Redpine's Data Licensing Play Has a Familiar Shape: Can Spotify's Model Work for Enterprise Data? The Spotify model is coming for enterprise data — and AI agents are the new listeners. Redpine, a Stockholm startup founded in 2024, has raised a €6.8 million seed round to build…
The Spotify model is coming for enterprise data — and AI agents are the new listeners.
Redpine, a Stockholm startup founded in 2024, has raised a €6.8 million seed round to build what its founders describe as Spotify for proprietary data. The comparison is apt, even if it's also convenient: the team knows something about building licensed-access platforms at scale. CEO Anders Hammarbäck previously worked at Antler and McKinsey; CTPO David Österdahl held engineering roles at Spotify and payments company Zettle. Their new bet is that the same licensing logic that displaced music piracy can be applied to the vast private data reserves enterprises have spent decades accumulating.
The mechanism is a headless API that lets AI agents query licensed sources and pay per token, with a revenue share flowing back to whoever owns the data. Redpine claims access to more than 100 billion tokens sourced from scientific journals, legal documents, financial filings, and clinical records. Data providers get a monetization path; AI agents get structured access to the 99% of information that never made it onto the open web.
The framing matters more than the number. "The internet is maybe 1% of the total data that's out there," Hammarbäck told Sifted. That observation is not new, but Redpine is wagering that AI agents make it actionable in a way previous waves of "enterprise data lake" tooling never managed. Rather than building another RAG pipeline or search overlay, they're constructing an economic layer: tokenize the source, meter the queries, split the revenue.
This puts Redpine adjacent to, but distinct from, the annotation and training-data businesses that have defined the data-for-AI industry so far. Scale AI, Appen, and Defined.ai are annotation-first platforms, built to label data for model training. Redpine is API-native and agent-first, designed for inference-time access rather than dataset construction, as The Next Web noted. The angels backing the round — from OpenAI and Perplexity — suggest the company is being evaluated as infrastructure for the next phase of AI development rather than another labeling play.
The Anthropic training-data settlement in early 2026 is almost certainly in the background of every enterprise data conversation happening right now. Rights holders have spent years watching their content scraped without compensation. Redpine offers them something different: a metered revenue stream, controlled access, and a place in the AI workflow rather than a lawsuit afterward.
Whether data owners want to be in the licensing business is the open question. Redpine is working with AI labs and the biotech firm AsedaSciences, which suggests the early customers are companies with proprietary datasets that need AI-native integration more than they need to protect a moat. Clinical trial data and scientific literature are obvious candidates: high value, structured, and currently inaccessible to agents without custom integration work.
The competitive implication is what makes this worth watching. If per-token licensing becomes the standard access pattern for AI agents querying proprietary sources, then the assumption that data ownership is a durable competitive advantage gets stress-tested. Companies that have spent years building proprietary datasets may find themselves with two choices: monetize through a licensing layer, or risk becoming irrelevant to AI-native workflows that simply route around them. Redpine is betting that most will choose the former. That bet is what the seed round is funding.
The 100 billion token figure is a marketing number, and the Spotify analogy is a founder talking point. The real substance is the economic layer Redpine is attempting to construct: a structured, metered, revenue-sharing interface between AI agents and the private data that still powers most of the global economy. That layer does not yet exist at scale. Redpine is building the plumbing and hoping the customers come.