AI powers engineering simulations at PhysicsX with $32M from stealth

Generative AI and how it accelerates consumer software and products are hot topics in artificial intelligence. Two theoretical physicists and a Formula One engineer co-founded PhysicsX, an AI startup that is now operating out of stealth to create and manage enterprise physical systems.

London-based PhysicsX has developed an AI platform to create and run simulations for engineers in automotive, aerospace, and materials science manufacturing, where model testing before production often causes development bottlenecks. This $32 million project emerges from stealth today.

General Catalyst leads Series A. Several intriguing financial and strategic backers are also in the round. They include Standard Industries, NGP Energy, Radius Capital, and Henry Kravis, KKR co-founder and chairman. The funds will support business growth and platform development. This is PhysicsX’s first outside funding.

PhysicX addresses a persistent but overlooked issue in manufacturing and physical production.

Every time a new idea is introduced, whether in an experimental lab or a live industrial environment, engineers must simulate how it will work before developing and improving it. Scientists and engineers are usually the ones who simulate and test, sometimes using AI but mostly manually.

“Simulating airflow across an object may take an hour or two, but something more complex may take a day or longer. This has computational and time costs. In an interview, Robin Tuluie, who co-founded PhysicsX with Jacomo Corbo, said that limits optimization depth.

They experience the pain points firsthand.

In his theoretical physics career, Tuluie has had two lives. He studied astrophysics with Nobel Prize winners. He then worked in racing as Renault’s head of R&D and Mercedes’ Chief Scientist, designing cars that won four Formula One world championships and earning him fame. He worked on automotive design at Bentley and Volkswagen for years.

Corbo, who earned his PhD from Harvard, founded and led QuantumBlack, McKinsey’s AI labs, working with Formula One and other automotive and industrial clients on difficult product engineering problems.

Corbo said the pair has assembled a team of 50 scientists, including mechanical engineers, physicists, and others, to build the PhysicsX platform, which targets automotive and other applications.

“We are building an enterprise platform to support a pretty broad range of domain applications tied to building and optimization problems and physics simulation bottlenecks,” he said. PhysicsX lets you predict a system’s physics with 10,000 to a million times greater accuracy and fidelity. Now be much more sophisticated about mining in high-dimensional space.”

PhysicsX’s arrival is timely in the world of deep learning and AI, especially in its application to the physical world.

Corbo believes that the physical turn will mark the next frontier of AI research and development. DeepMind released new research earlier this month on how it was applying very advanced machine learning to short- and long-term weather prediction.

“This is the first time AI models, these deep learning models, these geometric deep learning models, are overtaking numerical simulation for weather,” Corbo said. “That’s happening across physics more broadly. That enables many engineering applications, so we’re building a platform to do that across sectors and domain problems.

Companies have had trouble ripping out old infrastructure to adopt new IT and methods for digital transformation. Although PhysicsX is undergoing a “digital transformation,” the startup can avoid those challenges because its engineering and R&D applications do not require scaling across organizations.

It’s a new approach that will disrupt industrial companies’ development strategies. General Catalyst is betting on a hot area—AI — and breaking new ground by backing a startup that predicts its evolution.

According to General Catalyst MD Larry Bohn, “PhysicsX expands engineering boundaries in critical sectors, under the direction of a team deeply proficient in simulation engineering and machine learning.” We think PhysicsX can transform complex engineering with its credibility, customer relationships, and technical expertise. This supports our industrial transformation vision and gives PhysicsX the chance to define advanced industries.