TypeSafe AI raises $40M Seed

DCVC led the round. No other investors were disclosed. Diogo Almeida, a former OpenAI researcher, founded the company with Erik Gafni and Sasha Sheng. Its first model, Jev, is billed as a "System One Model". It outputs typed decisions with calibrated confidence instead of free text. TypeSafe says Jev matches frontier-model judgment on these tasks, with responses in 70 to 500 milliseconds versus 3 to 329 seconds for frontier models. The company has not published a valuation. One secondary outlet put it at $200M, citing Yahoo Finance. Early traction is the main support for the pitch. A Vercel engineer told TechCrunch that swapping an OpenAI model for Jev in a command-safety classifier made it five to 18 times faster, with better accuracy. TypeSafe says demand briefly left it unable to serve users from its API after launch. A secondary blog, citing The Information, reports talks to raise $1B or more at a $10B or higher valuation. That is unconfirmed, and no investors are named. For AI founders, the bet is that many production workloads need a fast, predictable classifier more than a chat model. Teams today often add a human review step to model output as a safeguard. If typed outputs with reliable confidence scores hold up, latency and cost become the competitive ground for agent and automation products. Startups that wrap general-purpose chat models for those tasks could face a cheaper, faster alternative.