Why Amazon and Microsoft's stocks could be in trouble due to AI's destructive economics

By Christine Ji

While demand for AI computing remains high, the economics are deteriorating for Amazon and Microsoft, which now must spend six times more capital to generate the same value as before, according to one analyst

Amazon's cloud asset base has grown 3.5 times since 2021, while operating profit has only risen 2.4 times, creating a drag on returns that Rothschild says is the lowest since 2015.

Investors have been willing to overlook Amazon.com Inc.'s and Microsoft Corp.'s hefty spending on artificial-intelligence infrastructure in hopes of a massive future payoff. But it's not so clear how much value these AI initiatives will add, and Wall Street could soon lose patience.

In a Tuesday report, Rothschild & Co Redburn analyst Alex Haissl downgraded shares of Amazon (AMZN) and Microsoft (MSFT) to neutral from buy, arguing that the return on investment for AI will be nothing like the initial "Cloud 1.0" shift, when companies began migrating their legacy IT operations to the cloud.

Haissl estimates that AI infrastructure requires six times the amount of capital to generate the same economic value as the original cloud shift, which could easily make AI projects "value-destructive." For every $1 a company currently spends on AI infrastructure such as graphics processing units, he said, the estimated net present value is just 20 cents. By comparison, mature traditional cloud investments generate roughly $1.40 for every dollar spent, according to Haissl.

Additionally, hyperscalers have lower pricing power and a higher risk of overbuilding today versus Cloud 1.0.

Investors may still believe that AI projects will improve their returns on investment over time, as shares of Microsoft and Amazon are still trading at similar levels to where they were during the less capital-intensive Cloud 1.0 era. But Haissl sees limited upside for these stocks going forward. He lowered his Microsoft price target to $500 from $560 and reaffirmed his $250 price target for Amazon.

Read: The bullish case for Amazon's stock builds even further thanks to these two factors

Expensive AI workloads are responsible for over 50% of growth at Amazon's AWS and Microsoft's Azure, according to Haissl, meaning that these companies' cloud business models are shifting to lower returns and heavier depreciation burdens.

In the case of Microsoft, Haissl estimates that the company generated $17 billion in revenue per gigawatt before its Azure business turned toward AI workloads. Today, that number is around $11 billion. Incorporating third-party AI model providers such as OpenAI also creates "value leakage" for Microsoft, creating a new cost that didn't exist when it owned the entire software stack.

Haissl said AWS is better-positioned than Azure because Amazon uses its custom Trainium chips to run AI workloads, leading to a higher internal rate of return as it avoids paying third-party chip suppliers. However, AI remains "dilutive to returns" at AWS, as Amazon's capital expenditures have led to a lower return on investment than anticipated.

According to Haissl, AWS' asset base increased 3.5x - from $64 billion at the end of 2021 to $224 billion at the end of the most recent quarter - but operating profit only rose 2.4x. That marked the lowest level of returns since 2015.

Although Amazon impressed investors last quarter with AWS growth surpassing 20%, Haissl argued that positive catalysts have become fully priced in, leaving the stock with limited upside.

Additionally, analysts have limited visibility into the timing of AI capacity coming online, making it difficult to determine the level of AI monetization. In the most recent October earnings report, Amazon reported adding 3.8 gigawatts of capacity over the last 12 months. Assuming that each gigawatt brings in $10 billion of revenue, that implies roughly $38 billion in potential annual revenue, Haissl noted. However, Amazon reported $4.6 billion in annualized revenue growth, leaving Haissl skeptical about when the rest of the monetization will occur.

For shares of both Amazon and Microsoft to turn around, the companies would need to show "sustained high growth paired with a meaningful reduction in capex," Haissl wrote.

Read on: Big Tech needs a staggering $1.5 trillion to fund the AI boom. This is the complex playbook it's using to get it.

-Christine Ji

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

11-18-25 1115ET