.

But that's only part of the story.

Many of today's contracts were signed before AI infrastructure became as scarce as it is today. As those agreements expire, they are likely to be renewed in a market where compute capacity commands materially higher pricing.

That creates two powerful growth drivers at the same time:

• Existing backlog gradually turning into revenue.

• Future renewals benefiting from stronger pricing.

When both volume and pricing move higher together, revenue growth becomes much easier to sustain.

Microsoft proved massive AI spending doesn't have to come at the expense of cash flow.

Microsoft continues to invest at an extraordinary pace.

Management expects quarterly CapEx to exceed $50 billion next quarter, with annual spending around $175 billion, while 2027 investment plans remain broadly unchanged.

Yet despite that scale of spending, Microsoft still expects to remain free cash flow positive, supported by the strength of its core business.

That may be one of the most important signals from this earnings season.

Customer demand still exceeds available compute capacity, according to management.

In other words, Microsoft isn't building infrastructure in hopes that customers eventually arrive.

The customers are already there.

Another detail worth noting is where the money is being invested.

Roughly two-thirds of quarterly CapEx went toward shorter-lived assets such as CPUs and GPUs, giving Microsoft flexibility as hardware evolves. The company also extended the estimated useful life of its data centers from 15 years to 25 years, improving long-term capital efficiency.

Microsoft also confirmed it will be among the first cloud providers deploying next-generation AI infrastructure based on AMD Helios and NVIDIA Vera Rubin, reinforcing that the next investment cycle is already taking shape.

Meta isn't seeing a slowdown either.

Meta kept annual CapEx in an extraordinary $130–145 billion range while emphasizing that industry-wide compute remains in short supply.

Management went even further.

The company disclosed that it has received offers to purchase compute capacity at significant premiums to its own cost, highlighting just how valuable existing AI infrastructure has become.

Meta also explained that additional compute isn't simply being reserved for future possibilities.

It already has multiple internal use cases capable of generating attractive returns from additional capacity.

Perhaps the clearest message came from management's view of the broader market:

Industry capacity remains below demand, and that imbalance is expected to persist for the foreseeable future.

Google just added even more fuel to the story.

Alphabet delivered perhaps the clearest signal that the AI infrastructure race is still accelerating.

Management raised full-year 2026 CapEx guidance to $195–205 billion, up from the previous $180–190 billion range.

At the same time, the company reiterated that capital spending is expected to be significantly higher again in 2027.

That's not the language of a company preparing to slow investment.

It's the language of a company that still sees demand expanding faster than available infrastructure.

Google Cloud continues to benefit from strong AI-driven demand, and management made it clear that additional investment is necessary to keep up with customer requirements.

Rather than stepping back after another year of record spending, Alphabet chose to increase its investment plans once again.

Three earnings calls. One message.

Strip away the headlines, and all three companies are pointing in the same direction.

Microsoft says customer demand continues to exceed available capacity.

Meta believes industry-wide compute will remain constrained for the foreseeable future.

Google raised CapEx guidance again while already preparing for another step higher in 2027.

Different companies.

Different business models.

The same conclusion.

The AI infrastructure buildout is still accelerating.

Why I'm still constructive.

Many investors continue to focus on how much money hyperscalers are spending today.

I'm more interested in what those investments become tomorrow.

The infrastructure being built today won't simply support existing AI workloads.

It becomes the foundation for future cloud services, enterprise AI, advertising, inference, developer platforms and entirely new applications that don't yet exist.

Meanwhile, the core businesses funding these investments remain exceptionally healthy.

Cloud continues to expand.

Enterprise software continues to grow.

Advertising continues to generate enormous cash flow.

Those businesses are financing the next generation of AI infrastructure while AI itself gradually becomes another major revenue engine.

If capital expenditure growth naturally moderates over the next several years—as infrastructure deployment matures—while the revenue generated from today's investments continues to scale, the free cash flow profile of the hyperscalers could improve meaningfully.

That's why I believe the recent AI selloff has focused too heavily on today's spending and not enough on tomorrow's earnings power.

The market may be debating the size of AI CapEx.

The companies writing the biggest checks don't appear to be debating demand at all.