TypeSafe exits stealth with $40M seed to build AI for software, not people
What's the deal? TypeSafe AIDealroom has a profile for this one. Try Dealroom →, a San Francisco frontier AI lab, has emerged from stealth with a $40 million seed round led by DCVC. The company is building "machine-native, composable AI" — intelligence meant to run inside software rather than chat with humans.
Who's behind it? Founder and chief executive officer Diogo Almeida is a former OpenAI researcher and co-inventor of RLHF, the technique behind ChatGPT. He is joined by co-founders Erik Gafni and Sasha Sheng.
What's the endgame? Almeida argues that people should not be the only consumers of AI. "Most intelligence should eventually live inside software, running quietly in the background," he said. The company wants intelligence to become a standard building block developers can layer into applications.
The product: TypeSafe's first model, Jev — named after Jevons Paradox — is in early access for select developers. It runs at under 100 milliseconds of latency and is up to 100 times faster and cheaper than other frontier models, the company says. Jev can process hundreds of outputs in parallel from a single prompt and returns calibrated confidence scores, letting software decide when to act autonomously and when to defer.
Why it matters: Today's frontier models hallucinate, shift methods between requests, and add variability that production systems can't tolerate. TypeSafe pitches a fix — intelligence reliable enough to test, constrain, and build on.
By the numbers: The round is exceptionally large for its stage. It ranks in the 99th percentile of all-time seed rounds in AI, based on a sample of 28,473 deals.
"TypeSafe is approaching one of the biggest remaining challenges in AI: turning increasingly capable models into technology that developers can reliably build into products at scale," said James Hardiman, general partner at DCVC.
The signal: Investors are betting big on AI infrastructure that treats intelligence as a software component rather than a conversational assistant. If that thesis holds, the next wave of AI value may come not from chatbots, but from models working silently inside the products people already use.
Read more: morningstar.com
Image credit: Typesafe Ai
