Google is acquiring data from a bankrupt airline, and it has a good reason
Google has won a tender to acquire the internal data of the American airline Spirit for $10 million. The tech giant intends to use this massive dataset to train its AI models.
When an airline collapses and goes bankrupt, we are used to seeing a fire sale of all its assets like planes, spare parts, office equipment, and maybe some used duty-free carts. But we are in 2026, and such companies have other assets that are no less important and no less expensive.
10 million dollars and all the data is yours
From documents filed in a court in the United States, it appears that Google won a recently closed tender and will acquire all the data and internal information of the American airline Spirit for 10 million dollars, after outbidding offers from competitors like the startup Mercor, which specializes in selling data for the purpose of training models.
Spirit was an American low-cost airline that ran into difficulties following the COVID-19 crisis in 2020 and failed to recover from huge losses, mounting debts, and fuel costs. In May 2026, it was permanently grounded, entered bankruptcy proceedings, and began selling its assets to cover debts, and this is where Google enters the picture.
So what does Google have to do with the servers and data of a failed airline? The answer is, of course, high-quality data, which Google urgently needs for training its models.
While the web is flooded with AI Slop and generic texts (not least because of Google itself), tech giants and AI companies are hungry today for high-quality and deep data that shows real processes of entire organizations, and especially how people really work. Spirit's massive database is a dream for anyone who needs to train models on organizational information, and it includes about 100 million emails, half a billion Microsoft Teams messages, and almost 37 million files on Drive and SharePoint. Add to that 600 thousand tickets that were opened and documented, 30 million lines of proprietary code, 370 thousand commits, and operational data on hundreds of thousands of flights, and you have a lot of high-quality organizational data.
What about passenger privacy?
The interesting, or disturbing, part, depending on who you ask, is the privacy issue, of course. After all, the employees who feature in this data did not agree to the sale. Google reassures that the information will undergo full anonymization before it reaches their servers. But if you look at the fine print of the deal, you see that Google is the one choosing the company that will perform the data cleaning, it is the one that will characterize exactly what to clean and keep, it pays for the service, and it is the one that approves the final product. In addition, the contract requires maintaining "referential integrity," or in simple words, the logical connection between the data must be completely preserved. So if, for example, a specific employee sent an email about a bug, opened a ticket, and then wrote code that fixes it, this specific sequence will remain. This is perfect for training AI agents on solving complex problems in the real world, but it also makes anonymity… a little less anonymous.
By the way, when Meta tried in the past to collect internal information from employees, to understand processes and especially how humans work and conduct themselves, it encountered a protest that caused it to scrap the plan. But in an airline that has gone bankrupt, there is simply no one to protest or oppose the move.
What will happen with passenger information?
Until now we have talked about organizational information and internal data, but what happens with customer data? Well, data such as personal profiles, order history, and conversations with the service center are not included in the current deal, but before you get excited… it turns out this is not because of the company's concern for their privacy. Spirit simply reserves the right to sell them separately to other business entities in the tourism industry. Have a pleasant flight.
