TypeSafe AI has exited stealth with $40 million in funding to develop a composable AI model architecture.
SAN FRANCISCO, Sept. 16, 2026 — TypeSafe AI, a frontier artificial intelligence lab building machine-native, composable AI, has emerged from stealth with $40 million in seed funding led by DCVC. Founded in 2024 by former OpenAI researcher and co-inventor of RLHF/ChatGPT Diogo Almeida, alongside Erik Gafni and Sasha Sheng, the company is developing a new class of intelligence designed to give developers reliable, efficient AI they can integrate directly into software systems.
While today’s frontier models have grown increasingly capable at reasoning and conversing with humans, the traits that power a strong AI assistant do not always translate to production environments. These models can hallucinate, shift methodologies between requests, and introduce unpredictable variability into systems that require consistent behavior. TypeSafe is taking a different approach, engineering intelligence that functions as a composable software primitive for applications demanding semantic judgment and intelligent decision-making.
“TypeSafe was founded to pursue an alternative path for AI research, focused on machine-native AI,” said Diogo Almeida, co-founder and CEO of TypeSafe AI. “I spent years working on models designed to make AI better at interacting with people. But if AI is going to fundamentally change how work gets done, people can’t be the only consumers of intelligence. Most intelligence should eventually live inside software, running quietly in the background. TypeSafe is building for that future.”
Jev, named in reference to Jevons Paradox, serves as TypeSafe’s inaugural model built around this philosophy. It delivers frontier-level intelligence with under 100 milliseconds of latency, operating up to 100 times faster and at a fraction of the cost compared to other frontier models. Currently available in early access for select developers, Jev can process hundreds of outputs in parallel from a single prompt and provides calibrated confidence scores, enabling developers to determine precisely when software should act autonomously versus when it should defer.
“TypeSafe is approaching one of the biggest remaining challenges in AI: turning increasingly capable models into technology that developers can reliably build into products at scale,” said James Hardiman, General Partner at DCVC. “Diogo, Erik, and Sasha bring the technical depth and conviction to rethink how models are built for software. Their approach to composable intelligence will unlock an entirely new generation of applications.”
TypeSafe envisions a future of automation where deterministic software is augmented by AI wherever semantic judgment is required. By making intelligent decisions fast, affordable, and reliable enough to test, constrain, and layer upon, the company is positioning intelligence as a standard building block for software development. This foundation aims to empower developers to construct increasingly sophisticated, intelligent applications that no single entity could build through top-down design alone.
About TypeSafe AI
TypeSafe AI is a frontier AI lab building machine-native, composable AI: intelligence designed to operate as a primitive within software. Its models provide developers with fast, predictable, and efficient capabilities for applications requiring semantic judgment and intelligent decision-making. Founded in 2024 and headquartered in San Francisco, TypeSafe has raised approximately $40 million in funding led by DCVC. Early access to its first frontier model, Jev, is currently waitlisted at typesafe.ai.
Source: TypeSafe AI