Science
Environments built around real scientific work (literature synthesis, experiment design, data analysis, wet-lab protocol planning) usually authored and verified by PhD-level domain experts. As labs chase 'AI scientist' capabilities, expert-written science environments have become some of the most expensive data in the market.
4 startups
| Name | Domains | Location | Team | Funding | Raising |
|---|---|---|---|---|---|
| Latch Biology data infrastructure turned life-science environments and benchmarks | Science | San Francisco | 11-25 | Series A, $15M (2022) | - |
| ReasonCore Science and code reasoning environments and benchmarks | Science, Coding | San Francisco | 11-25 | - | - |
| Sepal AI Science-domain environments; acquired by Mercor | Science | San Francisco | 1-10 | Acquired by Mercor (Feb 2026) | No |
| Tacit Labs Life-science and long-horizon environments for AI models | Science, Long Horizon | San Francisco | 1-10 | - | - |