US Court Lets Google Buy Spirit Airlines Data for $10 Million to Train AI
A bankruptcy court in New York has allowed Google to buy the internal data of the shuttered Spirit Airlines for $10 million. The package includes emails, chats, customer support call recordings and flight archives, and Google wants to use it to train AI. The court agreed only after Google promised to strengthen privacy protections, since unions and more than 120 lawmakers had objected. The archive of a bankrupt company is becoming an asset for training models.
AI-processed from Habr AI; edited by Hamidun News
When a company goes under, it doesn't just leave behind furniture and aircraft. It also leaves a digital archive: correspondence, documents, call recordings and technical logs. Now that archive can be sold, and the buyer can use it to train neural networks. A federal bankruptcy court in New York approved exactly such a deal: Google will pay $10 million for the internal business data of budget airline Spirit Airlines. Reuters reports this.
What Google Actually Bought
Spirit Airlines suspended operations this spring. Over the summer, as part of the asset liquidation, Google won an auction for its internal data. The list includes emails and chats, files from OneDrive and SharePoint, customer support call recordings, ServiceNow tickets, and archives of information on flights, fuel and parts purchases, and crew details.
Google said it intends to use these materials to develop its products and train artificial intelligence models. For AI, such an archive is valuable because it is live working correspondence and the real operational processes of an airline: how people solve flight problems, how they communicate with customers, how they plan repairs and purchases. The public internet usually does not provide details at this level.
Why the Court Initially Hesitated
The deal drew objections. Unions and more than 120 American lawmakers expressed concern. At the end of October, politicians asked Google to remove as much employee information as possible. According to them, the data may include hundreds of millions of emails and employee messages in Microsoft Teams. This is especially sensitive for former employees: in work chats, people write not only about schedules but also about health, conflicts and dismissals.
Also in October, the data protection ombudsman recommended that the court approve the sale on the condition that Spirit's customers' personal information be removed from it. In the first half of November, Google signed an agreement with the flight attendants' union. The document describes a mechanism that separates work data created by employees from their personal information. The court approved the sale because Google committed to implementing additional privacy protection measures.
What This Means for the Industry
The deal reveals a market that has rarely been discussed: the corporate data of bankrupt companies as a source of training material. Licensing agreements with publishers and platforms are expensive, while the archive of a closed company goes for auction prices. If this scheme becomes routine, liquidators will start putting not only property but also digital storage up for sale.
There are several practical takeaways.
- Product teams and founders should think in advance about what happens to company data in a bankruptcy or business sale. Terms in privacy policies and employee agreements then become a real document, not a formality.
- If you collect training data, buying a bankrupt company's assets is possible, but without a clear separation of work and personal data you risk protests, lawsuits and reputational damage.
- Companies that store correspondence in cloud services such as Teams or SharePoint should limit retention periods: the less surplus data accumulates, the less will end up in a potential sale.
Limitations and Open Questions
The report does not make clear how reliably the data separation mechanism works or who will verify its implementation. It also does not name the specific models or products Google plans to train on this data. The flight attendants' union received guarantees, but other employee categories, judging by the description, may be protected less well. These details will determine whether the deal becomes a template for future sales or a reason for new restrictions.
For now, a precedent has been set: an American court has allowed the internal archive of a bankrupt company to be turned into raw material for AI, requiring in return that privacy protections be strengthened.
- Nov 21, 2026Apple Is Building an AI Agent and Making It Harder for Outsiders to Get Access: Permission Now Takes 14 Steps, Not 4
- Nov 21, 2026Gemini 4 Neon shrank the TS Rust binary from 47.44 to 7.97 MB in under 10 hours
- Nov 21, 2026A swimming team captain handed ChatGPT three scheduled tasks and cut her routine work by a multiple
- Nov 21, 2026MIRA Neural Network Renders Rocket League at 20 FPS, but Ping Hits 460 ms
Key Questions Answered
What Google Actually Bought?
Spirit Airlines suspended operations this spring. Over the summer, as part of the asset liquidation, Google won an auction for its internal data. The list includes emails and chats, files from OneDrive and SharePoint, customer support call recordings, ServiceNow tickets, and archives of information on flights, fuel and parts purchases, and crew details. Google said it intends to use these materials to develop its products and train artificial intelligence models.
Why the Court Initially Hesitated?
The deal drew objections. Unions and more than 120 American lawmakers expressed concern. At the end of October, politicians asked Google to remove as much employee information as possible. According to them, the data may include hundreds of millions of emails and employee messages in Microsoft Teams. This is especially sensitive for former employees: in work chats, people write not only about schedules but also about health, conflicts and dismissals.
What This Means for the Industry?
The deal reveals a market that has rarely been discussed: the corporate data of bankrupt companies as a source of training material. Licensing agreements with publishers and platforms are expensive, while the archive of a closed company goes for auction prices. If this scheme becomes routine, liquidators will start putting not only property but also digital storage up for sale. There are several practical takeaways.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.