Amazon’s AI Data Center Push Turns Engineers Into Critics

Amazon AI data center layoffs have moved from a workplace grievance into Seattle land-use politics, after Amazon engineers backed a city moratorium on large data centers while the company has announced about 30,000 corporate role reductions. Amazon has told investors in the 2025 shareholder letter that it is planning about $200 billion in capital expenditures this year, with Amazon Web Services (AWS, Amazon’s cloud computing arm) at the center of the artificial intelligence buildout.

Seattle’s draft response would pause applications for new or expanded data centers for a year while officials write rules for electrical load, water use, utility rates and land use. That turns Amazon’s internal headcount debate into a public infrastructure fight, with engineers now testifying on the side of city regulation.

Workers Took the AI Fight to City Hall

Patrick Schloesser, a software engineer at AWS, gave the hearing its plainest line. He spoke as a worker at the company building much of the compute capacity that the city is being asked to host, and he tied the layoff number to the speed of the AI buildout.

The leaders at my company have laid off 30,000 corporate employees in the last eight months. What that tells me is that Big Tech is desperate to build as much compute capacity as it can, as fast as it can.

Schloesser’s testimony did something a normal zoning hearing rarely does. It brought severance emails, cloud contracts and power substations into the same room. Other speakers focused on noise, heat, water and utility bills, but the employee comments gave the hearing a workplace edge that Amazon cannot answer with a technical note about efficient cooling.

Amazon told CNBC that it respects employees’ right to express opinions. The Seattle documents leave the companies behind the pending data center interest unnamed, so the testimony should not be read as proof that Amazon is one of the applicants. The political target is broader: the firms turning AI demand into a race for land, power and permits.

Amazon’s Ledger Points to AWS

Beth Galetti, Amazon’s senior vice president of People Experience and Technology, published the October corporate workforce memo that put the first reduction at approximately 14,000 roles. Her January organizational update added approximately 16,000 more. Those two memos account for about 30,000 corporate roles and describe the cuts as part of reducing layers, increasing ownership and removing bureaucracy.

The financial side is moving in the opposite direction. In the first quarter results, Amazon said net sales rose 17 percent to $181.5 billion, while AWS sales rose 28 percent to $37.6 billion. Free cash flow fell to $1.2 billion for the trailing 12 months, driven mainly by a $59.3 billion year over year increase in purchases of property and equipment, net of proceeds from sales and incentives. Amazon said the increase primarily reflects investments in artificial intelligence.

That combination explains the employee backlash better than a debate over AI ethics alone. Workers are watching a company shrink corporate teams while its cloud arm buys the physical capacity needed to sell AI services for years.

A Permit Pause With Teeth

Seattle’s data center moratorium bill, CB 121214, is still listed in committee, but the draft language is direct. It would stop the filing, acceptance, processing or approval of applications to establish or expand data centers in all zones across the city. It also creates a definition for a data center as a facility used primarily for computer and networking equipment with capacity above 20 megavolt-amperes (MVA, a measure of apparent electrical capacity).

• 365 days – The moratorium would expire after one year unless the council extends it under state law or ends it early.
• 20 MVA – The electrical capacity threshold separates large facilities from smaller server rooms and legacy sites.
• Three-quarter council vote – The bill is written as an emergency ordinance with an immediate effective date if it clears the required vote and mayoral approval.

The draft also orders the Seattle Department of Construction and Inspections to require electrical capacity information in land-use permit applications. That is a small administrative change with a large practical effect: a developer would have to put power demand into the permit file at the start, before public officials spend months debating a project with incomplete load data.

The findings section names the risks the city wants time to study: energy and water infrastructure, utility affordability and reliability, jobs and economic development, public health and the environment. The bill also notes that several smaller data centers already operate downtown.

Why Utility Bills Enter the Layoff Debate

Data centers turn a software boom into a utility problem. A model training run may look abstract to consumers, but the facility behind it needs power connections, backup systems, cooling and land. Seattle’s draft bill gives those physical demands a public process, and the worker testimony gives them a labor politics frame.

A separate Oton Technology report on AI data centers’ water use covered the same municipal tradeoff now facing Seattle: host cities receive construction activity and tax arguments, then they inherit questions about water, heat, power bills and grid upgrades. That is why local fights over data centers keep landing at utility boards and planning departments.

• Large-load rates: Seattle expects to consider separate electricity rates for new high-demand customers, including data centers.
• Permit intake: the draft bill would require applicants to disclose electrical capacity as part of a complete land-use application.
• Public hearing: state law requires a hearing within 60 days after adoption of the moratorium.
• Work plan: city departments would study grid reliability, water use, utility rates, land use, jobs, the economy and public health.

Those items explain why engineers showing up at City Hall is more than symbolic dissent. A worker inside the AI supply chain can describe the pace of the buildout. A city can decide whether its grid and ratepayers carry that pace.

Moratorium Bills Move Into Statehouses

Seattle is part of a larger political turn. The National Conference of State Legislatures, in the June data center moratorium tracker, listed lawmakers in 14 states considering bans or pauses on new facilities. The bills vary widely, with some still active, some failed and Maine’s over 20-megawatt moratorium vetoed.

• Fixed pauses: Georgia, Vermont and Oklahoma proposals set end dates that would stop new projects while agencies study impact.
• Study-linked pauses: Minnesota, New York and Pennsylvania versions tie new permits to reports on power, water, local costs or utility rules.
• Failed or vetoed efforts: several bills have already stalled, showing that data center skepticism has spread faster than final legislation.

The jobs side of the AI shift is also spreading. Oton Technology’s report on Meta’s AI restructuring in Nairobi tracked how data work and content-processing jobs can disappear when large platforms change contracts and automate more of the production chain. Amazon’s Seattle hearing adds a different version of the same labor problem, with engineers challenging the physical expansion behind the AI services they help build.

For tech companies, the statehouse map creates a planning problem. Power access, local consent and utility pricing are becoming part of the AI deployment schedule, alongside chip supply and model development.

Seattle Turns the Complaint Into Policy

Amazon has spent years making AI sound like an operating system for the next version of the company. Its filings and letters now put hard figures beside that story: cloud growth, property and equipment purchases, custom chips and long-term customer commitments. Employees hearing the layoff message from inside the company are taking those same figures to local officials.

Seattle gives that pushback a rare venue. A Slack thread can be muted. A public hearing becomes part of a legislative record. The city’s bill names electrical capacity and permit categories, leaving company names out, which lets the council regulate the shape of the buildout without proving which cloud provider wants which site.

That is why the Amazon engineers’ testimony carries beyond one hearing. It joins two questions that tech companies have kept in separate folders: how many people they need to run the AI business, and how much public infrastructure the AI business will consume. The draft requires a three-quarter council vote and mayoral approval before it can take effect immediately.