Google has agreed to pay ten million dollars for a tranche of internal business data from bankrupt American carrier Spirit Airlines, with the intention of using the material to train its artificial intelligence models and improve its products.

The dataset is substantial. Court filings put the volume at roughly 100 million emails and 500 million Microsoft Teams messages, alongside spreadsheets, calendars, marketing materials, HR documents, financial databases, audits and presentations. Google secured the data through a bankruptcy auction, though a federal judge must still approve the transaction, with a hearing scheduled for Wednesday. Should the court decline to sanction the deal, AI hiring platform Mercor is lined up as runner-up bidder, having offered 7.5 million dollars for the same dataset.

Google has moved to address privacy concerns, stating that a third party will strip all personally identifiable information and customer records from the material before it changes hands. The company described the acquisition as a means of improving enterprise-focused products and AI capabilities and said no personal data will reach its systems directly.

The sale forms part of the broader wind-down of Spirit’s remaining assets through the US bankruptcy courts. Spirit ceased operations earlier this year after a government rescue package fell apart. The airline had been in discussions with the Trump administration over a deal worth up to 500 million dollars, but a key group of creditors rejected the proposed terms. Rising jet fuel costs, partly tied to geopolitical tensions involving the United States, Israel and Iran, pushed the carrier’s projected 2026 operating margin towards negative 20 per cent. The collapse left around 17,000 employees and contractors without work and marked the first disappearance of a major American carrier since Midway Airlines folded in the aftermath of the September 2001 attacks.

Spirit had filed for bankruptcy for a second time in August 2025, having reported losses of nearly 257 million dollars after emerging from its first insolvency process just months earlier in March.

For Google, the purchase underlines how the race to accumulate high-quality training data for large language models is drawing tech companies into unexpected corners of the market, including the wreckage of corporate insolvencies.