Following Amazon's (AMZN 1.43%) fourth-quarter earnings report, the stock experienced a severe sell-off. Amazon stock now trades roughly 23% below its all-time highs at just 25.8 times this year's earnings estimates, which is close to its lowest valuation in the modern era on a price-to-earnings basis.
While revenue beat expectations in the fourth quarter, including a nice acceleration in the Amazon Web Services business, the company's massive $200 billion 2026 capital spending forecast sent investors running. After all, Amazon made just $139.5 billion in operating cash flow in 2025, up 17% from the prior year.
So, barring a massive tilt higher, it's likely Amazon may even post negative free cash flow in 2026. No wonder many conservative investors fled the scene.
However, in a recent interview, AWS CEO Matt Garman explained why investors shouldn't worry. In fact, they should probably be greedy on the news of this turbocharged spending.
Image source: Getty Images.
NASDAQ: AMZN
Key Data Points
Garman predicts undersupply for "years"
In a recent interview with CNBC's John Fortt, Garman said:
Even with all this investment, my best estimation is that we will be capacity-constrained for the next couple years. We will sell every single server and every single bit, and we'll wish that we had more. And that is where the state of the world will be for at least the next couple of years...
If Garman is even close to correct, then investors shouldn't fear the $200 billion in spending. In fact, they should want Amazon to invest even more.
After all, many questioned Amazon's past strategy of perpetual investment in the growth of its e-commerce distribution and fulfillment network, which ate up all of Amazon's profitability in its early days. However, over the long run, that enormous physical footprint cemented Amazon's moat as the undisputed leader in U.S. e-commerce. Today, Amazon's e-commerce division is profitable, reporting $35 billion in operating income last year, up 23% year over year.
The same was also said when Amazon began investing in AWS data centers, and now that business is even larger than the e-commerce business in terms of profit, with $129 billion in revenue last year, up 18%, along with $45 billion in operating income.
If Garman is correct and believes that Amazon will still be undersupplied after spending $200 billion largely on AI computing power, then Amazon should be able to charge an adequate price for that compute.
Not only is all that spending not a bad thing, but it's actually the sign of the best type of business to own, according to Warren Buffett. In his 1992 letter to shareholders, Buffett wrote:
Leaving the question of price aside, the best business to own is one that over an extended period can employ large amounts of incremental capital at very high rates of return. The worst business to own is one that must, or will, do the opposite-that is, consistently employ ever-greater amounts of capital at very low rates of return.
While we aren't sure whether Amazon is 100% guaranteed to earn high returns, the fact that Garman said Amazon still expects to be undersupplied for years means it will likely be able to charge adequate prices for all that compute and generate appropriate returns. After all, AWS this year garnered a 35.4% operating margin in 2025, which is a healthy margin indeed. So, the odds are that Amazon is going to generate high returns on that spending.
A more diversified customer base
The other point Garman made in the interview was that those future revenues and profits are a much safer bet than those of peers, because Amazon's demand is diversified among a lot of different types of companies. Amazon is, after all, the first mover in cloud computing and remains the largest platform today.
While other cloud companies have reported massive backlogs that rival Amazon's, a large portion of some companies' backlogs is concentrated in just one customer: OpenAI. That's why Microsoft (MSFT 0.36%) and especially Oracle (ORCL 2.00%) didn't get much credit for their massive backlog increases disclosed over the past six months; future contracts with OpenAI accounted for a large percentage of that total. Given that investors are now questioning whether OpenAI can fund its massive future commitments, vendors have faced investor skepticism.
And while Amazon also serves OpenAI, OpenAI's total future commitment to AWS is only $38 billion, which is a relatively small portion of Amazon's $244 billion backlog.
In the same CNBC interview, Garman elaborated that Amazon's customer diversity lowers the risk of this new spending:
And so others have some growth, very concentrated in some of these very large customers. And we have many of them, the OpenAIs, the Anthropic of the world building on us as well. But we're also, I think, for the long-term health of the business, very focused on how do we get every single start-up out there building on top of AWS? How do we get every bank? How do we get every healthcare company? How do we get every retail company? How do we get every manufacturing company building on top of AWS in the cloud? And so, for that, and there, we see AWS winning the vast majority of those times.
Look for Amazon's growth to accelerate
In the fourth quarter, Amazon Web Services' year-over-year revenue growth rate accelerated from 20% in Q3 to 24% in Q4. Given its massive investments scheduled for 2026, which Garman noted will yield results in 2027 and even into 2028, AWS should see a requisite acceleration in revenue and profits.
While this year's spending may cloud the near-term performance of Amazon stock, long-term investors shouldn't hesitate to buy with both hands. The bet is likely to pay off.
