Who's liable when a $3.2B AI data center catches fire?

Share
Who's liable when a $3.2B AI data center catches fire?

Two stories this afternoon about the gap between what AI companies promise and what anyone can actually hold them to — one measured in megawatts and dry hydrants, one in a president's AI-generated feed.


A fire at a half-built AI data center in western New York exposed how thin the accountability really is. In early June, crews responding to a blaze in an unfinished building at the Lake Mariner campus in Somerset found no working alarm, no suppression system, and three dead hydrants — and the chemical safety sheets they're legally entitled to see had reportedly burned in the fire. Barker Fire Department chief Steve Matisz told reporters his crew went in "kind of blind." Two months later, Ars Technica reports, the water problems haven't been fixed. Construction hasn't slowed.

What makes Lake Mariner a template rather than an anecdote is the ownership stack. The $3.2 billion, 500-megawatt campus sits on land TeraWulf leases from a company owned by TeraWulf's own CEO; AI cloud firm Fluidstack will run it; Google holds warrants for a future 14 percent equity stake and backstops Fluidstack's lease payments; Anthropic's compute demand is what the thing exists to serve. TeraWulf says it owns operational safety. But nobody in that chain appears able to verify the site's own clean-energy claims, and a source familiar with Anthropic's thinking said the company had no details on what it would need from TeraWulf or Fluidstack to check — or whether it had ever asked.

The promises keep slipping. TeraWulf's 2024 CSR report called Lake Mariner "95 percent zero-carbon"; its 2026 sustainability policy now says "low- and zero-carbon," and the number is 91 percent — but the company says that reflects the regional grid mix, not any power contract, and it buys no renewable certificates at all. The 2019 jobs pitch was 165 permanent positions and $85 million in capital investment; the 2024 planning board hearing put full buildout at 35 to 40 jobs. Governor Kathy Hochul has said the quiet part out loud: data centers "do not deliver significant, long-term jobs." Meanwhile the legislature's Responsible Data Center Development Act has sat unsigned for three months, and even Hochul's own moratorium doesn't reach Lake Mariner — it was permitted before the order took effect.

Our read: this is what "diffuse responsibility" looks like at scale. Anthropic's ratepayer pledge is genuinely good policy — it pays for grid cost increases — but it covers prices, not hydrants, noise, or whether the electrons are clean. And when a facility's risk is spread across a landlord, an operator, a warrant holder, and a tenant, the local fire chief is the one holding the bag. If there's no hydrant, as Matisz put it, that fire is going to burn under their asses.


President Trump spent a Sunday posting dozens of AI-generated images to Truth Social — including one where he claims to have made "hundreds of billions" on stocks for the U.S. CNBC counted more than 10 hours of posting: fabricated scenes of Iran's destruction, Trump on horseback with George Washington, Robert De Niro in a "Trump is my president" shirt, and maps relabeling the Strait of Hormuz, the Moon, and Lake Ontario as American territory. The stock-gains claim came with no evidence, and the White House didn't respond to a request for comment.

It's tempting to file this as pure spectacle, but the context is the part that matters. Trump Media already sells "Truth API," a $100,000-a-month product that gives subscribers early access to announcements that can move markets; senators asked the SEC in July to investigate whether that violates securities law. A president who posts unverified market claims alongside synthetic images, on a platform his company monetizes for speed, is a real disclosure problem — not a memes problem.


What to watch: whether Hochul signs the Responsible Data Center Development Act before the legislature reconvenes, and whether anyone in Congress picks the Truth API question back up.

Should a data center's corporate backers be legally on the hook for the local infrastructure their buildout breaks? Tell us in the comments.

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t