Amazon CEO Andy Jassy says AWS can become the first business ever to reach a trillion dollars in annual revenue. He also told investors it will cost a lot more to get there. According to TechCrunch AI, Jassy repeated that goal on Amazon’s quarterly earnings call on Thursday. He confirmed roughly $220 billion in cash capex for 2026 and said Amazon will spend “meaningfully more” in 2027.
Wall Street didn’t like the price. Shares fell about 6% in after-hours trading.
📌 The Key Takeaways
- The goal: AWS becomes a trillion-dollar revenue business “in the next five to 10 years.”
- Where it stands now: a $184 billion annualized revenue run rate, up 39% from the same quarter last year.
- The forecast: growth “north of 40 percent” in both 2027 and 2028.
- The cost: about $220 billion in 2026 capex, most of it for AWS, and more in 2027.
- How the market took it: analysts kept pushing on margins and returns, and the stock dropped about 6%.
💰 Why the Spending Keeps Climbing
AI workloads run on physical things: data centers, chips, power and networking. That means cloud providers have to pay for capacity before the revenue shows up. Jassy said so directly, calling the AI boom a “once in a lifetime opportunity” that justifies the size of the commitment.
“We’re going to invest to continue being the leader,” he said. “We can fund this plan, and I think it will forever change the size of Amazon. It will forever change our revenue.”
The confidence matters. A company spending at this level usually hedges its guidance. Jassy did the opposite. He made a multi-year growth call and said Amazon is still “in the very early days” of enterprise AI adoption.
📈 What Makes the Growth Numbers Stand Out
The 39% growth rate is the most impressive part of the report. AWS is already huge, and companies this big rarely speed up. This is also the sixth straight quarter of steady growth, which gives Jassy’s 40%-plus forecast some support.
For scale, look at what the trillion-dollar target would take. Going from $184 billion to $1 trillion means growing AWS more than fivefold. Even at 40% a year, that’s a five-year run with almost no slowdown. It’s possible, but there’s very little room for things to go wrong.
⚖️ The Investor Problem
All year, investors have worried about how much Amazon is spending on AI infrastructure. TechCrunch AI reports that analysts on the call kept asking Jassy the same three questions:
- Why does the buildout need to be this big?
- What signals show the demand is real and will last?
- What profit margins should they expect at the end of it?
Jassy’s answer was that “the strategy is working.” The stock drop shows not everyone agrees yet. The worry isn’t about demand today. It’s about timing: if AI demand cools before the capacity pays off, Amazon is left with very expensive data centers it doesn’t need.
🌐 The Bigger Picture
Amazon isn’t the only one spending like this. Microsoft just reported another record quarter, driven mostly by its cloud business, and it’s spending heavily too. Every hyperscaler (the giant cloud providers like AWS, Azure and Google Cloud) is in the same spot. They’re making a lot of money from AI and spending even more to keep up with it.
What stands out is that none of them seem willing to slow down. Whoever runs short on capacity loses enterprise customers to whoever doesn’t. That’s why Jassy treats the capex as a cost of staying in the lead, not an optional bet.
🔭 What Practitioners Should Watch
- Capacity: more AWS capacity should mean fewer GPU shortages and easier access to AI compute over the next two years.
- Pricing pressure: heavy spending usually comes with pressure to earn it back. Watch how AWS prices its AI services.
- Enterprise lock-in: Jassy is betting companies will run both their “core and AI workloads” on AWS. Expect more bundled offers that make moving your AI stack elsewhere harder.
- 2027 guidance: the actual 2027 capex number will tell you how sure Amazon really is.
Jassy has set a high bar: 40% growth for two more years and a business bigger than any before it. Over the next few quarters we’ll see whether the spending turns into that revenue. The full earnings call coverage is available at TechCrunch AI.