PhysicsX sits in a part of the AI economy where errors cannot be hidden behind a polished interface. Its customers design aircraft, semiconductor processes, energy systems and other physical products where a bad prediction can become an expensive engineering decision.
That is why the backgrounds of co-founders Robin Tuluie and Jacomo Corbo matter. Tuluie worked in Formula One and automotive R&D, including at Mercedes and Bentley. Corbo previously co-founded QuantumBlack, the applied-AI company later acquired by McKinsey, and worked on race strategy in Formula One. PhysicsX emerged from the overlap between physical engineering and machine learning.
The bottleneck is simulation time
Traditional engineering depends on numerical simulation to test how designs behave before hardware is built. Those calculations can be extremely accurate, but they can also be computationally expensive and slow enough to limit how many designs engineers explore.
PhysicsX trains models to approximate parts of that physics more quickly, allowing teams to search a wider design space and reserve the most expensive simulations for validation. The value proposition is therefore not that AI replaces physics, but that it changes the iteration loop around physics.
Industrial AI attracts a different investor set
The company's funding reflects how strategically important that layer has become. Its June 2026 Series C raised $300 million at an approximately $2.4 billion valuation, led by Temasek with participation from investors including industrial and semiconductor groups.
Applied Materials, Siemens and Nvidia-related capital appearing around the company is significant. These are organisations that understand the cost of engineering cycles and the value of compressing them. PhysicsX is selling into budgets tied to R&D productivity rather than generic enterprise software seats.
A British AI company built around industry
Britain's AI story is often told through foundation-model safety, fintech and London software. PhysicsX gives the ecosystem a different shape: AI built for industrial engineering, with customers whose economic output sits in factories, aircraft, energy systems and chips.
If the company scales, its most important contribution may be showing that the UK's strength in applied mathematics, motorsport engineering and enterprise AI can combine into an exportable industrial software category.