Glostarep

Redpine Raises €6.8M to Unlock AI Agents’ Access to Non-Public Data

Redpine Raises €6.8M to Unlock AI Agents’ Access to Non-Public Data

Stockholm-based startup Redpine has announced a €6.8 million seed round to solve one of AI’s most pressing blind spots giving AI agents access to non-public data that sits beyond the reach of the open internet.

The round was led by NordicNinja, with participation from fellow Nordic investors Luminar Ventures and Node.vc. Redpine, founded in 2024 by former Antler partner Anders Hammarbäck and early Spotify employee David Österdahl.

The problem the company is tackling is stark. In accord to CEO Hammarbäck, the data available on the public internet represents only about 1% of all existing data. The rest lives in archives, scientific databases, clinical guidelines, financial records, legal case files, and institutional repositories largely inaccessible to AI systems.

Streaming Data Like Spotify Streams Music

Redpine has built an API interface that connects AI companies and agents directly to premium, rights-cleared data. Its model operates like a data equivalent of Spotify licensing content from rights holders and sharing revenue with them based on consumption. The startup charges per token, per word, and only for what actually used.

Österdahl, who helped build Spotify’s technical streaming platform, sees a direct parallel. The company is already working with leading international AI labs, biotech research firm AsedaSciences, and has angel investors from OpenAI and Perplexity on board.

For Hammarbäck, giving AI agents access to non-public data is not just a business opportunity it’s about preserving the value of human-created content in an era where AI-generated material is flooding the web.

Redpine plans to use the fresh funding for global expansion and further platform development, positioning itself as a critical infrastructure layer for the next generation of agentic AI systems.

Leave a Comment

Your email address will not be published. Required fields are marked *