Amazon не снижает расходы на дата-центры ради ИИ — инвесторы только «за»

Amazon is not slowing down data center investments for AI — and the market approves, reports TechCrunch on July 30, 2026. Unlike most tech companies where high CapEx weighs on stock prices, cloud providers in the AI era operate under a different logic.

Why investors aren't afraid of Amazon's spending?

Cloud providers occupy a unique position in the AI ecosystem: they provide infrastructure for the entire industry without competing directly in the model race. Amazon Web Services earns revenue on every compute request — whether it's training a model, deploying it to production, or storing data. Growing AI adoption by corporations and startups automatically translates into revenue growth for the cloud platform.

Investors describe this position as the "picks and shovels strategy": in a gold rush, it's more profitable to sell tools to all participants than to mine for gold yourself. In the AI economy, the "shovels" are compute capacity, GPU clusters, and managed cloud services. Amazon's CapEx expansion is perceived not as margin pressure, but as building capacity to meet confirmed demand.

A key factor is the durability of demand itself. Back in 2022–2023, some analysts allowed that the AI hype might pass quickly. That didn't happen: enterprises across industries are systematically migrating workloads to the cloud and embedding AI tools into production processes. Demand for compute is growing not in cyclical spikes, but along a sustained upward trajectory.

How Wall Street's perception of CapEx has changed

A few years ago, major capital expenditure announcements from tech companies were met with market skepticism: high CapEx weighed on operating margins and worried investors accustomed to "asset-light" growth models. Today the picture is different.

Cloud provider capacity is fully loaded, and queues for available GPU clusters stretch for months. In such an environment, restraining construction means consciously forgoing revenue and ceding ground to competitors. The market perceives aggressive CapEx not as recklessness, but as foresight: a company that builds sufficient capacity today gains a competitive advantage in the next cycle. For Amazon, which holds the leading share of the public cloud market, this argument is particularly compelling — scale allows for more efficient distribution of infrastructure costs.

Who else is ramping up data center investments

Amazon is not the only hyperscaler in this race. Microsoft, whose partnership with OpenAI requires constant expansion of compute capacity, and Google with its own AI development vertical are also scaling their capital programs. All three are competing for the same limited resources: construction sites in regions with cheap energy, access to network infrastructure, and GPU supplies for AI computing — a scarce asset that limits the pace of scaling.

As TechCrunch notes, investors approve of this race from all participants, because the infrastructure being built today will directly determine AI capabilities for the next decade.

What this means

The market's positive reaction to Amazon's rising spending reflects a structural shift in investor logic: cloud infrastructure has moved from the category of "heavy costs" to that of "strategic assets" in the AI era. For the AI industry as a whole, this means that compute power will continue to concentrate among a small number of hyperscalers — and this trend will only intensify.