AI & MACHINE LEARNING DATA ACQUISITION

Inside Google's US$10m Data Acquisition of Spirit Airlines

TM
Techmediaglobal
| 5 min read
US$10M
WINNING BID
100M
INTERNAL EMAILS
500M
TEAMS CHATS
7.5B
TRANSACTION RECORDS

Google has won a bankruptcy court auction for the corporate data of defunct carrier Spirit Airlines, paying US$10m for a trove of internal records the tech giant plans to use to train its AI models. The deal outbid AI recruitment firm Mercor and has reignited debate over how far companies will go to source specialised training data.

How the Deal Came Together

Spirit Airlines declared bankruptcy and ceased operations entirely on 2 May 2026, leaving the carrier to liquidate its remaining corporate assets in an effort to pay down roughly US$8.1bn in outstanding debt.

According to Bloomberg Law, Google emerged as the winning bidder for the airline's digital repositories, beating out a US$7.5m offer from Mercor. The US Bankruptcy Court for the Southern District of New York has named Mercor as the designated backup buyer should the primary agreement fail to close.

Court filings indicate Google intends to use the acquired material to train its large language models, gaining access to what is described as an unusually large and highly specialised operational dataset.

What's Inside the Dataset

Court records show the internal-communications trove alone spans 100 million emails and 500 million Microsoft Teams chats.

The transaction also includes decades of historical travel data: 7.2 billion records covering competitors' flights, 7.5 billion passenger transaction records dating back to 2008, and more than 175,000 employee records going back to 1986.

Beyond communications and travel history, Google gains proprietary pricing-curve data, marketing campaign records, project management documents and revenue figures, along with detailed records on aircraft operations, employee productivity, audits, fraud and human resources.

Privacy Safeguards and Regulatory Concerns

A transaction of this scale naturally raises concerns for anyone who has ever interacted with the carrier, whether as a customer, employee, contractor or investor. Notably, the bankruptcy agreement explicitly excludes personal passenger data: Google will not gain access to Spirit's 97.5 million passenger profiles, and the 50.2 million customer records tied to the Free Spirit loyalty programme remain entirely outside the deal.

The court has further mandated that all data transferred to Google must first be scrubbed of personally identifiable information by an independent third party.

Even so, the deal raises broader questions about whether selling customer-service interactions to train third-party AI models could become a wider industry pattern. Adil Tahiri, Chief Technology Officer at Konecta, notes that the acquisition gives new meaning to the familiar call-centre disclaimer about calls being recorded for training purposes, and questions whether it signals AI firms increasingly looking to partner with or acquire customer-service organisations.

"For Google, that data is clearly a gold-mine, worth more than a staggering US$10m."

— David Fischer, Chief Revenue Officer, Luware

Do Companies Even Need to Buy This Data?

Not every organisation needs to spend millions on external datasets. Fischer argues that most enterprise contact centres already sit on valuable repositories of real-time conversational data, and that every call, chat and customer interaction can fuel an AI feedback loop without a costly acquisition.

With strong privacy safeguards, proper governance and human oversight, businesses can responsibly put their own existing interactions to work refining self-service tools, supporting human agents and improving customer experience from within. For companies chasing a new vertical entirely, though, an external acquisition like this one offers a faster route to sector-specific dominance — in this case, a potential runway for Google to build and commercialise aviation-specific AI models.

Key Takeaways

  • Google won a bankruptcy court auction for Spirit Airlines' digital assets, paying US$10m and outbidding a US$7.5m offer from Mercor.
  • The dataset includes 100 million emails, 500 million Teams chats, and billions of travel, transaction and employee records.
  • Personal passenger data is explicitly excluded — 97.5 million passenger profiles and 50.2 million loyalty records stay out of the deal.
  • An independent third party must scrub all transferred data of personally identifiable information before Google receives it.
  • Industry voices are split on whether the deal signals a coming wave of AI firms acquiring or partnering with customer-service organisations.
  • Experts note most enterprises already have valuable conversational data in-house and don't necessarily need costly external acquisitions to benefit from it.
Tags: Google Spirit Airlines Data Acquisition AI Training Data Large Language Models Bankruptcy Data Privacy Aviation Tech