Proximal, a San Francisco-based AI data infrastructure startup, has emerged from stealth and disclosed a $15 million seed round led by General Catalyst at a $300 million valuation. The financing also included SV Angel, Chemistry and Diede van Lamoen of Go Global Ventures, alongside individual investors including Liam Fedus, Kevin Weil and Erik Bernhardsson.
The seed round was completed toward the end of 2025, before Proximal publicly announced the financing on September 29, 2026. General Catalyst has also listed Proximal among its portfolio companies and described itself as the lead investor in the $15 million seed round.
Started in 2025 by Calvin Chen, Justus Mattern and Navid Pour, Proximal is building infrastructure for creating training and post-training data for advanced AI systems. The company converts unstructured data, including AI agent execution traces and artifacts from human workflows, into evaluations that can identify capability gaps and generate targeted training data.
Proximal initially focused on software engineering, working with AI labs and enterprises on data for training coding agents. The company said it has since reached $200 million in annualized revenue. The Information reported that Proximal calculates the figure by annualizing revenue generated during the current quarter, rather than reporting $200 million in completed trailing-12-month revenue.
The company is now expanding its research into additional technical domains, including computational drug discovery, specialized semiconductor design and the modernization of legacy software used in critical infrastructure.
