Amazon.com Inc. (NASDAQ: AMZN) has spent the past year pouring money into the future at a pace few companies could stomach. The tech giant now heads into its Q3 earnings report later this month with the stock near $260, up around 5% since the end of September, and with Wall Street’s patience—and its own wallet—feeling pretty stretched.
With a capital-spending spree that’s on track to top $220 billion this year, investors want proof the money is turning into profit, not just bigger data centers and longer backlogs. The question hanging over the report is easy to ask and hard to answer: Will all of Amazon’s AI investments be justified in the stock’s price when the numbers finally land?
AWS Is Doing the Heavy Lifting
A major part of the bull case going into this month's earnings centers on Amazon Web Services (AWS). The cloud division has been growing north of 35%, operating income is climbing faster still, and the company’s margins have expanded as well. For a business of Amazon’s size, that kind of growth is almost unmatched, and it’s the engine that makes everything else possible.
Further support for the bullish case comes from the fact that the company’s AI demand is looking more real than theoretical. Amazon is reportedly sitting on a cloud backlog approaching half a trillion dollars, underpinned by commitments to its custom Trainium chips and swelling usage from the likes of Anthropic and OpenAI.
The custom chip angle matters more than you might think. By leaning on Trainium, Amazon can strategically lower its reliance on NVIDIA Corp. (NASDAQ: NVDA), saving money and reducing the risk of bottlenecks.
A $220 Billion Bet That Has to Pay Off
However, this is where the caution creeps in. That $220 billion spending plan is putting substantial pressure on free cash flow, even as Amazon's revenue and operating income continue to grow. The question is how quickly those investments will begin generating sufficient returns. A backlog, after all, is just a promise of future revenue, not money in the bank.
The near-term worry is margins. As Amazon’s new data centers come online, depreciation, energy bills, and interest on debt-funded spending could all weigh on profitability well before the matching revenue ramps. AI hardware also ages quickly, raising the risk that today’s expensive kit needs replacing before it has earned its keep.
Then there is customer concentration risk. Amazon has secured major cloud commitments from Anthropic and OpenAI, but it has not disclosed how much of its overall AWS backlog comes from either company. While those agreements support future demand, heavy reliance on a handful of major AI developers could introduce risks if their spending slows.
Why the Decart Talks Matter
Into this already expensive picture comes news that Amazon is in talks to buy AI startup Decart for around $7 billion, far above the roughly $4 billion valuation it carried in the spring. The news broke on Sunday, Oct. 11, and the deal isn’t finalized yet, but it says a lot about how aggressively Amazon plans to expand its AI stack.
The potential strategic benefits are easy to see. Decart’s Lucy model edits live video in real time. At the same time, its Oasis system generates simulated worlds for training robots and self-driving vehicles, the sort of technology that could feed Twitch, Prime Video, advertising, and warehouse automation. Amazon is reportedly already Decart’s largest customer, so this is less a bet on a stranger than a move to own what it already relies on.
The price, though, is hard to wave away. Paying such a premium could make sense if Amazon can scale Decart's technology and talent across AWS, rather than simply betting on AI models that may never become major platforms. It’s worth noting that Anthropic reportedly looked at a deal of its own with Decart before walking away, and that was when the startup carried a $6 billion price tag.
What the Report Has to Prove
Still, for all the hand-wringing over spending, the broad strokes of Amazon’s strategy are hanging together pretty nicely right now: build the infrastructure, lock in the orders, design the chips, back the leading model makers, and add software that makes the whole machine run harder. Get it right, and today’s capital intensity looks like an investment phase, rather than a permanent drag.
That is why the bar sits so high this month. Sure, Tigress Financial just reiterated its Buy rating on the stock with a $385 price target, pointing to nearly 50% upside, and the stock carries a MarketBeat consensus rating of Moderate Buy. But Amazon’s Q3 will need to show that its strong revenue and operating income growth can continue, while providing investors with greater confidence that its massive AI investments will eventually translate into stronger free cash flow.
In other words, the bull case is still very much intact, and for long-term investors, there’s a lot to like about the AI build-out. Despite the stock's gains over the past year, investors may need more evidence that Amazon's massive AI investments are paying off before shares can establish a sustained move to new highs.