Athletes Are Training Data and No One Is Paying Them for It
There’s a paradox that has been quietly operating within the sports entertainment industry for years: the most valuable assets of the business—athletes themselves—are the only participants in the process who don’t receive compensation when their data fuels artificial intelligence (AI) models. Each sprint of Kylian Mbappé, every inflection of LeBron James's voice, every biometric pattern captured during a Champions League final, enters the training datasets of AI systems without contracts, without royalties, and, in most cases, without the athlete's knowledge.
On March 16, 2026, a newly formed company called Callandor Group announced what it describes as the first registry dedicated to sports intellectual property in the AI era. The proposal is straightforward: to create an infrastructure that maps the ownership of videos, voices, performance data, and biometrics, allowing athletes and sports organizations to license that content transparently, in compliance with regulations, and, most importantly, for compensation. The team behind the project includes executives with backgrounds at Sony, MGM, Amazon Studios, and NASA’s Jet Propulsion Laboratory, along with operational ties to the digital divisions of FC Barcelona.
The strategic question isn’t whether the technology works. It’s whether the market is ready to adopt an infrastructure that, by definition, raises the cost of access to data that is currently obtained for free.
The Market No One Has Wanted to Structure
For decades, broadcasting rights were the financial backbone of professional sports. Leagues sold broadcasting windows, clubs negotiated shares, and athletes received salaries that partly reflected that audience value. This model worked as long as content was linear and consumption was passive.
AI broke this logic without asking for permission. Generative systems require data on movement, voice patterns, performance sequences, and body geometry to produce avatars, automated narratives, tactical simulations, and interactive experiences. All of that material exists in video archives that clubs, leagues, and platforms have amassed for decades. The vast majority of that archive has been used without a licensing framework regulating it.
Callandor CEO Michael Fisk describes it plainly: data from stars like Messi or LeBron James fuels AI models with zero transparency and no consistent royalty framework. This is not an accusation of bad faith; it’s a description of the regulatory void that existed before instruments like the European AI Regulation began imposing traceability obligations on high-risk systems.
What Callandor is building is not, strictly speaking, a product for end consumers. It’s a layer of infrastructure, what in financial terms is called a reference market: a standardized mechanism that allows dispersed and unpriced assets to gain exchange value. Phil McKenzie, a strategic advisor for the project and co-founder of Goldfinch, a platform that deployed over £300 million across more than 300 entertainment projects, compares it to what the financing of cinema credit was fifteen years ago: a massive base asset without infrastructure to monetize it.
The analogy is technically accurate. Before structured digital music rights markets existed, record labels also were using catalogs without artists receiving proportional compensation. What changed was not the technology, but the existence of auditable records.
What the Technical Architecture Reveals About the Business Model
The technical core of the platform is what the company calls the Event Horizon API: an interface that mediates AI queries about sports content, verifies ownership, applies the appropriate licensing contracts, and generates usage traceability to activate royalties. The CTO, who has experience with NASA’s Perseverance rover mission, is the technical figure ensuring that this mediation layer is auditable.
From a business model perspective, the architecture is elegant because it doesn’t compete with anyone in the existing value chain. It doesn’t take revenue away from leagues, doesn’t displace streaming platforms, and doesn’t pit clubs against each other. It positions itself as the settlement mechanism that everyone needs but nobody has built.
The real risk is not technical. It’s adoption. For the registry to function as a market standard, it needs a critical mass of registered assets before data buyers have an incentive to pay for access. To attract that critical mass, it needs clubs and athletes to perceive that registering generates more value than the cost of the process. The company is betting on clubs with the highest global visibility, specifically the digital divisions of FC Barcelona—Barça Media, Barça One, and Barça Digital Assets—as legitimacy anchors that attract the rest.
This is not an arbitrary strategy. The five major European football leagues concentrate global audiences with high regulatory alignment: the European AI Regulation, with full compliance obligations projected for 2027 on high-risk systems, turns data traceability into a legal requirement, not just an ethical preference. Whoever registers their assets before that date won’t be buying an optional service; they’ll be ensuring regulatory compliance.
When the Invisible Asset Becomes a Revenue Stream
Fisk uses a phrase that merits analysis: athletes are the new code; if athletes are the software, we build the app store. Beyond marketing hype, the metaphor accurately describes the revenue model that Callandor is attempting to enable.
An app store doesn’t sell software; it charges for each transaction it facilitates. In that logic, Callandor does not directly monetize sports data. It monetizes the volume of queries that AI systems make about that data. Every time a generative model accesses the movement sequence of a registered player, the system activates a contract, registers the transaction, and distributes royalties. Callandor’s revenue is the intermediation margin on that flow.
This model has an attractive financial property: fixed costs are those of building and maintaining the registry and the API; variable revenues scale with the volume of AI queries, which by definition grows as the generative entertainment industry expands. They do not accumulate physical assets. They do not finance content. They transform a data governance problem into a usage-based revenue structure.
The risk of this model is the speed of standardization. If within the next 24 months a consortium of leagues or tech platforms builds their internal traceability system, Callandor loses the neutral intermediary position that justifies its existence. Therefore, the move towards large European clubs is not purely commercial; it’s a race to establish itself as a reference registry before the market decides to build its own alternatives.
The Capital That’s Not Yet on the Balance Sheet
Training rights and broadcasting rights took decades to become financial assets with market valuation. Training rights, as Callandor calls the use of athlete data to fuel AI models, are at a similar turning point.
What Callandor is building does not change who owns that data. It changes whether that ownership has a price, if that price is auditable, and whether there is a mechanism for it to flow to the one who generated the asset. True leadership in this market doesn’t lie in burning capital competing for the same broadcasting rights that all platforms are already bidding for, but in having the audacity to eliminate the opacity that makes a multibillion-dollar asset have no market price, and build the infrastructure that establishes it.
