The San Francisco startup wants to make AI outputs provable, not just probable, starting with law, healthcare, and tax compliance.
Pramaana Labs, a San Francisco-based AI startup, has raised a $27M seed round led by Khosla Ventures to build what it calls a verification layer for artificial intelligence. The company’s thesis is deceptively simple: in industries where being wrong carries real consequences, AI shouldn’t just guess. It should prove its work.
The round also includes backing from BoldCap and Founders Future.
What Pramaana actually does #
Pramaana Labs is building technology that translates complex domain knowledge, think tax codes, clinical guidelines, legal statutes, and safety constraints, into formally verifiable representations. In English: instead of an AI saying “this tax deduction is probably valid,” Pramaana’s layer would let it show its reasoning chain in a way that can be machine-checked against actual rules.
The company was co-founded by Ranjan Rajagopalan, Krishnan Raghavan, and Sanjay Ganapathy Subramaniam. Their target verticals are deliberately narrow and deliberately high-stakes: statutory tax reasoning, legal compliance, healthcare safety, and autonomous systems.
The approach borrows from a discipline called formal verification, which has deep roots in hardware design and aerospace engineering. Chip manufacturers have used formal methods for decades to prove that processors won’t produce incorrect calculations.
The verification gap in AI #
Pramaana held its inaugural Verification Summit on June 10, 2026, in San Francisco, headlined by Vinod Khosla himself.
What this means for investors #
A $27M seed round is eye-catching, but it’s worth putting in context. Seed rounds in AI have ballooned considerably, with frontier model companies raising hundreds of millions at formation. Pramaana’s raise is more modest, reflecting its infrastructure-layer positioning rather than a compute-hungry model training play.
The crypto-adjacent angle is worth noting, even though Pramaana itself has no token or blockchain component. The concept of verifiable computation has been a core research area in blockchain for years, from zero-knowledge proofs to verifiable delay functions. Pramaana’s formal verification approach shares philosophical DNA with these efforts: the idea that trust should be mathematical, not institutional.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our