Vayuh at Exponential Risk London
Physics-Informed AI
for Catastrophe Risk
Machine-learned storm dynamics — not hand-coded approximations.
Trusted By Industry Leaders
Why It Matters
A New Standard in Cat Modelling
Learned Physics
Neural operators trained on high-resolution simulations capture storm dynamics no hand-coded model can express.
10M+ Event Sets
Underwrite solar farms, specialty portfolios, and emerging risks with statistically robust, custom event sets at scale.
Global Generalization
Deploy immediately in EU, Australia, Canada, Mexico — same model, no rebuild.
Higher Granularity
Resolution that finally matches modern risk accumulation tools — discriminate risk at individual-asset level.
About
Insurance-Grade Risk Intelligence
Vayuh combines atmospheric physics with modern AI to build catastrophe models purpose-built for the insurance industry. Our hybrid approach blends atmospheric dynamics with machine learning — delivering results 100x faster than traditional numerical weather models.
The founding team brings peer-reviewed expertise in data assimilation, severe storm physics, and high-performance computing to the problem of SCS loss estimation.
Mayur Mudigonda
Founder & CEO
Former UC Berkeley AI researcher and Berkeley Lab scientist. ACM Gordon Bell Prize 2018.
LinkedIn →Meet us at Exponential Risk
The Mermaid London · March 10–11, 2026