Google buys bankrupt Spirit Airlines' data for $10M to train AI

When an airline goes bankrupt, you usually expect rivals to fight over jets, gates, and landing slots. But in the generative AI era, the most prized scrap metal in bankruptcy court turns out to be spreadsheets and internal emails.

Following its full shutdown in May, discount carrier Spirit Airlines has agreed to sell its massive enterprise dataset to Google for $10 million. The data will be used to train Google's AI models, according to recent bankruptcy court filings.

Digging through a dead carrier's digital closet

Spirit is the first major U.S. airline in 25 years to halt operations entirely rather than get folded into a competitor. That rare total collapse left behind a peculiar asset: decades of real-world corporate operating logs.

Google's $10 million haul covers a staggering amount of everyday operations. The dataset includes internal emails and employee communications, internal spreadsheets, transaction logs, booking and frequent flyer histories, and even internal HR files. Court filings state that all personally identifiable details have been scrubbed, and Google confirmed to CNN that it will not receive personal info.

Google was not the only AI player circling the estate. AI-powered hiring startup Mercor.io put up the second-highest bid at $7.5 million, likely eyeing Spirit's vast HR and talent records. Bankruptcy Judge Sean Lane is expected to rule on the sale this Wednesday.

Why it matters

Frontier AI developers have largely exhausted the public web. To make agentic models that actually understand corporate workflows, pricing strategies, and team logistics, tech giants need private, uncurated enterprise exhaust. Spirit Airlines is essentially the first major test case where an entire company's corporate graveyard is sold directly into an AI training cluster.

We think this is an absurdly cost-effective move for Google, but one that raises serious privacy precedents. For $10 million, Google gets millions of authentic scheduling decisions, operational friction points, and customer service exchanges that synthetic data cannot replicate.

The catch? Most passengers who booked a cheap flight or job seekers who applied to Spirit never imagined their transactional footprint would end up feeding big tech's neural nets. Even with PII removed, treating corporate bankruptcy as an open-air data liquidation market creates a playbook that hospitals, retailers, and banks could soon follow.