Spirit ceased operations on May 2, 2026, after its second Chapter 11 filing, citing high debt, fuel costs, and failed financing efforts. The airline is now auctioning remaining assets in the U.S. Bankruptcy Court for the Southern District of New York to address roughly $8.1 billion in liabilities. Google outbid AI data firm Mercor, which offered $7.5 million and was named the backup buyer. A hearing on the sale is scheduled for Wednesday.
What’s in the dataset
Court documents describe an extensive enterprise archive that includes about 100 million emails, 500 million Microsoft Teams chats and collaboration records, millions of OneDrive and SharePoint files, calendars, spreadsheets, and IT tickets. It also covers operational data such as flight records, crew pairings, fuel and maintenance information, revenue management, pricing models, booking patterns, and roughly 30 million lines of custom software code from Spirit’s systems. Some records date back years, including employee-related files from the mid-1980s.
Google has emphasized what is not included. The purchase excludes customer profiles, Free Spirit loyalty program records, credit-card information, and other personally identifiable passenger data. A third party, designated and paid for by Google, will scrub remaining personally identifiable information before any transfer, with the buyer agreeing not to attempt re-identification.
More Context
Spirit’s collapse followed years of financial strain in the ultra-low-cost airline sector. The data sale is one of several asset auctions; other pieces, such as airport slots, have gone to different buyers. The de-identification process and court oversight are intended to address privacy concerns, though the scale of the archive has drawn attention to how bankrupt companies’ digital remnants are becoming training fuel for large AI models.
A bankruptcy judge will review the transaction this week. If approved, the anonymized data would transfer after the third-party scrubbing is complete. If the Google deal does not close, Mercor remains the designated backup. The episode illustrates both the hunger for specialized training data and the unusual ways that data is now changing hands.
