Google has agreed to buy a large batch of data from Spirit Airlines for $10 million. The deal was part of Spirit's bankruptcy proceedings and closed after an auction last Friday.
The data includes Spirit's software code, internal messages, and financial and operational records. A Google spokesperson said the company plans to use the information to improve its products and AI models.
The purchase is a small example of a much larger trend. Artificial intelligence companies are buying up corporate data of all kinds as they look for new ways to train their systems.
Why Companies Want This Data
AI models learn by studying huge amounts of information. The more data a model has, the more examples it can draw from when writing text, answering questions, or making recommendations.
For years, AI companies pulled much of this material from the open internet. That practice has become harder as legal challenges over scraping pile up.
As a result, firms are turning to companies willing to sell their records directly. The five biggest cloud providers plan to spend $800 billion combined on AI infrastructure this year, and AI labs are paying huge sums to use those facilities.
That kind of money makes a $10 million data deal look small to a company like Google. For the seller, though, it can be a large payout.
Mercor, an AI data start-up, also bid on the Spirit Airlines data. It offered $7.5 million but lost the auction to Google.
Bigger Deals Already In Place
Other companies have already struck licensing agreements with AI firms. Cloudflare bought Human Native in January, a marketplace that connects content creators with AI developers.
Two months later, News Corp signed a deal with Meta Platforms. The agreement gives Meta access to News Corp content for $50 million a year.
Smaller firms are getting involved too. micro1, another AI data venture, said it committed more than $20 million to license operational data over an 11 day period this month.
Reddit signed licensing deals with OpenAI and Google back in 2024. Each deal brings Reddit around $60 million a year.
John Wiley and Sons, a publisher of academic material, has licensing agreements with Amazon Web Services, Microsoft, and Anthropic. Those deals have generated more than $110 million since 2024.
Not every company wants to sell. Opera, the web browser maker, has avoided data deals over privacy concerns. Its finance chief said trust with users matters more than short term income.
Other executives worry that AI companies could use their data to build competing products. Hostelworld's chief executive called his company's data "the crown jewels" during a call last month.
Reddit's situation shows where this market may be headed next. The company is currently in talks with Google, and some investors expect the new agreement to be worth more than the original deal signed in 2024.
