PhysicsX raised $300 million in a Series C round at a $2.4 billion valuation — more than double its valuation twelve months ago. The round was led by Temasek, the Singapore sovereign wealth fund, with participation from M&G Investments, Intrepid Growth Partners, and existing backers including NVIDIA, Siemens, Applied Materials, Atomico, and General Catalyst.
The company compresses engineering simulation from hours or days to seconds using AI models trained on physical systems. Its platform is already deployed across aerospace, defense, semiconductors, automotive, energy, and materials manufacturing — the sectors that build the world's most critical hardware and can least afford slow development cycles.
The simulation workflows at the heart of hardware development are slow, expensive, and difficult to scale. PhysicsX replaces numerical solvers that take hours or days with AI models that predict physical behavior in seconds — and the technology has reached an inflection point where GPU economics and model architectures make it viable at production scale.
From the Race Track to the Factory Floor
PhysicsX was founded in 2019 by Jacomo Corbo and Robin Tuluie, both former Formula 1 engineers. Corbo was previously chief scientist and co-founder of QuantumBlack, McKinsey's AI division. Tuluie served as head of R&D at Renault (Alpine) F1 and vehicle technology director at Bentley Motors. The company emerged from stealth in 2023 with a $32 million Series A led by General Catalyst, and has since raised over $500 million in total funding.
The Formula 1 connection is not incidental. In motorsport, engineering teams have minutes between practice sessions to simulate, decide, and act. The tolerance for slow computation is effectively zero. That pressure shaped the company's approach: build AI that can predict physical behavior reliably enough to make real-world design decisions, not just generate academic benchmarks. "Almost every hard problem in the physical economy comes down to how fast and how well engineers can work through the underlying physics," Corbo told the company's investor base. "For decades, that has been the binding constraint on hardware innovation. Physics AI removes it."
We are giving engineers the ability to explore thousands of designs where they once managed a handful, in seconds rather than weeks, across the most demanding industries in the world.— Jacomo Corbo, Co-Founder & CEO, PhysicsX
The company's platform integrates directly with existing computer-aided engineering tools including ANSYS, Siemens NX, and OpenFOAM. Its proprietary models — called Large Physics Models (LPMs) and Large Geometry Models (LGMs) — are trained exclusively on customer data and deployed with uncertainty quantification for production use. This is not AI bolted onto existing solvers as an afterthought; the platform treats AI as a core primitive, running inference alongside traditional numerical methods and choosing the fastest path depending on the required precision.
What $300 Million Buys
PhysicsX has doubled year-over-year recognized revenue, tripled booked revenue, and more than doubled its customer base and headcount to over 300 people in the past twelve months. The new capital will accelerate global expansion — the company is opening offices in the US and Singapore while keeping its headquarters in London — and fund development of larger pre-trained physics AI models.
The timing reflects a structural shift in the industrial software market. Model architectures and GPU economics have matured to the point where physics AI at production scale is viable. The broader simulation software market stood at $26.5 billion in 2025 and is projected to reach $70.7 billion by 2033 at a 13 percent CAGR. Within physics AI specifically, the market is expected to grow from $236 million in 2025 to $683.5 million by 2035. Competitors Neural Concept and Rescale also raised large rounds in 2025, but PhysicsX differentiates by embedding AI across the full product lifecycle rather than accelerating isolated simulation steps.
The company has secured deep infrastructure partnerships to support that ambition. NVIDIA provides GPU-accelerated compute through its PhysicsNeMo framework and the broader CUDA ecosystem. Siemens collaborates on data center power infrastructure applications. CoreWeave supplies dedicated GPU clusters for training. T-Systems, the Deutsche Telekom arm, is hosting PhysicsX models on its European Industrial AI Cloud, providing sovereign infrastructure for industrial customers in Europe.
High-fidelity physics simulation has always been powerful, but it has also been slow, costly, and the preserve of a small group of specialists. Physics AI changes that in every dimension. It makes high-fidelity simulation dramatically more efficient, augments and improves on pure simulation results with real-world data, and opens applications that were never practical before.— Robin Tuluie, Founder & Chairman, PhysicsX
What makes the PhysicsX story worth tracking beyond the headline number is what it says about the direction of industrial AI. The company is not selling a tool that helps engineers simulate faster. It is building a platform that changes what engineering organizations can attempt in the first place — the difference between exploring a handful of design variants and exploring thousands. In aerospace, the difference between a design cycle measured in months and one measured in days. In semiconductors, the difference between optimizing a chip package for performance and optimizing it for manufacturability, thermal behavior, and cost simultaneously.
For an audience of engineers and technical decision-makers, the message is straightforward: the bottleneck in hardware innovation is no longer physics. It's time. And PhysicsX is compressing it.
Correction (June 22, 2026): A prior version of this article misstated PhysicsX's total funding. The company has raised over $500 million across Series A, B, and C rounds. (June 8, 2026)