Amazon CEO Andy Jassy has revealed that AWS, the company's cloud-computing division, has already booked most of its large-capacity contracts through 2027, with significant commitments extending into 2028. Speaking on a video podcast hosted by venture capital firm Andreessen Horowitz, Jassy described demand for AI infrastructure as "insatiable," noting that even AWS capacity for 2026 is effectively sold out.
The comments reframe the narrative around AI cloud computing. Instead of worrying whether customers will show up, the key question now is how fast Amazon can build data centers and secure enough advanced chips to deliver on what's already been reserved. This is a shift from earlier concerns about AI demand being speculative or overhyped.
What's driving the capacity crunch?
AI workloads require massive computing power, and companies across industries are racing to secure access to the specialized chips and infrastructure needed to train and run large language models. AWS, as the largest cloud provider, is a primary beneficiary of this demand. Jassy's comments suggest that customers are signing multi-year contracts to lock in capacity, a trend that mirrors what other cloud providers have reported.
Amazon has already pointed to roughly $220 billion in capital expenditure for 2026, and Jassy hinted the company could spend even more. This level of investment underscores the scale of the opportunity—and the risk. Building data centers and acquiring chips requires enormous upfront spending, while the revenue from those investments arrives gradually over time.
AWS does keep some compute capacity available for on-demand users and for services like Bedrock (a platform for accessing AI models) and EC2 (virtual servers). But Jassy said the "big chunks" of longer-term capacity are largely committed, meaning most of the new infrastructure is already spoken for.
What it means for investors
For investors, the news is a double-edged sword. On one hand, pre-booked multi-year capacity acts like a backlog, giving AWS unusually clear visibility into future demand. This can make growth look less fragile and reduce uncertainty about whether the massive investments will pay off. It also signals that AI demand is real and sustained, not just a passing trend.
On the other hand, the focus now shifts to execution and cash flow. Amazon must build data centers efficiently, secure enough chips, and ramp up utilization to turn those bookings into revenue. The accounting treatment also matters: depreciation on new assets will hit the income statement before the associated revenue fully lands, which could pressure operating margins and free cash flow in the near term.
Investors are likely to watch Amazon's third-quarter results, scheduled for October 29th, for updates on timelines—when the booked capacity will come online and how quickly it will translate into revenue. The company's ability to manage this transition will be a key driver of its stock performance in the coming quarters.
The broader context is that AI cloud demand is reshaping the competitive landscape. Rivals like Microsoft and Google are also investing heavily in AI infrastructure, and Oracle's recent investor day highlighted similar concerns about margins. Meanwhile, Nvidia's $1.1 billion prepayment to secure chipmaking capacity underscores the supply constraints that could limit how fast any provider can expand.
For everyday investors, the key takeaway is that Amazon's AI bet is large and long-term. The company is spending heavily now, with the payoff expected over several years. This is typical for capital-intensive industries, but it means that near-term earnings may be less impressive than the long-term growth story suggests. As always, it's important to consider how this fits into your overall portfolio and risk tolerance.
Amazon's situation is not unique. SAP recently reported a record quarter with a 26% jump in cloud backlog, and Nokia saw AI and cloud sales surge 123%. These examples show that the AI infrastructure boom is broad-based, but they also highlight the challenge of turning hype into profits.
As the earnings season unfolds, Amazon will be one of several tech giants reporting, and investors will be parsing every detail about AI capacity and spending. The message from Jassy is clear: demand is not the problem. The question is whether Amazon can deliver.
