AI verification startup Pramaana Labs has successfully raised $27 million in seed funding, according to reports from Business Standard. The funding round was led by Khosla Ventures and included participation from early backers Pushmeet Kohli, vice-president at Google DeepMind, and Sriram Rajamani, corporate vice-president at Microsoft CoreAI. The company plans to use the capital to train its formalisation and prover models, expand its AI research team, and onboard domain experts across regulated sectors including taxation, healthcare diagnostics, cybersecurity and financial compliance. The round also included backing from BoldCap, Founders Future, Accel, Nexus Venture Partners, Premji Invest, and Unbound, with the company holding its inaugural Verification Summit on June 10, 2026, in San Francisco, headlined by Vinod Khosla himself. As reported by Business Standard, the funding round was primarily led by Khosla Ventures, with BoldCap and Founders Future also providing financial backing.
As reported by Business Standard, Pramaana's platform converts domain-specific knowledge into machine-verifiable outputs by translating rules and regulations such as tax codes, clinical protocols and financial compliance frameworks into formal language that machines can process. User queries are then translated into formal statements and checked through a proof engine before answers are returned. The system either provides machine-checkable proof supporting answers or identifies specific rules preventing valid conclusions, with the platform designed to withhold responses when proof cannot be established. The company's technology borrows from formal verification, a discipline with deep roots in hardware design and aerospace engineering, where chip manufacturers have used formal methods for decades to prove that processors won't produce incorrect calculations. According to the latest reports, this means that rather than general statements like "this tax deduction is likely valid," Pramaana offers a rationale that can be cross-examined against established rules, establishing a verification layer for artificial intelligence. The platform still runs on a conventional LLM, giving it the flexibility to answer natural language questions and tackle complex problems that conventional computers can't handle, but includes a deterministic layer on top ensuring the LLM's work checks out.
According to Business Standard, co-founder and CEO Ranjan Rajagopalan stated that "AI has an accountability gap" and that many world problems are unformalised rather than unsolvable. He emphasized that Pramaana encodes rules into machine-reasonable form to address domains where mistakes can affect health, money or freedom. 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 approach shares philosophical DNA with verifiable computation concepts in blockchain, from zero-knowledge proofs to verifiable delay functions, with the idea that trust should be mathematical, not institutional. As reported by TechCrunch, Rajagopalan describes the rules of domains like tax law as "math in the sense that you have a lot of rules that you need to abide by," noting that once codified, reasoning becomes deterministic. Pramaana asserts that AI technology should demonstrate its decision-making process instead of making assumptions or guesses, establishing a fundamental mission to enhance AI's decision-making accuracy in high-stakes sectors.
As reported by Business Standard, the founding team combines expertise in formal methods and large-scale AI systems. Rajagopalan previously led Google Maps Moderation, while co-founder Krishnan Raghavan worked on Glean Assistant at enterprise search company Glean. Sanjay Ganapathy, formerly a staff research engineer at Google DeepMind, contributed to the Gemini family of AI models. All three founders are alumni of IIT Madras. General partner Sathya Narayanan from BoldCap noted that the team is addressing one of the most important AI challenges — trust — with auto-formalisation infrastructure having potential to become a critical building block for future AI systems. The founding team includes Ranjan Rajagopalan, Krishnan Raghavan, and Sanjay Ganapathy Subramaniam, with all three founders being alumni of IIT Madras.
According to recent reports, Pramaana Labs is building technology that translates complex domain knowledge into formally verifiable representations, targeting deliberately narrow and deliberately high-stakes verticals including statutory tax reasoning, legal compliance, healthcare safety, and autonomous systems. The company's research ecosystem includes collaborations with academics from IIT Delhi, IIT Madras, UC Berkeley and Stanford University's Centaur Lab, with tax formalisation efforts advised by former US Internal Revenue Service Commissioner Danny Werfel and supported by researchers from Yale Law School and Stanford. For each use case, Pramaana will build its own LEAN-style formal verification system, overseen by domain experts. For tax law, the company is working with Werfel, while professors from IIT Delhi, IIT Madras, and UC Berkeley oversee the cybersecurity and drug discovery system. A $27 million seed round is notable in the current AI landscape, reflecting the company's infrastructure-layer positioning rather than a compute-hungry model training play, with the raise being more modest compared to frontier model companies raising hundreds of millions at formation. The company's focus on critical sectors where validation and accuracy are paramount aligns with the study of verifiable computation, long a topic of interest in blockchain discussions, though Pramaana itself does not integrate tokens or blockchain into its structure.
