Spirit Data Sale Draws Union Objection as Analysts Weigh Google's AI Motives

Google's agreement to buy the internal records of a collapsed airline has drawn a formal objection in bankruptcy court, and the analysts reading the filing say the data could sharpen the search company's understanding of how travel pricing actually works.

The deal itself was reported last week. Per an August 14 court filing, Google was the highest bidder at auction for Spirit Airlines' business data, agreeing to pay $10 million. What has emerged since is a clearer picture of what is in the dataset, who else wanted it, and why not everyone is satisfied with the privacy safeguards.

The purchase covers employee emails, chats and messaging data, plus records from Spirit's finance and accounting systems, aircraft operations, revenue, website analytics and loyalty program. Before any transfer, the filing states that a third party will deidentify the data, removing or transforming anything that could be linked to a consumer.

"We acquired part of an enterprise dataset from Spirit Airlines, which can be helpful in improving our products and AI models," a Google spokesperson told PhocusWire, adding that the company "will not receive any personal information from this dataset."

Why the data is valuable

Eric Léopold, founder and managing director of Swiss consultancy Threedot, said the sale gives Google access to proprietary airline data of a kind its models rarely see. "AI models tend to hallucinate when they don't know something, often because they were never exposed to actual facts or data," he said. "Coding or mathematics sit on public information and work well with LLMs. Aviation runs on larger proprietary datasets that AI models usually don't see."

Timothy O'Neil-Dunne, principal at consultancy T2Impact, was blunter about the route Google took. "That is great training data, something that they could have received from their own scanning of emails but not legally," he said, calling the global view of an airline that went through a crisis a unique opportunity.

Gemma Timmons, director of operations and chief of staff at aviation data provider OAG, pointed to the runner-up bid as a signal of the data's worth. Mercor, a company that buys data and supplies it to frontier AI labs, bid $7.5 million. She also noted that the dataset spans a technology stack Google does not own, including Microsoft 365, SAP, Navitaire and UKG, which gives Google a detailed view of how a large regulated business runs across its people and systems.

Léopold flagged the limits. Spirit was a U.S. low-cost carrier, so the records say nothing about non-U.S. or full-service operations.

The agentic angle

Google is currently running a U.S. test of agentic hotel booking in AI Mode, with flight booking planned. The Spirit filing does not mention that program, but Timmons said pricing, booking-curve, transaction, refund, voucher and disruption data could help Google model how a fare behaves through servicing and disruption. That is the part of the journey an agent has to get right before travelers trust it with money.

For operators watching Google move from referring bookings to taking them, that is the line worth tracking. O'Neil-Dunne is not convinced agentic booking is the goal, since Google already profits from informing bookings. He sees a different payoff: "The purchase provides a treasure trove of historical data that allows Google to understand pricing decisions, not just observe them. This means in the future, Google can use this data to determine whether a price is a good or bad one."

The privacy fight

Former Spirit flight attendants have filed an objection in bankruptcy court, arguing the filing does not address whether the contents of records will stay confidential. Their union, the Association of Flight Attendants-CWA, took issue with language requiring deidentification to preserve "referential integrity across the data set," meaning links between records survive the scrubbing.

"With a group the size of the Spirit Flight Attendant population, our union has significant concerns that it may be possible that information about identifiable individuals or small identifiable groups can still be reconstructed and determined," the union said.

Timmons noted that privacy terms appear to be priced into the outcome. Mercor offered $10 million if it could run deidentification in house, and sellers designated its $7.5 million third-party option as the alternate. The court declaration says privacy terms could be outcome-determinative.

The court has not yet approved the sale.