TypeSafe AI announced a financing round that brought in $870 million, pushing the company's valuation to $7.5 billion. The round was led by Andreessen Horowitz and included Sequoia and existing backer DCVC. The capital influx follows the September 15 launch in the United States of the company’s new AI system, Jev.
The startup says that roughly one-third of Fortune 500 firms have already integrated Jev into their operations, a pace that the company describes as unusually fast for enterprise-level AI adoption. This claim suggests that large corporations are moving quickly to replace traditional text-based models with the new approach.
Jev runs on a transformer-based design but does not function as a large language model. Instead of generating text, it emits probability scores that TypeSafe markets as “calibrated decisions.” The firm argues that this output format enables substantially higher processing speed and consumes far fewer tokens than comparable language models.
The company positions Jev as a tool built for automation rather than for creating prose or code. By focusing on decision-making outputs, TypeSafe claims the model can be embedded directly into workflow engines, reducing the need for intermediate text handling layers that are typical of conventional LLM deployments.
Co-founder Diogo Almeida, a former OpenAI researcher, told TechCrunch that “we have been super good at human language for four years, but it’s not useful for automation because computers speak a different language.” The statement underscores the team’s belief that traditional language proficiency does not translate into effective machine-level automation.
TypeSafe AI was founded in 2024 by a trio that includes Almeida, Sasha Sheng,a former Meta research engineer,and Erik Gafni, an engineer and entrepreneur. The founders assembled the company to explore non-textual AI capabilities, aiming to create models that communicate in the format computers naturally understand.