Google is paying $10 million to acquire Spirit Airlines’ business data to use for training its AI models. The dataset includes internal communications, emails, spreadsheets, bookings, frequent‑flyer information and employee HR records, and it was sold after Spirit halted operations in May; the sale attracted a competing bid of $7.5 million from Mercor. Company documents state the data will be stripped of details that could identify individuals before Google receives it, and Google will not receive personal information through the purchase.
Google acquired a dataset from Spirit Airlines that includes internal communications and company records such as emails, spreadsheets, booking records, frequent‑flyer information and employee human‑resources files. A separate report indicates the dataset also contains Microsoft Teams messages and calendar entries. The dataset further includes marketing records, productivity files and operations records. These materials together comprise parts of Spirit’s enterprise dataset offered for sale.
Company documents state that identifiable personal details will be removed from the dataset before Google receives it. Those documents state Google will not receive personal information through the purchase. The dataset’s inclusion of employee HR records is addressed by those de‑identification steps. The descriptions specify removal of details that could identify individuals prior to transfer.
Taken together, the materials sold to Google cover communications and operational records while the seller says identifying details will be stripped beforehand. The company states Google will not receive personal information as part of the acquisition. The data transfer is described as completed only after those de‑identification measures are applied.
Spirit Airlines halted operations in May and is selling its remaining assets through a bankruptcy process. The bankruptcy sale includes the airline’s business data among the assets being offered, and the offering covers parts of Spirit’s enterprise dataset that had been maintained as company records. The bankruptcy sale could end with another carrier buying the company, including its data. Mercor placed a $7.5 million bid for the data.
The proposed sale requires court approval. U.S. Bankruptcy Judge Sean Lane will consider the deal at a hearing on Wednesday. The submission of bids for the data is proceeding as part of Spirit’s broader asset liquidation under the bankruptcy case. The legal review is being handled through the bankruptcy court process.
The bankruptcy proceedings remain active. Further court action is pending.
The Spirit Airlines dataset sale to Google is described in reporting as part of a broader trend of companies paying for large datasets to support AI development. In February 2024, Reddit gave Google access to about two decades of user-generated content in a deal reportedly worth $60 million a year. In May 2024, OpenAI reached an agreement to bring Reddit content into ChatGPT. These transactions are cited alongside the Spirit sale as examples of commercial agreements that provide large-scale content to technology companies.
In January, the Wikimedia Foundation announced agreements with Microsoft, Google, Amazon, Meta and others for large-scale access to Wikipedia content. The Spirit deal, involving enterprise records and communications, is presented in the reporting as another commercial data transaction. Reporting notes that the Spirit dataset sale proceeded through a bankruptcy process and included de‑identification measures before transfer.
Reports list these arrangements together as instances of paid access to datasets by AI and technology firms. The referenced deals are dated February 2024, May 2024 and January in the reporting.
Google’s $10 million purchase of Spirit Airlines’ business dataset underscores the financial scale of recent deals to license data for AI training. The transaction, which the seller says will remove identifiable personal details before transfer so Google will not receive personal information, is cited in reporting as part of a broader movement of companies licensing large datasets to develop and improve AI models.
