AI infrastructure spending could surpass $1.3 trillion in 2027, according to S&P Global, and investors are right to question how much of it will ever return a profit. But two operators are already converting that capex into revenue at scale: Amazon and Microsoft, as reported by Yahoo! Finance on October 25, 2026.
The skepticism is rational. Not every model, agent, or AI-infused product will generate substantial revenue. What separates the winners is not the size of the spend but control of the distribution layer, and right now two protocols own it.
Amazon bets $220 billion on AWS
Amazon raised its capital expenditures to $220 billion this year, with most of it flowing into AI infrastructure. The bet is backed by a market position that competitors cannot realistically dislodge: AWS holds 28% of the cloud computing market, according to Synergy Research Group.
AWS now runs at an annualized revenue rate of $169 billion, with $16.6 billion in operating income and an operating margin above 39%. CEO Andy Jassy said he expects AWS revenue to reach $600 billion within a decade, a target that, if approached, would make the current spend cycle look cheap in retrospect.
Microsoft spends $175 billion, Azure accelerates
Microsoft is running a parallel program. Its capex reaches $175 billion this year, and Azure is responding: the service carries a $100 billion annualized revenue run rate, up 41% year over year, with management guiding to 45% cloud revenue growth in fiscal 2027.
Intelligent cloud sales, which include Azure, account for roughly 44% of Microsoft's total revenue. Management has stated plainly that demand continues to exceed available capacity, which signals that the constraint is supply, not appetite.
Why a duopoly compounds
AWS at 28% and Azure at 21% form what is effectively a two-node network at the base layer of the AI economy. Their leads were established years ago, and the capital requirements of entry now function as a permission barrier no challenger can clear.
Every AI workload, from training runs to inference to agent orchestration, ultimately executes on rented compute. The landlords of that compute collect rent regardless of which model wins. This is the structural insight the market is repricing: the value of AI may accrue less to model builders than to infrastructure owners.
Microsoft holds the ecosystem edge
Between the two, Yahoo! Finance identifies Microsoft as the stronger position, and the reasoning is architectural. Microsoft has sold more than 30 million paid seats for its Copilot AI agent software, a 50% jump in several months.
Because Copilot integrates with Azure, Microsoft can offer enterprise customers a full stack: agents, cloud, and enterprise software in one contract. Amazon matches the compute layer but not the application layer. In ecosystem terms, Microsoft controls both the protocol and the client.
What the spending wave means for the market
The broader signal for investors is that capex skepticism and capex returns are not contradictory positions. Most AI spending will not pay off. The market is in a filtering phase, and the filter is operating margin. Companies that own the infrastructure layer and show unit economics, like AWS's 39% operating margin, are already demonstrating return. The Motley Fool notes that both companies have accelerated cloud revenue through AI spending.
Yahoo! Finance's analysis, also syndicated by BizToc, frames the question correctly: not whether AI spending is excessive in aggregate, but which operators hold positions that convert spend into durable revenue.
FAQ
How much are companies spending on AI infrastructure?
AI companies have already invested hundreds of billions of dollars in capital expenditures this year, and S&P Global projects spending could surpass $1.3 trillion in 2027.
How much is Amazon spending on AI?
Amazon raised capex to $220 billion this year, with most of it going to AI, while AWS generates a $169 billion annualized revenue run rate.
Which cloud provider is bigger, AWS or Azure?
AWS leads with 28% of the cloud market, per Synergy Research Group, while Microsoft Azure holds 21%.
Why is Microsoft considered the better AI stock?
Microsoft pairs Azure growth with more than 30 million paid Copilot seats, giving it an AI software layer that Amazon currently does not match.