To The Moon

The biggest debate in technology right now is no longer whether AI demand exists.

It does.

The real question investors are asking is:

How much are companies willing to spend to capture that demand — and will the returns justify the investment?

Based on the latest earnings reports, the answer from the biggest technology companies is clear:

They are still aggressively building.

The AI factory buildout keeps accelerating

$Amazon.com(AMZN)$ became the latest hyperscaler to show the scale of this investment cycle.

The company reported $53 billion in capital expenditures during the second quarter, up 69% year over year, as it expanded AI data centers and supporting infrastructure.

That follows a similar pattern across the industry:

• $Meta Platforms, Inc.(META)$ increased costs significantly as it accelerated AI infrastructure spending.

• $Microsoft(MSFT)$ continued expanding its AI capacity.

• $Alphabet(GOOG)$ raised its investment plans again.

Together, Amazon, Google, Meta and Microsoft are expected to spend an enormous amount on AI infrastructure over the next two years as they build data centers and secure advanced computing resources.

This is no longer a traditional software cycle.

AI has transformed the tech industry from an asset-light business into one where physical infrastructure has become the foundation of growth.

Data centers are becoming the new factories.

The market is worried. The companies are focused.

The reaction from investors has been mixed.

Some companies have been rewarded for showing discipline around spending.

Others have been punished when investors worried that costs were rising faster than near-term profits.

Google faced pressure after reporting negative free cash flow in a quarter as infrastructure investment accelerated.

Meta also saw its shares fall after raising the lower end of its annual CapEx outlook and warning that costs would continue climbing.

But management teams across the industry have remained consistent:

They believe the opportunity is large enough to justify the spending.

The argument is simple:

If demand for AI compute is greater than available supply, then building more capacity directly creates more revenue opportunities.

Infrastructure is not being built for a hypothetical future.

Customers are already waiting for access.

Amazon shows both sides of the story

Amazon's latest results highlight the current AI investment dynamic.

The company generated strong operating performance:

• Revenue reached $200.6 billion, up 20% year over year.

• Profit more than tripled to $62.6 billion.

• AWS revenue grew 37% to $42.2 billion.

• Operating cash flow over the past year reached $161.4 billion, up 33%.

At the same time, heavy infrastructure investment pushed free cash flow lower, reaching negative territory after major spending on data centers and facilities.

This is the tradeoff every hyperscaler is facing:

Spend heavily today to secure AI capacity, with the expectation that those assets generate revenue for years.

The strongest signal: demand is still ahead of supply

The most important message coming from these earnings calls is not the size of CapEx.

It is the reason behind it.

Companies continue to say they do not have enough computing capacity to meet demand.

Microsoft highlighted strong Azure momentum, with Azure revenue surpassing a major milestone and growth accelerating.

The company continues expanding its global data center footprint, adding capacity across regions and preparing for the next generation of AI infrastructure.

Meta delivered a similar message.

The company is actively securing land, power and infrastructure for future AI workloads because available compute remains constrained.

The industry is effectively preparing years ahead because waiting could mean losing customers.

The backlog is becoming the hidden driver

One reason the spending continues is the growing amount of contracted demand already in the system.

Amazon, Google and Microsoft have accumulated significant backlogs, much of which is connected to AI-related demand from major customers and leading AI companies.

That backlog provides visibility.

However, it also means execution matters.

Companies need to convert infrastructure investment into profitable AI services over time.

The next phase of the AI race will not be determined by who can spend the most money.

It will be determined by who can turn that spending into durable revenue and cash flow.

Why the AI buildout is still moving forward

The current cycle looks expensive because it is expensive.

But history shows that major technology shifts often require massive upfront investment before the economic benefits become obvious.

The internet required networks.

Cloud computing required data centers.

AI requires compute infrastructure at an entirely different scale.

The hyperscalers are not treating AI as a side project.

They are building the foundation for what they believe will become one of the largest technology platforms in history.

The market may continue debating the cost.

But the companies spending the money are making their position clear:

They would rather build too much capacity than risk having too little.