Amazon shares sank 11% in extended trading on Thursday after the company laid out plans to spend as much as $200 billion on capital expenditures, a figure that has sharpened Wall Street’s unease about how far, and how fast, Big Tech is willing to go in the race to dominate artificial intelligence infrastructure.
The scale of the forecast left investors stunned. Amazon spent about $131 billion on property and equipment in 2025, already a steep increase from roughly $83 billion the year before. The new outlook implies another dramatic step up, placing Amazon well ahead of its megacap peers and more than $50 billion above what analysts had penciled in. In a market increasingly sensitive to cash flow discipline, the reaction was swift.
The selloff comes at a moment when the AI narrative is shifting. Since OpenAI’s release of ChatGPT in late 2022, technology companies have justified rising spending by pointing to explosive demand for compute. Entering 2026, that logic is facing tougher scrutiny as commitments grow larger and timelines for returns remain uncertain.
Alphabet said this week it could spend up to $185 billion next year, while Meta has guided for capital expenditures of as much as $135 billion. Rather than peaking, the AI investment cycle appears to be accelerating.
On Amazon’s earnings call, analysts pressed management to explain when shareholders might begin to see tangible payback. CEO Andy Jassy said he was “confident” that the investments would deliver strong returns on invested capital, particularly through Amazon Web Services, but he declined to offer specific milestones.
That lack of precision was central to investor concerns. Evercore ISI analyst Mark Mahaney urged Jassy to bridge the gap between conviction and visibility, asking how the company gets from today’s spending surge to sustained long-term returns. Jassy responded by framing the investment as demand-driven rather than speculative.
“This isn’t some sort of quixotic, top-line grab,” he said, arguing that Amazon is responding to concrete customer needs. According to Jassy, demand for AI compute on AWS is so strong that growth is being constrained by capacity rather than interest.
The numbers from AWS lend some support to that claim. The cloud unit reported revenue growth of 24% to $35.6 billion in the most recent quarter, beating expectations and marking its fastest pace of expansion in 13 quarters. Jassy said AWS could have grown faster if it had more infrastructure in place, a point he used to justify the aggressive buildout.
To close that gap, Amazon added nearly 4 gigawatts of computing capacity in 2025 and expects to double its available power by the end of 2027. That expansion requires massive upfront investment in data centers, networking equipment, and custom chips, locking Amazon into a capital-heavy path that investors worry could weigh on margins if demand cools or pricing weakens.
Beyond near-term financials, analysts are also questioning how the structure of the AI market will evolve. Barclays analyst Ross Sandler asked whether spending remains concentrated among a handful of AI-native labs or whether enterprise adoption is broad enough to support sustained returns on infrastructure.
Jassy described the market as increasingly polarized. On one side are large AI labs consuming enormous amounts of compute. On the other hand, enterprises are adopting AI as a tool for productivity gains and cost control. Between them is a broad middle of companies experimenting, piloting, and gradually scaling applications.
“That middle part of the barbell very well may end up being the largest and most durable,” Jassy said, suggesting that enterprise demand, rather than hype around foundation models, will underpin long-term growth.
Still, the market reaction suggests investors are no longer willing to take that outcome on faith. Amazon’s stock drop underscores a broader shift in sentiment: enthusiasm for AI remains, but tolerance for open-ended spending is narrowing. With interest rates still elevated and competition intensifying across cloud and AI services, investors are increasingly focused on execution, efficiency, and timing.
The challenge for Amazon now is to convince the market that its AI ambitions will follow the same arc as AWS did in its early years, when heavy investment eventually produced one of the most profitable businesses in tech.
