Google won a bankruptcy auction for Spirit Airlines’ business data and software code for $10 million, according to an August 14 court filing. The winning bid followed a brief escalation with AI recruiting firm Mercor, which had offered $7.5 million. Judge Sean H. Lane must still approve the sale at a hearing set for August 19, 2026. What follows is not primarily about Spirit Airlines, or even about Google. It is about the assignment of tradeable value to the recorded interior of a failed organization.
In this article
What Google Actually Bought
The deal grants Google access to approximately 3.4 million payroll records, 100 million emails across 80,000 accounts, and employee information stretching back to 1986, according to the filing. It includes 100 million emails, 500 million Microsoft Teams records, 17.1 million OneDrive files, 20.6 million SharePoint files and 667,563 IT tickets, as well as 516 source code repositories containing about 30 million lines of code, 372,585 commits and 43,170 pull requests, together with bug reports, review discussions, tests and build logs.
Google will also gain access to Spirit’s pricing models, booking curves, flight behavior data, inflight purchase and Wi-Fi sales records, refund histories, and travel package information. The records encompass about 7.5 billion transaction entries dating to 2008, 7.25 billion observations of competitors’ flight prices, data on 763,391 Spirit flights and over 5 million crew pairings. The package is comprehensive enough to reconstruct most operational decisions Spirit made across nearly two decades.
The deal explicitly excludes all customer and loyalty data – covering 97.5 million passengers, 52.4 million loyalty members, and 740,000 co-branded cardholders – and the acquired data will be scrubbed of personal information. Any data Google receives will be rigorously scrubbed of any personally identifiable information by a third party before receipt. The data is not about passengers. It is about the people who ran the airline.
The Context: Two Bankruptcies and a Shutdown
Spirit Airlines filed for chapter 11 twice: first on November 18, 2024, and again on August 29, 2025. It emerged from a pre-arranged process roughly four months later on March 12, 2025. Spirit Airlines stopped flying on May 2, 2026. Its second Chapter 11 case shifted from a proposed reorganization to a court-supervised wind-down.
The ultra-low-cost carrier plunged into bankruptcy after a $3.8 billion deal to sell itself to JetBlue was blocked by a federal judge and after previous efforts to combine with rival Frontier Airlines fell apart. The airline’s collapse was not sudden. It was structural, predictable, and well-documented in the financial press for more than a year before the first filing. By the time Spirit stopped flying, it had accumulated losses exceeding $2.5 billion since 2020.
The disposal of its digital assets followed standard bankruptcy procedure. Aircraft were returned to lessors, airport slots were sold, and attorneys continued handling asset disposition through the summer of 2026. At Asset Hub, a specialized broker that handles liquidation data from bankrupt companies, comparable internal data deals typically go for $10,000 to $100,000 per company. Dividing Spirit’s $10 million deal by the upper end of that range yields a multiple of 100x.
The 100x Multiple and the Emergence of a Market
The 100x multiple should be read as a number signaling that the AI training-data market has begun seriously pricing a new asset class: corporate internal communications. This is the structural observation worth isolating. Google did not acquire Spirit’s data because it contained unique insights into airline economics. It acquired the data because it represents a large, time-sequenced corpus of workplace decisions, software modifications, and operational responses – material that is difficult to synthesize and expensive to generate.
Together, those materials preserve connected sequences of workplace activity: problems emerge, employees discuss them, software changes, operational responses follow and results accumulate. Such sequences may be particularly useful for training AI systems expected to perform multi-stage office, engineering and logistical work.
That is increasingly important as technology companies search for additional sources of training materials beyond publicly available internet content. The bidding war itself – brief as it was – confirms latent demand. AI data company Mercor also participated in the bankruptcy auction, initially offering $7.5 million after Google entered with a $5 million bid. Google ultimately raised its offer to $10 million, securing the winning position. Mercor was named as backup buyer in the event Google’s bid failed court approval.
The $10 million figure is modest by acquisition standards but significant as a price signal. It establishes a floor for similar transactions and makes explicit what was implicit: that the documented interior of a large failed enterprise has measurable value to firms building models meant to automate or assist with enterprise work. Cases in which the internal data of failed companies becomes the target of a bidding war among AI companies during liquidation proceedings may well spread beyond Spirit going forward.
Employee Data, Legal Constraints, and Congressional Attention
The asset schedule filed with the court lists about 100 million emails across 80,000 accounts, 500 million Microsoft Teams items, 175,658 employee records, 3,426,618 payroll records and 148,018 employee tax forms, while excluding 97.5 million passenger profiles. The exclusion of passenger data was deliberate and legally necessary. The Code was specifically amended to restrict or prohibit the sale of customers’ personally identifiable information when in violation of a debtor company’s existing privacy policy.
Section 363(b)(1) provides that if a company has disclosed a policy to its customers that prohibits the transfer of customer data and the policy is in effect on the date of filing bankruptcy, then the company may not sell customer data unless the sale is consistent with the policy or a bankruptcy court approves the sale of customer data after the appointment of a consumer privacy ombudsman and notice and hearing. Spirit’s privacy policy did prohibit such transfers. The customer data was carved out accordingly.
Employee data, however, occupies a different legal category. Rep. Steven Horsford and Sen. Elizabeth Warren led a letter signed by 121 members of Congress on October 8, warning that removing names and email addresses “does not necessarily make a dataset anonymous.” The lawmakers wrote to Spirit CEO Dave Davis and Google CEO Sundar Pichai. They said the sale could include about 100 million emails, 500 million Teams messages, employee records, timecard records, payroll and tax information and employment contracts.
The sale agreement lets Google pick or approve the firm that scrubs the data, requires that firm to preserve “referential integrity across the data set,” and bars Google only from intentionally linking the data back to people. The phrase “referential integrity” is important. It means that relationships between records must be preserved even as identifying details are removed. A payroll record may lose an employee’s name, but it must retain the connection to that employee’s timecard, email threads, and project history. Otherwise the dataset loses most of its value for training models on how work is coordinated.
The gap between technical de-identification and practical anonymity is the core concern. Workforce activity generates patterns. Those patterns can be re-identified if combined with auxiliary information, particularly in a dataset as large and interconnected as Spirit’s. Congress has not yet imposed constraints on the sale of employee data in bankruptcy comparable to those governing customer data, but the October letter signals legislative attention.
What This Reveals About Email as an Architectural Problem
Email was never designed to be monetized in bankruptcy. It was designed for communication between identified parties. The structure of email – centralized, server-stored, persistent, and tied to organizational domains – makes it an asset in ways that end-to-end encrypted messaging is not. When Spirit’s operations ceased, its email remained. When its employees dispersed, their correspondence did not follow them. The architecture created a residual asset.
The same is true of Microsoft Teams, Slack, and most enterprise collaboration platforms. They are built for organizational control, not individual custody. Employees do not own the messages they send from a corporate account. The organization does. When the organization enters bankruptcy, those messages become part of the estate. Google’s bid simply makes that dynamic visible and assigns it a dollar figure.
For individuals who prioritize operational security over convenience, the implications are straightforward. Centralized email infrastructure creates long-lived records that can be transferred, subpoenaed, breached, or sold. Minimizing the sensitivity of what is recorded in those systems reduces exposure. For communications that require stronger assurances, tools that do not retain server-side plaintext – such as PGP-encrypted email or systems built on forward-secret protocols – shift the risk surface.
Open-source post-quantum infrastructure – such as PQCServer, released under AGPL-3.0 – provides primitives for building systems that do not rely on long-term key retention or centralized message stores. Onion Mail, which routes traffic over Tor and offers optional PGP integration for users who manage their own keys, represents one approach to separating correspondence from institutional custody. These are not solutions to the bankruptcy-data problem specifically. They are architectural alternatives that treat message custody as a user responsibility rather than a service feature.
Trajectory
Google’s $10 million bid is an isolated transaction, but it establishes precedent. Bankruptcy courts will see similar motions. Competing AI firms will submit rival bids. Employees whose communications are included in such sales will have limited recourse under current law. The gap between customer data protections and employee data protections will remain until legislatures close it.
The market for enterprise failure as a training asset is now priced. That pricing will shape the next generation of data-retention policies, employment contracts, and corporate communication practices. Organizations that understand their internal correspondence as a future asset – or liability – will adjust accordingly. Those that do not will discover the adjustment has already been made for them, in bankruptcy court, after the fact.