The sale of a portion of Spirit Airlines' data assets to Google, developed as part of the airline's bankruptcy proceedings, is raising an issue that goes beyond this individual case: when a bankrupt company sells off its digital archives, who can truly claim to own every element?
Google has confirmed that it purchased part of an enterprise dataset from Spirit, maintaining that it could help improve products and artificial intelligence models. The company also clarified that the package will not include personal information. However, some of the airline's vendors dispute that the definition of the data being sold is sufficiently precise, fearing that third-party software, know-how, technical data, and trade secrets could be transferred to Google alongside Spirit's operational information.
At the center of the dispute is Springshot, a startup founded by Doug Kreuzkamp in 2011. For the past three years, right up until Spirit's final flight, the airline had been using its proprietary platform to coordinate airport activities, logistical processes, and operations necessary to keep services running smoothly. Springshot's technology is deployed across hundreds of airports worldwide and is used by both human operators and automated systems to manage inefficiencies and unforeseen issues.
According to Springshot, no one notified the company prior to the auction's preparation, and the sale agreement fails to adequately distinguish between data actually owned by Spirit and data that, while stored in its repositories or systems, incorporates the vendor's intellectual property. The limited objection filed in court last month calls for the sale to be put on hold until a transparent forensic analysis clarifies the scope of the material destined for Google.
The core issue: operational data or intellectual property?
In a large enterprise's IT systems, the separation is rarely clear-cut. A database may hold records generated by the client, yet be structured around a vendor's proprietary logic. A cloud platform can host the user company's information, while workflows, operational models, configurations, technical documentation, and software components remain subject to licenses, confidentiality clauses, or third-party rights.
It is precisely this overlap that alarms Springshot. In the contested documentation, the dataset categories reportedly include very broad phrasing, such as productivity and collaboration data, business systems and enterprise application data, as well as data related to workflows and processes. For the startup, these descriptions make it impossible to understand what information has been isolated, what may remain covered by contractual agreements, and what checks were performed prior to the transfer.
The position expressed by Kreuzkamp is that the physical availability of data and software on Spirit’s systems is not enough to transfer ownership within bankruptcy proceedings. The risk flagged is that a court-approved acquisition effectively turns into a transfer of trade secrets to a buyer that had no direct relationship with the original vendors and could use that information to develop competing tools.
In Springshot’s case, the concern is particularly sensitive because Google already operates in the market for enterprise tools based on Gemini. In August, shortly before the Spirit auction, Ryanair announced a five-year partnership with Google to share operational data intended to improve Gemini Enterprise. Springshot argues that using data and processes developed throughout its relationship with Spirit could expose it to direct competitive risk, even if Google has not stated any intention to build products that compete with the startup.
The objections do not come from a single vendor
Springshot’s challenge is not an isolated one. International Aero Engines LLC and IAE International Aero Engines AG filed a separate objection citing similar concerns. The companies state that the dataset could include their proprietary commercial information, technical data, and financial materials, all subject to confidentiality obligations under their agreements with Spirit.
For these vendors, the sale could cause damage that is difficult to repair: first, through the transfer to Google of information they never authorized to share; second, through the possibility that the data could feed into further operations or become accessible to third parties. The issue is therefore not merely determining whether Google can receive an archive from Spirit, but defining whether and how bankruptcy proceedings can override contractual restrictions established to protect proprietary industrial knowledge.
When asked about the objections, Google did not address the substance of the claims. It reiterated that the acquisition involves part of an enterprise dataset and that the material can help improve its products and AI models. The commitment not to receive personal data addresses a critical aspect of privacy protection, but it does not resolve the issue raised by the vendors: commercial confidentiality and intellectual property are governed by different rules than personal data.
A significant precedent for the data economy
Insolvency proceedings tend to maximize the value of assets available to creditors and the bankruptcy estate. In a modern enterprise, digital archives can be among the most coveted assets: they assemble years of operational data, edge cases managed in the field, configurations, process histories, and relationships between events that prove invaluable for training or refining artificial intelligence systems.
However, this wealth of information may comprise materials subject to different proprietary regimes. An airline may own its operational metrics, flight-related data, and internal activity logs; that does not automatically imply it can transfer, without verification, every component embedded in the infrastructure used to process them. This distinction is crucial, particularly for startups and specialized vendors, which often deliver products through platforms integrated into client systems and derive their value precisely from methods, interfaces, configurations, and processes that are difficult to separate after the fact.
This is the origin of the request for a forensic examination prior to the actual transfer of the data. Such an audit would be expected to identify the content belonging to Spirit, exclude or restrict access to material subject to third-party rights, and verify compliance with confidentiality clauses. Springshot, however, fears that the absence of a clearly defined scope will turn the bankruptcy into a shortcut to acquire strategic data without the standard negotiations, licensing, and authorizations required on the market.
The phrasing used by the startup—bankruptcy cannot become a new land grab for AI data—sums up a conflict destined to recur. Companies acquiring assets from bankruptcy proceedings view datasets as an increasingly valuable resource; technology providers, for their part, need certainty that a client's insolvency will not cause the boundaries of their intellectual property to evaporate.
For now, the tangible issue remains the petition to halt or condition the sale on a more rigorous review. The Spirit case could clarify how detailed transfers of corporate data must be in bankruptcies and what protections are due to the parties who helped build those systems. At stake is not merely the fate of an operational archive, but the relationship between the data market, contract law, and the development of artificial intelligence products.
