PhysicsX raised $300M at a $2.4B valuation in a Temasek-led round to bring AI-powered engineering simulation to aerospace, semiconductors, and manufacturing.
When Jacomo Corbo and Robin Tuluie left the high-pressure world of Formula 1 racing, they carried with them an unusual obsession: the belief that the most important bottleneck in modern manufacturing was not raw materials, labor, or capital. It was physics simulation.
Engineering teams at aerospace companies, semiconductor manufacturers, and automotive giants were spending days, sometimes weeks, waiting for supercomputers to model how a blade would handle stress at 30,000 feet, or how heat would move through a chip the size of a thumbnail. Corbo and Tuluie thought AI could collapse that wait into seconds. Five years later, the world's biggest investors are betting $300 million that they were right.
PhysicsX, the London-based startup the two founders built, announced on June 8, 2026, that it had closed an oversubscribed Series C financing round at a valuation of approximately $2.4 billion. The round was led by Temasek, Singapore's state-owned sovereign wealth fund, which first invested in the company in 2025 and has been closely involved in guiding PhysicsX's expansion into Asian markets.
New investors M&G Investments and Intrepid Growth Partners joined the round, along with a roster of returning backers that reads like a who's-who of the technology and industrial sectors: NVIDIA, Applied Materials, Atomico, General Catalyst, Siemens, July Fund, NGP, and Radius.
The round was oversubscribed, meaning investor demand exceeded the amount PhysicsX was seeking to raise, a meaningful signal in an environment where venture capital has grown considerably more selective since 2022.
What PhysicsX Actually Does
At its core, PhysicsX builds what it calls "Large Physics Models." The name is a deliberate echo of the large language models that power AI chatbots, but the analogy only goes so far. Instead of predicting the next word in a sentence, PhysicsX's models predict how physical objects and systems will behave under real-world conditions: aerodynamic forces, thermal stress, fluid dynamics, electromagnetic behavior.
Traditional engineering simulation software, long dominated by established players like Ansys and Dassault Systemes, runs numerical calculations that can take hours or days to complete even on powerful computing infrastructure. PhysicsX replaces much of that heavy computation with AI inference, delivering results in seconds. The company argues its models maintain the accuracy that engineers need for high-stakes decisions, a claim that its growing customer list across aerospace, automotive, semiconductor manufacturing, energy, and materials science appears to support.
"Almost every hard problem in the physical economy depends on how fast and how well engineers can work through the underlying physics."
Jacomo Corbo, CEO and Co-Founder, PhysicsXIn aerospace, the company says it has already cut aircraft design cycles from months to days. In semiconductor manufacturing, where the demand for faster chip design is relentless, PhysicsX is helping clients significantly reduce the time needed to develop new equipment prototypes. CEO Corbo told the Financial Times that semiconductors are expected to become PhysicsX's largest industry segment by the end of 2026.
The Formula 1 Backstory
The origin story of PhysicsX is embedded in one of the most simulation-intensive sports environments on earth. Formula 1 teams run hundreds of thousands of aerodynamic simulations each season, racing to find a tenth of a second of lap time in the curvature of a wing or the angle of a diffuser. Both founders spent years operating in that world before pivoting to a broader industrial application.
Corbo served as chief race strategist at Renault F1 and later became co-founder and chief scientist at QuantumBlack, the AI consultancy that McKinsey acquired and that has since become one of the most recognized names in applied machine learning. Tuluie served as head of research and development at Renault's Alpine F1 team and later as vehicle technology director at Bentley Motors, one of the more demanding environments in precision engineering.
The company was founded in 2019, operated in stealth for several years, and emerged publicly in 2023 with a $32 million Series A led by General Catalyst. A $135 million Series B followed in 2025, anchored by Atomico and later extended with strategic backing from NVIDIA at a valuation approaching $1 billion. The trajectory from there to the current $2.4 billion is striking: a more than 2.4x increase in valuation in under twelve months.
The AI Boom Is Its Best Customer
There is a compelling and somewhat counterintuitive dynamic at the center of PhysicsX's growth story. The company is an AI startup whose most important customers are the companies building the physical infrastructure that the rest of the AI industry depends on.
Every large language model requires a data center. Every data center requires cooling systems, power turbines, compressors, and custom chip packaging. Every chip requires precision manufacturing equipment designed through iterative simulation. The AI boom, in other words, is not just creating demand for software. It is generating an enormous and urgent need to design, test, and refine the physical machines that make AI possible at scale. PhysicsX sits squarely in that supply chain.
Corbo himself has described the company as currently "very supply-side limited," telling the Financial Times that PhysicsX is having to moderate how quickly it rolls out its platform to existing customers simply because demand is outpacing its ability to deliver. That is an unusual problem for a seven-year-old startup to have, and investors have clearly decided it is the right kind of problem to bet on.
A Crowded and Well-Funded Field
PhysicsX is not the only company pursuing the industrial AI simulation market, and the competitive landscape includes some formidable opponents. Notably, Amazon founder Jeff Bezos has invested in Project Prometheus, a company founded in 2025 that specializes in physical AI and industrial automation and was reportedly endowed with $10 billion in starting capital. The European market is also seeing consolidation, with Paris-based Mistral AI recently acquiring Austrian startup Emmi AI to expand its industrial AI footprint.
On the incumbent side, traditional simulation software leaders Ansys and Dassault Systemes have spent decades building the tools that engineering teams rely on. Synopsys has agreed to acquire Ansys in a deal that underscores just how much strategic value the market places on simulation capability. PhysicsX's argument is that physics-informed AI models can leapfrog those legacy tools on speed without sacrificing the accuracy that makes simulation useful in the first place.
What Comes Next
The Series C proceeds will be deployed across several priorities. The company has grown from roughly 150 employees to more than 350 over the past year and plans to continue expanding its team. A new office is planned for Singapore, reflecting Temasek's role not just as a financial backer but as a bridge to Asian industrial markets, particularly in semiconductor manufacturing, where Singapore has a significant industrial presence.
Expansion in the United States is also a priority. The American aerospace, defense, and semiconductor manufacturing markets represent a large share of global demand for exactly the kind of engineering simulation PhysicsX accelerates, and the company's recent ranking second on Sifted's inaugural European AI 100 list suggests it is beginning to attract the international attention its technology warrants.
For Corbo and Tuluie, the journey from optimizing lap times in Formula 1 to optimizing jet engine blades and chip packages for a $2.4 billion AI company is, on reflection, less surprising than it sounds. Formula 1 has always been a laboratory for engineering ideas that eventually find their way into broader industry. This time, the idea that made it out is the ability to simulate physical reality faster than the physical world itself can reveal its answers.
0 Comments