This is general information, not investment advice.
You use "the cloud" constantly — every email, stream and app — usually without knowing whose machines you're renting.
What the cloud actually is
Cloud computing means renting computing — servers, storage, databases, software — from a provider's data centers over the internet, instead of buying and running your own hardware. You pay for what you use, like a utility. It comes in three tiers: IaaS (infrastructure as a service — raw virtual machines and storage you build on), PaaS (platform as a service — managed tools to build and run apps), and SaaS (software as a service — finished apps like Gmail or Salesforce, where the provider runs everything). Companies moved en masse because the cloud turns big upfront hardware costs into pay-as-you-go capacity that scales in minutes and reaches the whole globe.
The big three
The infrastructure market is dominated by three hyperscalers — operators running data centers at the scale of tens of thousands of servers across dozens of regions. Per Synergy Research: Amazon Web Services (AWS) holds roughly 29% of the global market, Microsoft Azure about 20%, and Google Cloud around 13% — together about 63% of all cloud infrastructure spending. The market is huge and fast-growing: global quarterly cloud spend reached around $107 billion in late 2025 and roughly $129 billion by early 2026 (up about 35% year on year), with Canalys tracking similar momentum. In their latest quarters, AWS reported about $37.6 billion in revenue (+28%), Google Cloud crossed $20 billion (+63%), and Azure grew around 39%.
The economics
Cloud is a high-margin, recurring-revenue business — the profile markets love. AWS has long been Amazon's profit engine, subsidizing its thinner retail margins; Azure anchors Microsoft's commercial cloud; Google Cloud turned profitable in 2023 and has widened margins since. The revenue is sticky because of vendor lock-in: once a company builds on a provider's proprietary tools and APIs, moving to a rival is slow, risky and costly — so customers tend to stay and spend more.
The AI connection
The cloud and the AI boom are now inseparable. Training and running large AI models requires vast clusters of specialized chips (mainly Nvidia GPUs) housed in these same hyperscaler data centers, so AWS, Azure and Google Cloud are the main gateways to AI compute. Microsoft's OpenAI ties, Amazon's Anthropic investment and Google's own models each bind AI strategy to cloud revenue — and to enormous capital spending (Microsoft and Amazon each guided to roughly $80 billion-plus of capex), reinforcing the barriers that keep smaller rivals out. (Anthropic, an Amazon partner, also makes this publication's AI tools; we note it for transparency.)
The risks
Concentration is the core worry: when a major AWS region goes down, big chunks of the internet can fail at once, because so much depends on one provider. Regulators in the EU and UK are probing whether hyperscaler dominance in cloud — and now AI — harms competition. Lock-in means migrations cost more than firms expect. And the capital intensity (billions in data centers, land and power) entrenches the big three even as Oracle and a wave of AI-focused "neoclouds" nibble at the edges. For investors and businesses, the takeaway is that a handful of companies now own the plumbing of the digital — and increasingly the AI — economy, with all the profit, power and systemic risk that implies.
