01
Attendee
Researchers, founders building in regulated domains, applied-AI engineers, and investors with active verification theses. Capacity is intentionally small (~120 in the room). Approval-based.
Apply to attend →
Launch Edition · San Francisco
Where AI Verification gets serious
A half-day summit on the verification problem in AI — moving from outputs that sound right to outputs we can prove right.
- Anchored by Vinod Khosla
- Hosted by Pramaana Labs
- Supported by BoldCap
Listen to leaders from
Mission
Why a Verification Summit
AI has crossed a threshold where its outputs influence decisions in law, science, software, finance, and policy. But the foundation underneath is still probabilistic — a model that's right most of the time, with no way to know which time is which.
Verification is how that changes.
Formal methods, automated theorem proving, runtime checks, alignment-by-construction, provable agents — these are the tools that turn "the model said so" into something an engineer, a regulator, or a scientist can stake a decision on.
The Verification Summit is the first single-track, technical-first gathering for this field.
We're curating it the way the early systems and security conferences were curated: small, sharp, and built for the people doing the work.
Q.E.D.
Speakers
In the room
Confirmed speakers at the Summit.
-
ConfirmedVinod Khosla
Founder · Khosla Ventures
Co-founder of Sun Microsystems and founder of Khosla Ventures. One of Silicon Valley's most consequential investors, with early bets on companies reshaping energy, health, and enterprise AI.
-
Panel ModeratorKanu Gulati
Partner · Khosla Ventures
Partner at Khosla Ventures, investing in AI infrastructure, robotics, and autonomous systems. PhD in ECE from Texas A&M and an MBA from Harvard. 10+ years as a research scientist and founding engineer across Intel, Cadence, and Heavy.ai.
-
Panel ModeratorSathya Nellore Sampat
General Partner · BoldCap
General Partner at BoldCap, a seed-stage AI fund backing outlier Indian founders building the next generation of global AI companies. Currently investing across autonomous systems, frontier labs, and physical AI.
-
ConfirmedKrishnamurthy Dvijotham
Senior Staff Research Scientist · Google DeepMind
Senior Staff Research Scientist at Google DeepMind, working on safe, reliable, and trustworthy AI. Previously founded the AI security research team at ServiceNow Research; his work spans AI agent security testing, formal verification of neural networks, and robustness against adaptive data poisoning.
-
ConfirmedSatnam Singh
Software Engineer · Harmonic AI
Software engineer and researcher in formal hardware verification, ML acceleration, and functional programming — with stops at Harmonic, Groq, Google, and Microsoft. Uses theorem provers (Lean, Agda, Coq) and compiler-optimization techniques to design and verify silicon and secure digital systems.
-
ConfirmedEve Bodnia
Founder & CEO · Logical Intelligence
Building deterministic AI for software and hardware correctness at Logical Intelligence. Quantum physicist and mathematician. UCSB PhD under Nobel laureate. Pioneered Google Quantum AI's brain research grant. Berkeley physics. 22 papers spanning dark matter to quantum mechanics.
-
ConfirmedSriram Rajamani
Corporate Vice President · Microsoft CoreAI
Corporate VP at Microsoft CoreAI and ACM Fellow. PhD in formal verification from UC Berkeley; co-built the SLAM/SDV software model checker that won the 2011 CAV Award. Now leads a product effort in Microsoft combining LLM agents with static analysis and verification tools to make agentic approaches scale with high quality.
-
ConfirmedCarina Hong
Founder & CEO · Axiom Math
Founder and CEO of Axiom Math, building the world's best AI for math reasoning engine applied to formal verification. AxiomProver recently scored 120/120 on the Putnam exam and published 8 papers solving open conjectures. Rhodes Scholar and Morgan Prize Laureate.
-
ConfirmedJason Gross
Co-founder · Theorem
Co-founder of Theorem, building AI that ships verified code. MIT CS PhD on the performance of proof assistants under Adam Chlipala. Author of Fiat Cryptography — the verified crypto behind a meaningful slice of the internet's TLS handshakes. Reported the plurality of bugs in Rocq (Coq), now a core dev.
-
ConfirmedSanjit Seshia
Cadence Founders Chair Professor, EECS · UC Berkeley
Cadence Founders Chair Professor at UC Berkeley, developing the mathematical techniques and tools that make AI and cyber-physical systems provably safe, secure, and trustworthy. A foundational voice in verified artificial intelligence.
-
ConfirmedMarco Pavone
Associate Professor · Senior Director, AV Research · Stanford / NVIDIA
Associate Professor in Stanford's Aeronautics & Astronautics Department, where he leads the Autonomous Systems Laboratory, and Senior Director of Autonomous Vehicle Research at NVIDIA. His research focuses on bringing formal methods and learning-based approaches to safety-critical robotics, spanning applications from autonomous driving to spacecraft autonomy.
-
ConfirmedRanjan Rajagopalan
Co-founder & CEO · Pramaana Labs
IIT Madras alumnus; previously co-founded Astra and was a software engineer at Google and Graviton, focused on high-performance architecture, ML models, and automated content moderation pipelines.
-
ConfirmedKrishnan Raghavan
Co-founder & CTO · Pramaana Labs
Built Glean's India search team from scratch and pioneered their enterprise conversational AI; previously a Staff Software Engineer at Google. Has spent his career turning complex AI models into verifiable, structured intelligence.
-
ConfirmedSanjay Ganapathy
Co-founder & Chief Scientist · Pramaana Labs
AI researcher; previously a Staff Research Engineer at Google DeepMind, where he architected Gemini's tool-use system and led post-training efforts to advance frontier AI capabilities.
Agenda
Run of show
Wednesday, June 10, 2026 · 3:30 PM – 8:30 PM
Full lineup revealed two weeks before the event.
-
01
4:30 PM
Welcome
Pramaana founders
-
02
5:00 PM
Fireside · The verification thesis
Vinod Khosla × Ranjan Rajagopalan
-
03
5:30 PM
Research talk · I · Verified AI: Past, Present, and Future
Sanjit Seshia (UC Berkeley)
-
04
6:00 PM
Panel · I · Formal Verification in the Wild
Sriram Rajamani, Carina Hong, Satnam Singh, Krishnan Raghavan
Moderated by Kanu Gulati -
05
6:30 PM
Research talk · II · Physical AI Deployment: Overcoming Challenges and Accelerating Validation
Marco Pavone (Stanford / NVIDIA)
-
06
7:00 PM
Panel · II · How AI changes what we can verify, and what verification means for AI
Eve Bodnia, Krishnamurthy Dvijotham, Sanjay Ganapathy, Jason Gross
Moderated by Sathya Nellore Sampat -
07
7:30 PM
Announcement · Pramaana Research Program (PRP) + Closing
Sanjay Ganapathy (Pramaana)
-
08
7:45 PM
Dinner + research booths
Open floor
Co-hosted by
Apply
Three ways in
02
Booth exhibitor
Open to research groups, labs, and early-stage companies with technical demos in the verification stack. Four to five booths during the reception. Approval-based.
Apply as a booth exhibitor →03
Speaker
If you're working on something the room should hear, get in touch. We're particularly interested in operator-side perspectives from regulated domains.
Speaker interest →-
01
Who is this summit for?
Researchers, founders, and engineers working on verification, formal methods, trustworthy AI, alignment, and the infrastructure underneath provable AI systems. We curate for depth over scale — single track, ~150 attendees, no parallel sessions.
-
02
How do I get an invite?
The Summit is invite-only. Request an invite through this site and our team will get back within 5 business days. We prioritize people building, researching, or investing in the verification layer of AI.
-
03
Can I apply to speak?
Yes — submit a talk proposal via Apply to speak. We're curating the Lightning Talks block from these. Submissions close May 20, 2026.
-
04
Is the Summit recorded?
Yes. We're not livestreaming this edition — the event will be recorded and published on our official social channels afterwards. Fully edited talks land on YouTube; shorter clips and highlights go up on LinkedIn and X.
-
05
Where is it?
San Francisco. Final venue announced once confirmed — invitees and applicants notified directly.
-
06
Is there a cost?
There is no ticket fee. The Summit is hosted by Pramaana Labs as part of our efforts to drive a much-needed conversation in the space of verification in AI.
-
07
Will there be future editions?
Yes — the Verification Summit is designed as an annual flagship for the field. 2026 is the first edition.
-
08
I lead a company working on verification or formal methods. How do we engage?
Apply as a booth exhibitor. The booth is your chance to put what you're building in formal verification in front of the researchers, founders, and operators most likely to care.
-
09
Will there be media coverage?
Yes — selected media partners attend under embargo. If you're media and want to attend or cover, reach out via this site.
Host
About Pramaana Labs
Pramaana Labs is building AI you can prove right. Backed by Khosla Ventures. We're hosting the Verification Summit as a public moment for the conversation we believe will define the next phase of the field.