Amazon's aggressive AI investments are spooking investors - but one analyst believes they're necessary for the company to fulfill accelerating cloud demand
Amazon's backlog-to-capex ratio suggests the company will be able to monetize its investments, according to an analyst.
With plans to devote $200 billion this year toward artificial-intelligence capital expenditures, Amazon.com could be the biggest spender of the Big Tech bunch.
Some investors are worried that the company is sinking billions into expensive data centers with little to show for it, but BNP Paribas analyst Nick Jones doesn't see it that way.
Amazon's (AMZN) heightened spending is "appropriate and necessary given demand levels and the size of the future opportunity," Jones wrote in a Tuesday note. He reiterated his outperform rating on the stock with a target price of $320, representing over 50% upside from current levels.
Shares of Amazon have fallen 8% since the beginning of the year.
Amazon and other hyperscalers such as Google are building data centers at a rate that's in line with the surge in contracted demand, or backlog, Jones pointed out. He also reiterated his outperform rating and $390 price target for shares of Alphabet.
More: Amazon hasn't been this cheap since 2008. Here's how the stock could rally 46% from here.
Even as Amazon's capital expenditures are expected to climb to levels that are over 100% of its cloud revenue, Jones highlighted that Amazon's backlog growth has accelerated in recent quarters.
"Concerns are overdone," the analyst said of elevated capex levels. Instead of focusing on Amazon's total spending, he said, investors should be tracking the company's backlog-to-capex ratio, which measures how contracted demand compares to infrastructure spending.
Jones estimates that every gigawatt of data-center capacity requires $50 billion to build, and generates $15 billion of revenue annually once online. With a robust backlog, he believes Amazon will be able to instantly convert available capacity into revenue.
Despite the steep initial investment, Jones thinks these estimates are likely conservative. Amazon can often obtain volume discounts on data-center construction and can deploy AI across its own core products to drive higher revenue per gigawatt than if the company were simply relying on its external cloud-services business alone.
Additionally, Amazon is seeing an upward trend in revenue per employee, a metric that Jones sees as a proxy for productivity gains. Companies are effectively replacing physical labor with digital labor as they scale AI and cloud workloads, he wrote.
Amazon's revenue per employee has climbed from under $300,000 in the first quarter of 2022 to over $540,000 in the fourth quarter of 2025, according to BNP Paribas estimates and company filings.
Comments