TypeSafe AI Hits $7.5B Valuation 24 Days After Launching Non-Text Model Jev

Quick Facts

• TypeSafe AI closed an $870 million Series A at a $7.5 billion valuation, led by Andreessen Horowitz, just 24 days after its $200 million seed round.
• Jev is a discriminative model that returns typed values and confidence scores, not generated text, and processes requests in under 700 milliseconds.
• About one-third of the Fortune 500 has already adopted Jev, and OpenAI launched a competing product called Decisions API three weeks after Jev went live.

TypeSafe AI has raised $870 million in a Series A round at a $7.5 billion valuation, less than a month after the San Francisco company launched its first AI model. The round was led by Andreessen Horowitz, with Sequoia Capital, existing backer DCVC, and a group of angel investors participating.

The speed of the valuation jump is without precedent in the AI sector. TypeSafe's seed round, led by DCVC, valued the company at $200 million. Twenty-four days later, the Series A closed at $7.5 billion, a nearly 37-fold increase.

A Different Kind of AI Model

Jev does not generate text. TypeSafe calls it a System One model, a decision engine that takes program state and a set of typed questions and returns structured answers with calibrated probabilities. It outputs three primitives: Choice, Score, and Noul, each paired with a confidence value.

The model was trained using what TypeSafe calls reinforcement learning for calibrated decisions, a method developed in opposition to the reinforcement learning from human feedback technique that underlies most large language models. TypeSafe trained Jev entirely on synthetic data over two years in stealth before its September 15 launch.

TypeSafe prices Jev at $0.042 per million input tokens, with output tokens free. The company claims response times of 70 to 500 milliseconds end-to-end, versus 3 to 329 seconds for comparable LLM calls, making it up to 200 times faster and 100 times more cost-efficient than frontier models.

The Founders' Backgrounds

CEO Diogo Almeida spent a decade in AI research at OpenAI and Google Brain. He is credited as a co-inventor of InstructGPT, ChatGPT, GPT-4, and the reinforcement learning from human feedback technique that made large language models commercially viable.

Almeida told Fortune he began thinking about Jev after helping build the technology behind ChatGPT. He described a gap between models' ability to answer human questions and their usefulness in automating routine work, and said he wanted to address AI's "massive over-promise under-deliver issue."

TypeSafe's other co-founders are Erik Gafni, CTO, and Sasha Sheng, COO. The company was founded in 2024.

Rapid Enterprise Adoption

Jev reached 1 trillion tokens processed within three days of launch, according to Andreessen Horowitz's investment memo. The firm called it the fastest-growing model it has seen.

Within 24 hours of appearing on Vercel's AI Gateway, nearly 13 percent of Vercel's paid teams had tried Jev. Vercel reported that Jev reached more than twice as many paid teams in its first day as any previous model on the platform. Cloudflare, LangChain, and Langfuse also integrated Jev within three days of launch.

Vercel engineers replaced an OpenAI classifier with Jev on a command-safety check and reported results five to 18 times faster with greater accuracy. About one-third of the Fortune 500 has already adopted the model.

Competitors Respond Quickly

Three weeks after Jev launched, OpenAI released a product called Decisions API, built on its existing GPT-6 Luna model. Amazon Web Services' Strand Labs also released its own decision model in response to demand from AI developers seeking tools built for computer automation rather than human conversation.

The competitive response came fast because Jev targets a specific bottleneck in enterprise AI deployment: the cost and latency of using generative models for classification tasks where only a structured decision is needed. TypeSafe built Jev specifically for that workload. OpenAI's Decisions API repurposes an existing generative model for the same job.

TypeSafe's path from stealth to multibillion-dollar company in under a month reflects a market that is actively searching for AI that works with code rather than with people.

Read more: The maker of non-text AI model Jev valued at $7.5B just weeks after launch