A Bankruptcy Auction Unlike Any Other
When Spirit Airlines ceased operations in May 2026 after years of financial struggle, the sale of its assets was expected. But alongside airport slots and equipment, a more intangible asset went on the block: its data. Google emerged as the top bidder
with a $10 million offer, beating out AI startup Mercor, which bid $7.5 million. The deal, part of Spirit's bankruptcy proceedings, is now pending a court hearing for final approval. This was delayed after the Association of Flight Attendants-CWA filed an objection, raising concerns about the sale of employee data. The sale highlights a new frontier in corporate liquidations, where a company's digital ghost—its emails, operational logs, and internal communications—can be one of its most valuable remaining assets.
What is 'Deidentified Data'?
The data package Google is buying is vast. It includes around 100 million employee emails, 500 million Microsoft Teams messages, 7.5 billion passenger transaction records, and 30 million lines of software code. Crucially, Google says it is not acquiring personally identifiable information (PII) like passenger names, loyalty program information, or credit card details. Before Google receives the data, a court-appointed third party is tasked with "rigorously scrubbing" it to anonymize the records. The goal is to de-identify the data so it can't be linked to a specific individual. However, labor unions and privacy advocates have expressed concern, arguing that with the vast datasets Google already possesses, re-identifying individuals or small groups from supposedly anonymous data could be possible.
Why Google Wants a Failed Airline's Brain
For Google, this isn't about air travel; it's about artificial intelligence. In a statement, the company said the dataset will be "helpful in improving our products and AI models." This kind of data is a goldmine for training AI. It provides a real-world blueprint of how a large, complex company operates—from internal problem-solving in Teams chats and emails to revenue management, flight operations, and marketing campaigns. An AI model trained on this data could learn to perform better at tasks like customer service, financial analysis, and managing logistics. As AI companies have largely exhausted public data from the internet, they are now seeking private and corporate datasets to give their models a competitive edge.
The Bigger Picture: Data as a Corporate Asset
The battle between Google and Mercor, a prominent AI data startup, signals a fierce competition for high-quality training material. Mercor's rival bid underscores how valuable this kind of comprehensive corporate data is to the burgeoning AI industry. This sale could set a major precedent. For other companies facing financial distress, their operational data could become a key asset to be sold off to creditors. It raises significant questions about the future of workplace privacy. Even if personal names are removed, the content of millions of internal emails and messages—the digital record of employees' work lives—is being sold as a commodity. The court's decision will be watched closely, as it will weigh the financial relief for Spirit's creditors against the broader ethical and privacy implications of the sale.
