Amazon’s AI Investments Are Creating a Whole New Business Model
Amazon is pouring hundreds of billions into AI infrastructure while free cash flow turns negative, yet something unexpected is showing up in its retail data that suggests this spending may be doing far more than powering cloud servers.
The artificial intelligence spending spree has reached a point where investors are right to demand more than promises. Amazon (NASDAQ:AMZN | AMZN Price Prediction), Microsoft (NASDAQ:MSFT), Alphabet (NASDAQ:GOOG), and Meta Platforms (NASDAQ:META) are committing hundreds of billions of dollars to chips, data centers, networking equipment, and power, creating a massive question around return on investment. The concern is straightforward: Can all that infrastructure eventually generate enough revenue and profit to justify the capital being deployed today?
Amazon is beginning to provide an unusually compelling answer. Its AI spending isn’t merely supporting an existing business. It may be creating entirely new sources of demand.
Amazon’s AI Spending Keeps Getting Bigger
Amazon spent $131 billion on capital expenditures in 2025, up from $83 billion in 2024, and initially expected roughly $200 billion of capital expenditures in 2026. Then it raised that figure to approximately $220 billion following its second-quarter results. It could spend as much as $628 billion by 2028.
That spending is already showing up in the cash-flow statement. Amazon generated $161.4 billion in operating cash flow over the 12 months ended June 30, but free cash flow was negative $7.6 billion after capital expenditures.
Granted, that is a legitimate concern. Investors don’t get to spend $220 billion without eventually demanding a return. But Amazon is also generating the revenue growth needed to make the investment case.
The AI Infrastructure Is Producing Revenue
Second-quarter AWS revenue jumped 37% year-over-year to $42.2 billion, its fastest growth in 18 quarters, while AWS operating income increased to $16.6 billion from $10.2 billion. Amazon also said its AI and chips businesses each surpassed a $25 billion annual revenue run rate. That’s important because Amazon isn’t simply building data centers and hoping customers arrive later.
CEO Andy Jassy has said much of the company’s 2026 AWS capital spending is already backed by customer commitments, with much of that capacity expected to be monetized in 2027 and 2028.
The infrastructure is therefore becoming a revenue-producing asset rather than just an expense line. And now something even more interesting is happening inside Amazon’s retail operation.
Agentic AI Could Change the Retail Math
Evercore ISI just raised its Amazon price target to $355 from $315.16 after its 14th Annual U.S. Online Retail Survey produced what it called the first survey evidence that agentic AI is additive to Amazon’s retail business.
The key figure is remarkable: 57% of Alexa AI users said they purchased a product they previously didn’t know about.
That’s different from AI simply making Amazon’s existing search engine better. Traditional e-commerce largely captures existing purchase intent. A shopper wants something, searches for it, compares options, and buys. Agentic AI can intervene earlier in that process by understanding what a customer needs and introducing products the customer wasn’t actively looking for. That creates new purchase intent.
If 57% of Alexa AI users are discovering products they didn’t previously know existed and then purchasing them, Amazon isn’t merely converting demand more efficiently. It is potentially creating incremental demand.
That’s not just another feature — it’s a new business model.
Key Takeaway
In short, investors shouldn’t ignore Amazon’s negative free cash flow. Spending $220 billion in a single year creates real execution and return-on-capital risks.
But the evidence is moving in the right direction. AWS is growing 37%, its AI business has surpassed a $25 billion annual run rate, and agentic AI is beginning to influence what customers buy.
Ultimately, the biggest payoff from Amazon’s AI investment may not come from selling computing power. Instead, it may come from using that computing power to make customers buy things they never intended to purchase.
Contact [email protected] for any questions or corrections.
After two decades of patrolling the dark corners of suburbia as a police officer, Rich Duprey hung up his badge and gun to begin writing full time about stocks and investing. For the past 20 years he’s been cruising the markets looking for companies to lock up as long-term holdings in a portfolio while writing extensively on the broad sectors of consumer goods, technology, and industrials. Because his experience isn’t from the typical financial analyst track, Rich is able to break down complex topics into understandable and useful action points for the average investor. His writings have appeared on The Motley Fool, InvestorPlace, Yahoo! Finance, and Money Morning. He has been featured in both U.S. and international publications, including MarketWatch, Financial Times, Forbes, Fast Company, and USA Today.
