TypeSafe AI, Maker of Jev, Raises $870 Million at a $7.5 Billion Valuation

TypeSafe AI, the San Francisco startup behind the Jev model [4], announced a funding round of $870 million at a $7.5 billion valuation on 9 October 2026, led by Andreessen Horowitz (a16z). Bloomberg reported it first; TypeSafe and a16z confirmed it later that day. The round is confirmed. The usage and performance figures around it are the companies’ own claims, which TSN has not checked. This is not investment advice.

What was announced

Bloomberg’s Natasha Mascarenhas and Rebecca Torrence reported at 14:00 BST that TypeSafe “has raised about $870 million at a $7.5 billion valuation in a financing led by Andreessen Horowitz” [1]. That was reported only until the early evening, when TypeSafe published its own blog post, “TypeSafe A raises Series AI” (the title as published). It says the company has “raised a really big series A¹ led by Andreessen Horowitz, with participation from Sequoia Capital, existing investor DCVC, and the tech illuminati (aka a bunch of the best angel investors)”. The footnote reads “$870 million at a $7.5B valuation, with Martin Casado joining the board” [2]. a16z published its own post, “Investing in TypeSafe AI”, the same day [3]. Bloomberg’s report, as relayed by Investing.com, says a16z “wrote the largest check” [4].

A valuation is the price investors and a company agree for a stake, not a measure of sales or profit. Neither TypeSafe nor a16z says whether $7.5 billion includes the new money.

Who TypeSafe is, and what Jev does

TypeSafe AI says it spent two years in stealth before launching Jev on 15 September 2026. Founder Diogo Almeida says he previously worked at OpenAI on methods behind ChatGPT (his own account) [5]. Bloomberg says he started TypeSafe in 2024 [4].

Jev is not a chatbot. TypeSafe calls it a “System One Model”, built “to make fast, structured decisions that software can use directly” [5]. In plain English: a program hands Jev some text or data plus a question with fixed possible answers, such as yes or no, one choice from a list, or a score on a scale. Jev replies with an answer and probabilities the program can use directly, instead of a paragraph it must parse. TypeSafe’s phrase is “unstructured state in, typed probabilistic decisions out” [5].

TSN’s earlier explainer, RAG vs Jev+RAG Explained, shows one use: a “judgment gate” that scores retrieved passages before a chatbot writes an answer. OpenAI and Perplexity launched similar models in early October: OpenAI’s Decisions API and Perplexity’s Open Decider.

The figures are company claims

• Fortune 500. TypeSafe’s blog says “A third of the Fortune 500 are getting their Jev on” [2]. Bloomberg says Jev “is now used by about a third of Fortune 500 companies, according to the startup, which declined to disclose any names” [1]. Almeida’s X post says “29.4% of the fortune 500 showed up” [6]. a16z says “25% of the fortune 500 enterprises have integrated Jev” [3]. The wording differs (“used”, “showed up”, “integrated”), and none of the sources says what counts.
• Users and volume. Bloomberg says the model “surpassed one million users in a matter of days, the startup said” [4]. Almeida says it “accidentally served trillions of tokens a day” [6]. a16z says Jev reached “1 trillion tokens generated in merely 3 days post launch” [3].
• Speed and cost. TypeSafe’s launch post claims 70ms to 500ms responses and its own pricing, then admits “We can’t prove it isn’t subsidized” [5]. a16z, which is now an investor, says Jev is “roughly 1/100 to 1/500 the cost of frontier models while being 100x faster for classification tasks at the comparable accuracy” [3]. The launch post says its workflow tests were made by TypeSafe’s own team, so “some bias could exist” [5].

What is not known

None of the sources TSN read gives TypeSafe’s revenue, profit, customer names, the size of the stake sold, or whether the valuation counts the new money. TypeSafe says only “We’ve saved customers millions of dollars in production already” [2], with no detail. No independent benchmark of Jev was found.

What this does not prove

• That Jev beats other models. The speed, cost and accuracy claims come from TypeSafe and its new lead investor, on tests TypeSafe designed [3][5].
• That a third of the Fortune 500 use Jev in production. No customers are named, and the sources give 25%, 29.4% and “a third” [1][2][3][6].
• That the valuation reflects what the business earns. A valuation is a negotiated price for shares, not a revenue figure.
• That “can’t hallucinate” means “always right”. TypeSafe’s wording is about the answer format: every reply fits the options the program declared [5]. A valid option can still be the wrong one.
• Investment advice. TSN is describing a funding announcement, nothing more.

The Bottom Line

TypeSafe AI has raised $870 million at a $7.5 billion valuation, led by Andreessen Horowitz with Sequoia and DCVC taking part. Bloomberg reported it first and TypeSafe and a16z have since confirmed it. Jev is a model that gives software typed, probability-backed decisions instead of prose, and the company says it is fast, cheap and widely used.

The money is confirmed. The adoption, speed and cost claims are not yet independently checked, and the figures do not agree with each other. The next evidence to watch is independent testing of Jev and named customers, not the size of the round.

Related on TSN: RAG vs Jev+RAG Explained: What the Judgment Gate Adds; OpenAI’s Decisions API and Perplexity’s Open Decider: AI That Answers With a Probability, Not an Essay

Sources

1. Natasha Mascarenhas and Rebecca Torrence, “Andreessen Horowitz Backs Jev Maker at $7.5 Billion Value”, Bloomberg (via Bloomberg Government), 9 October 2026, 14:00 BST (paywalled; TSN read the headline and opening paragraphs; reported). https://news.bgov.com/environment-and-energy/andreessen-horowitz-backs-jev-maker-at-7-5-billion-value
2. TypeSafe AI, “TypeSafe A raises Series AI”, TypeSafe AI blog, 9 October 2026 (company announcement), read 9 October 2026. https://typesafe.ai/blog/series-ai
3. Andreessen Horowitz, “Investing in TypeSafe AI”, a16z newsletter, 9 October 2026 (investor’s own post; published about 17:28 BST), read 9 October 2026. https://www.a16z.news/p/investing-in-typesafe
4. Investing.com, “TypeSafe AI raises $870 million led by Andreessen Horowitz – Bloomberg”, 9 October 2026 (a write-up of the Bloomberg report; source of the “largest check”, “one million users” and “started TypeSafe in 2024” details). https://au.investing.com/news/stock-market-news/typesafe-ai-raises-870-million-led-by-andreessen-horowitz–bloomberg-93CH-4682137
5. Diogo Almeida, “Introducing System One Models & Jev”, TypeSafe AI blog, 15 September 2026 (company launch post), read 9 October 2026. https://typesafe.ai/blog/introducing-system-one-models-and-jev
6. Diogo Almeida (@CompleteSkeptic) on X, 9 October 2026, 17:26 BST (founder’s own post; checked via the X API), and TypeSafe AI (@typesafeai) on X, 9 October 2026, 17:31 BST (company post announcing the round, linking the blog). https://x.com/CompleteSkeptic/status/2108594987177021737 and https://x.com/typesafeai/status/2108596156951642589