$37.59 billion. That's what AWS generated in a single quarter — up 28% year-over-year, the fastest rate of growth in 15 quarters. I added AMZN at $185 after Q4 2025 results. That thesis has not changed. This earnings report made it stronger.
The Number That Surprised Everyone
AWS grew 28% in Q1 2026. The Street expected somewhere between 22-24%. The beat was not marginal — it was structural.
Operating income for AWS came in near $11.5 billion, implying a margin above 30%. For context: AWS alone now generates more operating profit in a single quarter than most S&P 500 companies produce in a full year. Total Amazon revenue hit approximately $155 billion for the quarter, with EPS beating consensus by a meaningful margin.
The 28% growth figure is worth pausing on. AWS crossed $100 billion in annualized revenue while accelerating. The base effect was supposed to slow it down. It didn't. That tells you the demand is real, not a restatement of prior-period backlog.
The Competitive Position
The cloud war is a three-way race, and the spread is widening:
- AWS: $37.6B quarterly, +28% YoY — ~31% market share
- Azure: +33% YoY, with AI contributing 16 percentage points of that growth — ~25% share
- Google Cloud: $12.3B, +28% YoY — ~12% share
Azure's 33% growth looks better on the surface. But Azure started from a smaller revenue base, and AI contributions are inflating its headline number. When you strip the AI tailwind out of Azure's growth, organic momentum is closer to 17-18%. AWS grew 28% without needing to break that out separately — because the underlying business is accelerating on its own.
GCP at $12.3B is growing fast from a low base. The real story there is what happens to Google Cloud enterprise deals as Gemini matures. I'll cover that in the Alphabet Q1 piece. For now, AWS's moat holds.
Why the Capex Bet Is the Thesis
Amazon has committed over $200 billion in capital expenditure for 2026 — the largest single-year infrastructure bet in the company's history. Most analysts flag this as a risk. I flag it as the thesis.
Here's the logic: every data center AWS builds is a reason for enterprise customers to run more workloads on AWS, not fewer. The capital isn't being spent into a vacuum — it's being deployed against confirmed demand. AWS Bedrock, Amazon's managed AI platform, has seen accelerating adoption across enterprise accounts. Customers are not just experimenting with AI anymore; they are running production workloads at scale. The infrastructure needs to be there before those workloads can scale.
The capex-as-moat argument works like this: to compete with AWS in 2028, you would need to commit $200+ billion today. Microsoft is doing exactly that. Google is doing it. No one else is. The capital commitment is itself a barrier to entry — it pre-empts competition for the next five years of enterprise AI workloads.
Cloud capex is not a cost. It is a reservation system for future revenue. Every dollar locked into AWS infrastructure today is a customer commitment that compounds.
What I'm Watching
Three metrics govern whether this thesis holds or breaks:
1. AWS margin trajectory. Operating margins need to stay above 29-30%. If AI workload pricing compresses margins below that band as the company fills excess capacity, the thesis weakens. Q1 at 30%+ is exactly where it needs to be.
2. Bedrock adoption velocity. Enterprise customers moving from proof-of-concept to production AI is the demand signal that justifies $200B of capex. If adoption plateaus at pilot stage, that capex looks speculative. The Q1 data — acceleration, not deceleration — is the right signal.
3. Q2 and Q3 revenue guidance. AWS guided conservatively through 2025 and over-delivered every quarter. If guidance for Q2 2026 implies a deceleration back toward 22-24%, it suggests the Q1 beat was partially pull-forward. Watch the next earnings call language on AI backlog and enterprise pipeline.
The Risk I'm Watching Most
If enterprise AI spending disappoints — if companies find that their AI deployments don't generate the productivity gains that justified the spend — AWS margins compress as capacity underutilization rises. The $200B capex becomes a liability instead of an asset.
This is not my base case. But it is the scenario that would break my thesis. I will be watching AI productivity metrics across enterprise customers through 2026 and into 2027. If the signal turns, I'll act on it.
My Position
I added AMZN at $185 after Q4 2025 results. The thesis at entry: AWS is the most defensible infrastructure business ever built, AI workloads are accelerating enterprise cloud adoption, and Amazon's self-funding model — where AWS profits fund the data center buildout — creates a compounding advantage that neither Azure nor GCP can replicate at the same cost structure.
Q1 2026 confirmed each part of that thesis. The 28% growth removes the deceleration narrative. The 30%+ margin removes the margin-compression narrative. The $200B capex commitment — which the market initially punished as reckless — is now being repriced as a rational response to confirmed demand.
I am not adding here. The position is sized and the thesis is intact. The next decision point is Q2 guidance and whether Bedrock adoption metrics continue to accelerate. Until then, I hold.
— Ruslan Averin