The bankruptcy auction for Spirit Airlines' internal data archive has turned into a bidding war between Google, Mercor, and Micro1, with a judge set to rule on September 9, 2026, on both the highest offer and a privacy objection from the flight attendants' union. The defunct airline's archive includes roughly 100 million emails, 500 million Teams messages, and 30 million lines of code, plus employee records reaching back to 1986. Google initially won the auction on August 14, 2026, with a bid of $10 million. Days later, Micro1 sent Spirit's lawyers a $12.5 million offer, submitted after the bidding deadline, topping Google by $2.5 million. The bankruptcy judge pushed the approval hearing to September 9, 2026, to weigh Micro1's higher bid against the union's formal objection that preserved links between records could allow re-identification of individuals.
A Data Trove From a Fallen Airline
Spirit Airlines stopped flying on May 2, 2026, after its second bankruptcy, leaving over 17,000 workers out of work. The airline carried roughly $8.1 billion in debt at the time of its collapse. The archive now up for sale captures the inner workings of that 17,000-person workforce: internal emails, Microsoft Teams messages, and software code that show how the company operated day to day. Passenger profiles and frequent flyer accounts are excluded from the sale, according to the auction terms. A third party will de-identify the archive before Google receives it, stripping names, addresses, and other personal details. Google also agrees not to reverse the de-identification process. The sale agreement, however, requires links between records to stay intact, which is exactly what the union says creates a privacy risk.
The union fears that individuals in a 17,000-person workforce could potentially be re-identified from patterns in the data, even after de-identification. Because the links between records remain, someone with enough context could piece together who did what, when, and with whom. The union formally objected on those grounds, and the judge will hear that argument alongside the question of which bid should win.
The Bidding War: Google, Mercor, and Micro1
Google won the auction on August 14, 2026, with a $10 million bid. Mercor, an AI training company, had earlier bid $7.5 million for the same data. Micro1, an AI training startup, came in late with $12.5 million, a full $2.5 million more than Google's offer. Ali Ansari, CEO of Micro1, argued that Google's price was far too low for data that valuable. He said the archive holds operational details that are crucial for training advanced AI models, especially as public internet data becomes scarce.
The public internet shows what companies say about themselves, but almost never how a company runs. An archive like Spirit's captures work as it happened. The record of decisions, coordination, and real mistakes is worth more than what's left of the public internet. That logic drove Micro1 to submit its late bid, and it may drive the judge to accept it. Bankruptcy code leaves room for a late bid if it puts more money in creditors' pockets, and $12.5 million is more than $10 million by a meaningful margin.
Why AI Firms Want Failed Companies' Data
AI companies have historically built models on public internet data like websites, books, forums, and code. That supply has a ceiling and a blind spot. According to BTUAI, only about 15% of the world's knowledge has ever been digitized. The rest sits in private archives, corporate servers, and internal systems that never see the light of day. Internal communications offer rich insights into how companies truly operate, revealing decisions and processes that no public source can match.
The first battles over AI data were fought over scraping the public web in copyright suits. The next battles are over private archives of how businesses really work. A small market has formed this year around data of failed companies. Startups are selling internal records and messages to AI firms for creditor recovery. Bankruptcy treats data as an asset of the business, so when a company dies, its inbox becomes a salable commodity. Spirit found out what its inbox was worth only after it died, and never grasped the privacy risks its employees faced.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
The Legal Framework Lags Behind
Bankruptcy law was written in 1978, before the value of data or privacy concerns were imagined. The code treats data like a gate slot or a plane in bankruptcy proceedings, an asset to be liquidated for creditors. It has no mechanism for weighing the long-term privacy consequences of selling millions of internal messages. The judge's ruling on September 9, 2026, will test whether that framework can handle the modern reality of data as both treasure and liability.
The hearing has two big issues: the union's privacy objection and Micro1's extra $2.5 million bid. Until the judge rules, Google has won the bid, not the data. The union's objection could block the sale entirely, or the judge could accept Micro1's higher offer and push the union's concerns aside. Either way, the precedent is already visible regardless of the judge's decision. Your company's communication archive is now an asset, and someone will bid on it when you fail.
What This Means for Businesses
This case sets a precedent, treating company data as a valuable asset in bankruptcy. Businesses should re-evaluate their data retention and privacy policies. Leaders should ask uncomfortable questions about data retention, privacy protection, employee and vendor awareness, and ownership in contracts. If Spirit's archive can fetch $12.5 million, other failed companies' archives will fetch similar sums, and the workers whose messages are sold will have no say in the matter.
The union's objection is not just about Spirit. It is about every company that stores years of internal communications without thinking about what happens to them in a liquidation. The de-identification process sounds protective, but the preserved links between records are what makes the dataset valuable for training. That same value is what makes re-identification possible. The judge will have to decide which concern wins.
The September 9, 2026 hearing will resolve two questions. First, whether the union's privacy objection blocks the sale. Second, whether Micro1's $12.5 million bid, submitted after the deadline, should replace Google's $10 million offer. Google says the data will improve its products and AI models. Micro1 says the data is worth more than Google paid. The union says the data should not be sold at all without stronger privacy protections.
The archive includes approximately 100 million emails, 500 million Teams messages, and 30 million lines of code. Employee records reach back to 1986, covering decades of hires, firings, and personnel changes. The airline had roughly $8.1 billion in debt when it filed for its second bankruptcy, and over 17,000 workers were left out of work when it stopped flying on May 2, 2026. Those workers now face the possibility that their internal messages, performance reviews, and coordination records will be sold to an AI training firm.
Bankruptcy law, written in 1978, has not caught up with technology. It treats data as an asset without considering the privacy implications of selling it. The judge's ruling will set a precedent for every future bankruptcy involving a company with a large digital footprint. The AI data wars have found a new battleground in bankruptcy auctions, and the outcome of this case will shape how those battles are fought.