Amazon announced Thursday that it plans to spend a staggering $200 billion in capital spending this year, raising concerns that the AI spending boom is entering a more active phase.
The company’s capital spending forecast exceeded Wall Street expectations by more than $50 billion. This leaves some analysts unsure whether these huge investments will pay off.
“Strong long-term returns on invested capital — I think that’s what the market is talking about today,” Evercore ISI technology analyst Mark Mahaney said on a post-earnings conference call with Amazon executives. “So, could you tell us a little bit more about how you think investors can understand that?”
Amazon CEO Andy Jassy said new AI capabilities are being monetized as soon as they are introduced, calling it a “very unusual opportunity” and arguing that AI deployments are accelerating customers’ migration to the cloud. He added that AWS’s experience with demand forecasting helps limit wasted capacity.
“This isn’t some kind of weird top-line acquisition,” Jassy said during the call. “We are confident that these investments will provide a strong return on invested capital.”
Mr. Jassy’s comments did little to calm investors. Amazon shares fell more than 10% in after-hours trading after the company detailed its massive spending plans. Earnings forecasts were also lower than Wall Street expectations, raising concerns about ROI.
Big Tech’s AI infrastructure build-out has entered a new and surprising phase in recent days. Google said Wednesday that it expects capital spending in 2026 to be between $175 billion and $185 billion, significantly higher than Wall Street expectations. Meta and Microsoft also raised capital investment.
During Amazon’s earnings call on Thursday, other analysts sought assurances about the return on investment. JPMorgan’s Doug Anmas asked Mr. Jassy whether the company had “financial guardrails” in place for its spending plans.
Jassy said AI represents a “very unusual opportunity” that could reshape the scale of AWS and Amazon. He said the company’s in-house chips, Trainium and Graviton, are on track to generate more than $10 billion in revenue this year, adding that the partnership with Anthropic is “going very well.”
“We see this as a unique opportunity and, as we have done in recent years, we intend to invest aggressively to become a leader,” Jassy said.
Jassy describes the AI market as a “barbell,” with heavy spending by major AI labs on the one hand and productivity-focused enterprise applications on the other.
He said the new businesses in between – enterprise production workloads and AI-natives – are still in their infancy, but could become the “largest and most durable” source of demand as costs come down and adoption grows.
“Most of that demand hasn’t come to the middle of the barbell yet, but it will emerge over time,” he said.
Any tips? To contact this reporter via email, please specify the following address: ekim@businessinsider.com or on Signal, Telegram or WhatsApp 650-942-3061. Use a personal email address, non-work WiFi network, and non-work device. Here’s a guide to sharing your information securely.
