Google just bought a bunch of Spirit Airlines data for AI training

Spirit Airlines flew its last flight earlier this year, done in by soaring fuel costs during the Iran conflict and years of financial strain. Now its digital remains are being auctioned off piece by piece, and Google showed up with a checkbook.

The search giant placed a $10 million bid, according to Bloomberg Law, for a chunk of Spirit's enterprise dataset. Not the planes, not the routes, not the loyalty program. The data. Google confirmed the purchase in a short statement, saying the material could help improve its products and AI models, which is corporate speak for feeding the machine.

And there is a lot to feed it. Court filings describe hundreds of millions of Microsoft Teams messages and more than 100 million emails, alongside records tied to revenue, aircraft operations, employee productivity, audits and fraud investigations. Bloomberg's reporting adds even more texture: marketing campaigns, human resources files, strategic planning documents, project management records, pricing data pulled from over 7 billion competitor flights, roughly 7.5 billion passenger transaction records spanning nearly two decades, and more than 30 million lines of software code.

Google says the dataset will be scrubbed of personally identifiable information by a third party before it ever reaches the company, and that passenger profiles and loyalty details are excluded entirely. What remains is essentially the operational skeleton of a defunct airline, its emails, its code, its internal chatter, all of it potentially useful training material for large language models hungry for real-world business data.

The deal is a small footnote in Spirit's bankruptcy proceedings, but it says something bigger about where AI companies are now shopping for data. When a company folds, its assets used to mean planes, gates and slots. Increasingly, the most valuable line item might just be the archive.