Google, after the bankruptcy of Spirit Airlines, paid $10 million in an auction to buy a large portion of the airline's internal data; a deal that shows how valuable proprietary corporate data has become for companies active in the field of artificial intelligence.
According to ITPna, Spirit, which had been struggling financially since 2020, ceased operations in May 2026 and entered liquidation. During the sale of the company's assets to settle debts, a massive set of corporate data was also put up for auction and Google won that portion of the bidding.
Exactly what did Google buy?
The scale of this dataset is considerable. According to court documents, the purchased assets include more than 100 million emails and about 500 million items of data from Microsoft Teams. There are also over 17 million OneDrive files, more than 20.5 million SharePoint items, and over 30 million recorded customer service calls in the collection.
In addition, Google will gain access to more than 15 million customer conversation records, about 600,000 ServiceNow tickets, and more than 13.7 million active email addresses used in the Oracle Responsys marketing system.
Information on more than 11 million in-flight internet sales is also part of this collection.
But data related to Spirit's flight operations may be more attractive for developing specialized AI models. The set includes information on more than 763,000 flights, 5 million crew pairings, over 1.2 million fuel receipts, and purchase records for 787,452 parts.
Why are these data important for AI?
Public data are abundant for training AI models, but the value of proprietary and specialized data is that they relate directly to an industry and a set of real processes.
Google has said the purpose of buying these data is to improve products and develop AI services. The Spirit dataset could, for example, serve as a basis for developing models that have a more precise understanding of airline operations, flight scheduling, crew management, fueling, parts, customer service, or even financial and organizational processes.
The importance of such data increases as the AI industry moves toward smaller, specialized models. Instead of trying to understand all topics, such models can be trained on a specific domain like aviation and perform better on specialized tasks.
Interestingly, Google was not the only company interested in these data; according to a published report, Mercor, a company that provides data for training AI models, was Google’s rival in this auction; a fact that demonstrates the potential value of this dataset for training and developing AI systems.
Concerns about customer information
Given the large volume of this information, privacy concerns are natural. However, court documents say the data were de-identified before being offered in the auction, and Google has also committed to removing any remaining personally identifiable information from the set.
Nevertheless, anonymizing large and complex corporate datasets is not always simple. A collection that includes millions of emails, calls, files, and operational records can, if auxiliary information remains or multiple sources are combined, create a risk of re-identifying some individuals. For this reason, the manner of scrubbing and using these data will likely be important topics after the deal is finalized.
A new market for proprietary data
The purchase of Spirit's information is not just an odd transaction during an airline's liquidation; it may be a sign of a new market forming for proprietary and industrial data.
In an era when technology companies need high-quality data to train and improve AI models, real information derived from an organization's activities can acquire independent value beyond the company itself. Such data allow models to encounter real examples of business processes, operations, and human interactions.
On the other hand, the deal raises important questions about the ownership and use of data from organizations that go bankrupt: if a company shuts down, to what extent can the data produced by its employees and customers become a sellable asset, and what restrictions will a new buyer have on their use?
According to The Register, Google has said it purchased these data to develop AI products and services.
Ultimately, the Spirit deal shows that in an AI-driven economy, a company's value is not limited to its brand, equipment, or physical assets; data accumulated over years of operation can also become one of an organization's most valuable assets.