Proximal, a New York City-based AI coding data startup, has emerged from stealth with $15 million in funding at a $300 million valuation.
The funding round was led by General Catalyst, with participation from SV Angel, Go Global Ventures, and Chemistry. Several technology executives and investors also joined the round, including Liam Fedus, CEO of Periodic Labs; Kevin Weil, former Chief Product Officer of OpenAI; and Erik Bernhardsson, CEO of Modal.
Proximal plans to use the new capital to expand its engineering and research teams and scale its post-training and data research infrastructure.
The company also plans to accelerate expansion into new industry areas, including computational drug discovery, custom semiconductor design, and legacy software refactoring for critical public infrastructure.
Founded in 2025, Proximal is developing an AI infrastructure platform focused on automating the post-training feedback loop across complex operational domains.
The platform works with raw, unstructured real-world data, including AI agent execution traces and human workflow artifacts. It converts this information into evaluation environments designed to test AI systems in specific operational scenarios.
Proximal’s technology is designed to identify specific model reasoning failures and use those signals to generate targeted post-training datasets. These datasets can then be used to improve model performance and optimize deployment parameters.
With the new funding and $300 million valuation, Proximal plans to broaden its technology beyond AI coding into additional technical and industrial applications.
The company’s planned expansion includes areas where AI systems need to work with complex data, specialized workflows, and domain-specific requirements.
