TypeSafe AI raised $870 million at a $7.5 billion valuation on 9 October, weeks after launching its Jev model, TechCrunch reported.
Key points
- Andreessen Horowitz led the round, with Sequoia Capital and existing investor DCVC participating.
- TypeSafe AI says a third of Fortune 500 companies are using Jev.
- Jev produces probabilities rather than text.
Andreessen Horowitz leads the Series A
Andreessen Horowitz led TypeSafe AI’s Series A, joined by Sequoia Capital and DCVC, an existing investor. TypeSafe AI left stealth on 15 September with $40 million in seed funding from a round led by DCVC, pricing the company at about $200 million post-money, Crypto Briefing reported.
The Series A followed the seed announcement by less than four weeks. TypeSafe AI launched Jev on 15 September. Crypto Briefing reported that the model passed one million users within days of launch. That figure counts users; TypeSafe AI’s separate statement that a third of Fortune 500 companies use Jev is a company claim reported by TechCrunch.
TypeSafe AI was co-founded in 2024 by Diogo Almeida, formerly a researcher at OpenAI, former Meta research engineer Sasha Sheng and engineer and entrepreneur Erik Gafni. Almeida said in September that software automation calls for a different output from the human language produced by large language models.
Almeida is chief executive, Gafni is chief technology officer and Sheng is chief operating officer. There is also reported early investor interest in a potential further round of more than $1 billion at a valuation above $10 billion. That prospective financing is separate from the Series A announced on 9 October.
Jev returns probabilities instead of text
Jev uses a transformer architecture but is not a large language model. Its outputs are probabilities — described by TypeSafe AI as “calibrated decisions” — rather than generated text. The company positions Jev for automating tasks instead of writing prose or code.
Crypto Briefing described Jev’s inputs as typed questions, with a predictable output format intended for software. In that account, an application can consume the model’s response as a decision without first parsing a prose answer. Its reported output latency is under 100 milliseconds. TypeSafe AI also claims Jev runs significantly faster and uses far fewer tokens than large language models, TechCrunch reported.
TypeSafe AI’s claim about Fortune 500 use places large companies among Jev’s users. The reported user total, the company’s enterprise-use claim and its speed claims describe different aspects of adoption and performance: people reached after launch, organisations using the model and the time taken to produce an output.
Axiom’s $2 billion financing and rival products
TypeSafe AI’s round followed a $2 billion financing for Axiom Solutions International in a list of large startup deals, Traders Union reported. General Catalyst and Koch Equity were purchasing shares in Axiom Solutions International, a computing and electricity services company, from Flex at a starting enterprise valuation of $37.5 billion. Axiom’s figure is an enterprise value, while TypeSafe AI’s $7.5 billion figure is the reported valuation of its funding round.
Other companies have released products for structured decisions. AI Affairs previously reported that OpenAI unveiled a Decisions API for preset answers and confidence scores and that AWS released Strands Decider 2B for local decisions in AI agents.
AI Affairs also reported that Databricks launched ai_decide for structured decisions on governed data. Cloudflare released Clef decision models alongside a fine-tuning product.
