Amazon CEO Tries to Explain $200 Billion in AI Spending to Shareholders

In his annual letter to shareholders, Amazon’s CEO is already painting AI as a success for the company. He claims that AWS annualized revenues from AI just hit an annual run rate of $15 billion in a year in which the company is spending $200 billion on AI-related expenses. To be clear, this means that AWS AI revenues have exceeded one quarter of $15 billion one time, and during its best-ever quarter to date.

But let’s present that as a positive.

“Amazon is smack in the middle of this [AI] land rush, and companies are choosing AWS for AI,” Mr. Jassy writes. “Three years after AWS launched commercially [in 2002, 24 years ago], it had a $58 million revenue run rate. Three years into this AI wave, AWS’s AI revenue run rate is over $15 billion in Q1 2026 (nearly 260 times larger than AWS at that same point) and ascending rapidly.”

To be fair, Amazon is one of a handful of companies that might benefit financially from AI, given the popularity of its AWS infrastructure. But these numbers need to be put in context: In its best-ever quarter for AI revenues, AWS delivered approximately $3.75 billion (one-quarter of $15 billion) while Amazon’s AI spending hit about $50 billion (assuming the company divides its spending equally between the four quarters). That’s the delta between revenues–which are not profits, this part of the business is far from profitable—and spending, and that’s for one of the few companies in this industry that’s well-positioned to actually make a successful run at this market.

“We have never seen a technology more quickly adopted than AI,” he adds. “When ChatGPT launched in November 2022, it reached 100 million users in two months, four times faster than TikTok and 15 times faster than Instagram (ChatGPT already has over 900 million weekly active users). Both OpenAI and Anthropic have revenue run rates [meaning both have reported revenues, not profits, of about $7.5 billion one time each] reportedly approaching $30 billion. These are breathtaking numbers for companies this soon after their commercial launches.”

In Jassy’s words, Amazon sees a disproportionate inflection point in AI, and it is betting big.

“You want to invest as aggressively as you responsibly can,” he writes. “This will create investment spikes that will invite scrutiny, but the game-changers don’t typically accommodate smoother investment horizons.”

AI, Jassy says, is not over-hyped and that we are not in an AI bubble. Instead, the margins and return invested on capital expenditures (ROIC) are “appealing.” If anything, he says, AWS could be growing even faster if it wasnt for capacity restraints that prevent it from meeting all its customer needs. Two customers have reportedly asked him if they could buy all of Amazon’s Graviton GPU instance capacity for all of 2026.

“Our chips business is on fire,” he continues. “Virtually all AI thus far has been done on NVIDIA chips, but a new shift has started … Demand is booming.”

Jassy also explains why Amazon is spending so much on AI infrastructure: Customer commitments make that spending predictable. Amazon is not investing $200 billion on AI this year “on a hunch,” he writes, though I would argue that a “commitment” is just a promise and not booked revenues. “The recent OpenAI commitment (over $100 billion) is an example of this, but there are several other customer agreements completed (and unannounced), or deep in process. Of the AWS capex we expect to spend in 2026, much of which will be monetized in 2027-2028, we already have customer commitments for a substantial portion of it.”

In short, AI is a “once in a lifetime opportunity,” Jassy says, and the growth rate is unprecedented. Amazon, he says, will not be conservative here and will emerge with its leadership position in cloud/AI infrastructure intact as a result.