These days, there’s a time during earnings calls when you can practically hear the math break in between the prepared remarks and the analyst questions. In late April, Amazon Web Services reported $37.6 billion in quarterly revenue, up 28%. The response was more of a gradual recalibration than a celebration. All of a sudden, spreadsheets based on “high-teens growth” required a new column. Investors were quietly rewriting their notes after dismissing AWS as the cloud’s slow-moving giant.
The cloud industry may have been undervalued for the majority of the previous year. Not even close. By a great deal.
The odd thing is that the story has been pointing in the opposite direction lately. AWS was the laggard in August of last year. While Google Cloud was approaching 32% and Microsoft Azure was clipping along at 26%, its growth had plateaued in the high teens. In just one week, the stock fell 8%. Amazon was losing market share, Andy Jassy had missed the AI window, and the company known for creating modern cloud computing had somehow arrived late to the most significant event in its own backyard, according to analysts at Maxim and other places. It was odd to watch that consensus form because anyone who had worked in an enterprise IT department could see that demand existed; it was just hidden by contract cycles and capacity restrictions that take quarters rather than days to resolve.
The floor then shifted. In the fourth quarter of 2025, AWS revenue grew at its fastest rate in thirteen quarters, reaching 24%. The order backlog, which indicates what is on the horizon rather than what has already been reserved, increased 40% annually to $244 billion. Then April’s figure came in at 28%, or $37.6 billion in a quarter, with an annualized run rate that was well over $150 billion. The curve’s shape had simply altered.
AWS’s growth is not the only factor distorting cloud valuations. It’s the seriousness of it. Every comparable company is repriced by association when the largest player by absolute revenue begins to accelerate rather than decelerate—that is, when the sector’s base case needs to be redrawn. When you consider that Google Cloud is expanding from a much smaller base, its 63% growth in the most recent quarter appears astounding. Microsoft Azure, once the AI-cloud darling, now appears to be excellent rather than unique. The entire group of peers strays.
Walking through this, it seems like investors are pricing in something they can’t quite put their finger on. The OpenAI agreement is one aspect of it; over the course of eight years, Amazon increased the agreement by $100 billion, and AWS became OpenAI Frontier’s sole third-party cloud distribution provider. In addition to Amazon’s $25 billion investment in the company, Anthropic has committed to spending over $100 billion on AWS over the next ten years. The chip story is part of it; Trainium and Graviton have a combined yearly run rate of over $10 billion, growing at triple-digit percentages, with Trainium3 already shipping and almost all of its 2026 supply committed. These are all major issues. When stacked, they start to resemble the infrastructure of a completely different type of business.
The danger still exists. Amazon has increased its capital expenditures from $131.8 billion to $200 billion for 2026. That is a huge wager, and as depreciation increases, margins will be compressed. According to Jassy, a large portion of 2026’s expenditures will only be monetized in 2027 and 2028. This timeline puts investor patience to the test in a market that prefers quarters to years. The forward P/E ratio of 29.2x is high. Zacks gives the stock a hold rating.
However, the numbers continue to come in, and they consistently exceed any models. It’s difficult to ignore the fact that about six months prior to AWS’s most notable reacceleration, there was a lot of skepticism. The valuation math for the cloud industry is currently bending around the growth curve of a single company, whether that is a sign or a coincidence, and no one is quite sure how to price what comes next.
