Two F1 Engineers Just Tripled Their Revenue and Raised $300M by Solving a Problem Most People Don’t Know Exists
PhysicsX hit a $2.4 billion valuation this morning. The product is physics simulations that run in seconds instead of days. The real story is much bigger than that.
Engineering simulation is one of the most expensive, least glamorous bottlenecks in modern industry. If you’re designing a jet engine component, an EV battery pack, or a semiconductor node, you run physics simulations to test how it behaves under stress, heat, and vibration. Those simulations run on conventional computational tools and take hours — sometimes days — per iteration. Every hour is engineering time, compute cost, and product cycle delay.
PhysicsX, founded in London by two former Formula 1 aerodynamicists, just raised $300 million in an oversubscribed Series C led by Singapore’s Temasek at a $2.4 billion valuation. That’s more than double the company’s valuation from its Series B less than a year ago. NVIDIA, Applied Materials, Atomico, General Catalyst, and Siemens all increased their stakes. The company has about 350 employees, has doubled its customer count in the past year, and tripled its booked revenue.
The product replaces conventional physics simulations with AI models trained on simulation outputs and real-world industrial data. Results come back in seconds, not days. The fastest-growing vertical right now: the hardware needed to build and run AI data centers.
Let that land for a second. The AI boom is so large that it’s driving demand for faster AI tools to design the physical infrastructure that runs AI. The feedback loop is compounding.
Mental Model: Second-Order Thinking
The first-order read on PhysicsX is: AI makes engineering simulation faster, customers save time and money. That’s real.
The second-order read is more interesting. If simulation cycles compress from days to seconds, the cost of iteration collapses. Which means engineering teams can run 100x more design variants in the same time window. Which means product quality improves faster and R&D costs per breakthrough go down — not just incrementally, but structurally. The teams that adopt this first don’t just move faster. They compound their design advantage over every competitor still running conventional tools.
Third order: whoever controls the Large Physics Models — the foundational AI layer that generalizes across industries, physics domains, and hardware types — captures the switching cost that makes this a platform, not just a point tool. That’s what PhysicsX is building. And that’s why Siemens, NVIDIA, and Applied Materials are all in the cap table together.
The Contrarian Take
Every article about this round will call it a physics-AI play or a simulation acceleration story. I’d call it an industrial infrastructure bet.
The most durable businesses in software aren’t the ones with the best features. They’re the ones embedded deepest in a workflow that can’t stop. Aerodynamics testing at Airbus, chip thermal modeling at TSMC, battery simulation at a major OEM — those aren’t experiments. They’re mission-critical processes that run every day. When PhysicsX gets integrated into those pipelines, they don’t get ripped out. They get expanded.
Founders Jacomo Corbo and Robin Tuluie came out of F1, where physics simulation is as close to religion as engineering gets. They understood the problem from the inside. They built the solution before the AI data center boom handed them their best use case on a silver platter.
The hardest moat to build is the one where the customer can’t imagine running their workflow without you. PhysicsX is building that moat in a market that just got a very large tailwind.