Amazon engineers blast $200 billion AI buildout after 30,000 layoffs

Amazon’s own engineers just turned the company’s AI spending story into a public referendum. At a Seattle Land Use and Sustainability Committee hearing on Wednesday, software engineer Patrick Schloesser said Amazon is spending $200 billion on capital this year, with most of it going to data centers and AI, while “the leaders at my company have laid off 30,000 corporate employees in the last eight months.” His conclusion was blunt: “What that tells me is that Big Tech is desperate to build as much compute capacity as it can, as fast as it can.”

That is the core tension here, and it is bigger than one company. Amazon is not just spending heavily on infrastructure, it is doing so while trimming headcount and facing a growing public backlash over where, how, and whether those facilities should be built. Schloesser was one of three Amazon employees who spoke in favor of tighter regulation of local data center development. The hearing helped push Seattle City Council to approve a yearlong moratorium on local data center construction after public outcry over proposals for five large-scale complexes around the city. In other words, this is no longer an abstract debate about cloud computing. It is now a zoning fight, a labor story, and a political problem for Big Tech all at once.

Amazon says it has no plans to build data centers within Seattle city limits and says it remains committed to investing in local economic development and improving water and energy efficiency in its projects. Company spokesperson Margaret Callahan said, “We respect our colleagues' right to voice their opinions,” adding that Amazon “engage[s] regularly with community stakeholders to understand local priorities and address concerns transparently, supporting both technological innovation and the specific needs of each region where we operate.” The company’s statement matters because this kind of pushback does not stay inside the building. Once employees, residents, and city officials are on the same stage, capital plans become reputation plans, and reputation plans become regulatory plans.

The scale of the AI infrastructure race helps explain why this is becoming such a flash point. Hyperscalers like Amazon, Alphabet, Meta, and Microsoft have poured $700 billion into AI infrastructure this year alone, part of a wider spending blitz that is expected to reach $7 trillion by 2030. Amazon reiterated in April that it plans to spend $200 billion in AI capital expenditures for the rest of this year. Microsoft is spending $190 billion. The numbers are enormous because the prize is enormous: more compute means more room to train models, run AI products, and keep up in a race where falling behind can look existential. But the source material also shows the other side of the ledger. As data center spending balloons, tech companies have cut costs elsewhere, including in their workforces.

Amazon’s layoffs were tied by the company to a need to decrease bureaucracy and increase efficiency. Meta dismissed 10% of its staff last month after announcing earlier this year that it would double its AI capex of $72 billion from 2025. Oracle’s staff reduction, estimated at anywhere from 20,000 to 30,000 employees this spring, coincided with the company’s disclosure of $248 billion in future data center lease obligations. That sequencing is hard for workers, communities, and lawmakers to ignore. When companies say they need fewer people and more servers at the same time, they are making a bet about where value will be created next. They are also making a bet that local governments will tolerate the tradeoffs.

So far, tolerance is wearing thin. Environmental concerns around data center construction, including noise pollution and excessive water usage, have dented public support for AI infrastructure expansion. A recent Gallup poll found 70% of Americans opposed data center construction in their local areas, including nearly half who are strongly opposed. Legislators in 14 states are now considering bans on data center development, and a dozen are weighing moratoriums. Monterey Park, California, a city of 60,000 people just outside Los Angeles, voted overwhelmingly for a permanent data center ban on Tuesday, with 86% of votes supporting the restriction. Still, opposition does not always stop a project. In Saline Township, Michigan, a $16 billion data center tied to OpenAI and Oracle’s Stargate AI infrastructure initiative moved forward even after the town board and planning commission rejected the plans. The developer sued, the town settled, and construction was ultimately able to proceed.

Inside Amazon, the pushback is getting more specific. Senior software engineer Liesl Wigand argued at the Seattle hearing that local entities should control the terms of expansion: “Local governments, in collaboration with community stakeholders, should be setting the terms for data center buildout,” she said. “Let’s not let Big Tech burn Seattle to win the AI race.” Schloesser went further, arguing that tech companies should offset the energy they consume with renewable energy to support the U.S. grid and should provide “good jobs building these things” while paying “a new tax that funds city jobs every time you conduct a large layoff.” Amazon Employees for Climate Justice, the internal advocacy group that includes Wigand and Schloesser, has already pressed the company on these issues. In an open letter published in Nov. 2025, the group called on Amazon to power all data centers with 100% local renewable resources and increase AI working groups with participation from non-managers. It warned that “the all-costs-justified, warp-speed approach to AI development will do staggering damage to democracy, to our jobs, and to the earth.” For executives, that is the takeaway: the AI buildout is no longer just a capital allocation decision. It is now a question of whether your company can keep scaling fast enough to compete without triggering enough local, employee, and political resistance to slow the machine down.