TypeSafe AI has secured $870 million just weeks after launching Jev, an AI model designed to make decisions rather than hold conversations.
The San Francisco-based startup announced its Series A funding round led by Andreessen Horowitz, with Sequoia Capital, existing investor DCVC, and several angel investors also participating. The financing values TypeSafe at $7.5 billion, and Andreessen Horowitz general partner Martin Casado will join its board.
The raise follows Jev’s Sept. 15 launch and a viral promotional video that attracted interest from developers, executives and investors, according to Bloomberg. TypeSafe says the model surpassed 1 million users within days of its release. The company also claims that about one-third of Fortune 500 companies are using Jev, although it has not publicly identified those customers.
Jev takes a different approach to AI
Unlike ChatGPT and Claude, Jev is designed to make decisions rather than generate conversational responses, emails or code. Built on a transformer architecture, the model returns probabilities, which TypeSafe calls “calibrated decisions.”
That makes Jev a potential fit for software that needs to classify information, route customer requests or determine the next step in an automated workflow. Instead of generating paragraphs before producing an answer, the model is designed to return structured decisions directly.
General-purpose language models can perform similar tasks, but TypeSafe argues that Jev’s specialized design can reduce processing time and token consumption by producing structured decisions directly instead of generating longer text responses.
The startup was founded in 2024 by Diogo Almeida, a former OpenAI researcher, alongside former Meta research engineer Sasha Sheng and engineer and entrepreneur Erik Gafni.
Where the new funding will go
TypeSafe plans to use the money to expand its workforce, secure additional computing capacity and scale product deployments. The company says it already processes trillions of tokens daily.
Almeida told Bloomberg that TypeSafe is profitable after expenses, although the company declined to disclose its revenue. The startup also plans to introduce more models designed for machine use and add enterprise features requested by customers, according to its funding announcement.
What Jev could mean for enterprises and channel partners
Jev enters a market where businesses are seeking more efficient ways to automate routine decisions. For enterprises processing large volumes of customer requests or transactions, a specialized model could reduce costs and processing time if it delivers comparable accuracy to general-purpose language models.
The approach could also create opportunities for managed service providers and systems integrators building AI-powered applications. Rather than replacing existing AI platforms, Jev could complement them by handling specific tasks such as customer support routing, fraud detection, and automated classification.
However, TypeSafe’s efficiency claims require independent validation. Actual performance will depend on the workload, data quality, and deployment costs. The company’s reported Fortune 500 adoption also lacks detail about how extensively Jev is being used in production.
TypeSafe’s substantial funding gives it room to expand, but its long-term success will depend on whether enterprises and technology partners can demonstrate reliable performance and measurable savings over existing alternatives.
Other news: TD SYNNEX has expanded its PartnerFirst platform with AI agents and a Model Context Protocol (MCP) connector, allowing resellers and MSPs to automate tasks such as quoting, order tracking, and subscription management.