TypeSafe AI Raises $870M in Funding Round

The San Francisco startup, led by co-founder and CEO Diogo Almeida, is backed by Andreessen Horowitz and reportedly valued at $7.5 billion.

Updated on Nov. 1, 2026 in Artificial Intelligence

TypeSafe AI Raises $870M in Funding Round

Live Poll

Would you trust an AI tool to make routine decisions for you?

TypeSafe AI, a San Francisco-based artificial intelligence startup, raised approximately $870 million in a funding round reported by Bloomberg in October 2026. The round was led by Andreessen Horowitz, with participation from Sequoia Capital.

Why it matters

The funding supports TypeSafe AI's push into structured AI outputs for software developers, a segment of the broader AI market focused on usable machine-consumable results. The reported $7.5 billion valuation places the company among well-capitalized players in the Silicon Valley AI sector.

The funding round raised approximately $870 million at a reported $7.5 billion valuation. The extraction does not provide technical performance benchmarks for the Jev model, and the valuation has not been independently confirmed.

The players

Andreessen Horowitz

A prominent Silicon Valley venture capital firm that led TypeSafe AI's funding round.

Sequoia Capital

A leading venture capital firm that participated in TypeSafe AI's funding round.

Diogo Almeida

Co-founder and CEO of TypeSafe AI, and a listed co-author of the InstructGPT research paper, who also represented the Philippines at the International Mathematical Olympiad.

The details

TypeSafe AI launched Jev, a new AI model that returns structured outputs such as classifications, rankings, scores, choices and probabilities. The model is designed to be used directly within software applications and alongside other AI models rather than as a conventional chatbot.

Weeks before October 2026: Jev launched shortly before the funding round.
October 2026: Bloomberg reported the $870 million funding round.

Roadmap

TypeSafe AI's Jev extends a line of research on making language models follow instructions reliably, a lineage tied to the InstructGPT paper on which co-founder Diogo Almeida was a listed co-author. The push toward machine-consumable structured outputs reflects a broader industry shift from open-ended chat interfaces toward AI components embedded in software systems.

For software developers, Jev's structured outputs could mean AI decisions that plug directly into application logic without extra parsing. The practical effect on pricing, access, or adoption has not been detailed in the source extraction.

The takeaway

Developers evaluating AI tools may find value in models built to return structured, machine-usable outputs rather than free-form text. Watching how such products are priced and adopted will show whether this approach gains traction.