- Protege connects organizations holding proprietary data with AI developers for training and evaluation.
- Data types include de-identified clinical records, medical imaging, video and audio.
- Data holders earn a revenue share each time their data is used, per SiliconANGLE.
- It raised a $25 million Series A in August 2025 and $30 million from Andreessen Horowitz in January 2026.
- Type
- Governed data marketplace
- Founded
- 2024
- Data types
- De-identified clinical records, medical imaging, video, audio and more
- Seller model
- Revenue share each time data is used; help with valuation and compliance
- Funding
- $25 million Series A (August 2025); $30 million extension from a16z (January 2026)
What Protege does
Protege connects organizations that hold proprietary data with AI developers who need it for training and evaluation. Its partner network grew to hundreds of organizations through 2025. Data holders get help estimating the value of a dataset, with privacy and intellectual property compliance and with making data AI-ready, and earn revenue share payouts each time their data is used. Developers use a curated marketplace and compliance layer to find, request and combine datasets (SiliconANGLE).
Where it fits
Protege is a strong fit for health data holders: providers, imaging centers and others with de-identified records. Health data needs HIPAA de-identification by Safe Harbor or Expert Determination before licensing. See medical imaging data and de-identification for AI training data.
How it compares with a broker
A revenue share means you earn as the data is used, rather than a fixed fee. A broker typically negotiates a direct license with one buyer, which can mean a larger upfront payment, exclusivity or both. Which is better depends on the data and how many buyers want it.
Things to check before you sign
- How the revenue share is calculated and reported, and audit rights.
- Whether the platform holds a copy of your data, and where.
- Who performs de-identification and who signs the Expert Determination, if used.
- Termination: what happens to copies already delivered to developers.
Alternatives and related companies
Questions
How do data holders get paid on Protege?
SiliconANGLE reported that data holders earn revenue share payouts each time their data is used.