Google buys a dead airline's data for $10 million, but here's what it's not getting

Google has bought the leftovers of Spirit Airlines, winning a bankruptcy auction for the dead carrier's internal business data with a $10 million bid that it says will feed product development and AI model training. The sale still needs sign-off, with a hearing set for August 19 before Judge Sean Lane in the US Bankruptcy Court for the Southern District of New York.The company was quick to spell out what it had not bought. A spokesperson told Business Insider the company was buying internal data and custom software, "but we are not buying their customer or credit card information." Google added that anything it receives will be "rigorously scrubbed of any personally identifiable information by a third party before receipt."What is inside the Spirit Airlines dataset Google just boughtThe haul is enormous. Court filings list roughly 100 million emails and 500 million Microsoft Teams messages, about 30 million lines of code along with development metadata and software models, and more than 175,000 employee records going back to 1986. There is revenue data, aircraft operations, employee productivity, audit and fraud material, marketing campaigns, HR files, board presentations and budget documents. Filings also point to pricing from 7.2 billion competitor flights and roughly 7.5 billion passenger transaction records stretching back to 2008.What Google did not get: 97.5 million passenger profiles and an estimated 50.2 million Free Spirit loyalty records. One caveat sits in the paperwork, though. The estate kept the right to sell its customer list, including individual traveller spend aggregated by year, to buyers in the hospitality or travel business.What the fine print of the Google-Spirit data sale actually saysThe paperwork on that scrubbing promise is more specific than the statement suggests. Under the sale agreement filed on August 14, the deidentification agent must be acceptable to or designated by the buyer, Google covers the entire cost of the work, and Google gets a reasonable chance to review and comment on the process before it signs off. The standard in the document is a scrub reasonably satisfactory to the buyer.The certification also has to preserve referential integrity across the data set. In plain terms, the links between records survive even after the names come off, so one pseudonymous person still runs from an email to a support ticket to a code commit. That linkage is exactly what makes the archive useful for training agents that handle multi-step work.The auction that got Google here was competitive. Google opened at $5 million, AI data firm Mercor countered at $5.2 million and then offered $7 million if it could take the raw data first and anonymise it itself. Google closed at $10 million. Mercor stays on as backup buyer at $7.5 million.Get the latest technology news and updates. Download the TOI App.