PhysicsX confirms $20M from Nvidia at ‘nearly’ $1B valuation
The deal furthers a strategic partnership between the chip giant and the startup building an AI software engine for creating and testing hardware for defence, aerospace, manufacturing and others
PhysicsX — the London-based startup building an AI-based software platform to run better simulations for those designing, building and testing advanced hardware — is getting a little more firepower to expand its business.
Today the startup confirmed that NVentures, the VC arm of Nvidia, has made a strategic investment of $20 million into its latest round, bring the total funding for the Series B to $155 million.
The startup’s valuation appears to be staying largely flat with Nvidia’s infusion: it is currently at just under $1 billion, which was the same as it was in June 2025 when it initially announced the round at $135 million, led by Atomico with participation from Temasek, Siemens, Applied Materials, July Fund, General Catalyst, NGP, Radius Capital, Standard Investments, and Allen & Co.
For those who have been watching PhysicsX, this news might not come as a huge surprise.
For starters, the two had already been working together. Earlier this month the pair, along with Deutsche Telekom, said that PhysicsX would be a key partner on DT’s new Industrial AI Cloud, a service powered by Nvidia chips and other products that the carrier is building.
The three have a common goal: as manufacturing becomes more digitised, the hope is that more builders will adopt AI-based hardware and software to execute their work, and will in turn turn to clouds optimised for that work.
“The next industrial revolution will be AI-native — and the next generation of hardware will be imagined, designed, and proven in software. This investment reflects our investors’ shared conviction of that future,” said Jacomo Corbo, CEO & Co-Founder of PhysicsX, in a statement. “Together, we’re reimagining engineering and manufacturing and building the AI infrastructure to solve the most important industrial challenges of our time. The present can’t wait for tomorrow’s technologies and we’re working hard — with our customers, partners, and investors — to build the software that bridges that gap.”
PhysicsX might be optimised to work on Nvidia architecture, but it’s likely also heavily leaning on Nvidia chips to build and train its own platform. That compute does not come cheap, and so, as with other investments that Nvidia has made into AI companies, some of that funding is likely going to go towards defraying the cost of that compute, in a classic circular finance arrangement.
The other reason this may not come as a surprise is because rumours of the round surfaced at the time that the Nvidia-Telekom-PhysicsX deal was getting announced. Contacted for comment, neither Nvidia nor PhysicsX would comment on the other detail that was leaked at that time: that Nvidia was also committing to investing up to $100 million in PhysicsX between this Series B and future rounds. That extra, future $80 million will be predicated on how PhysicsX gets adopted by customers, meaning Nvidia’s investment placeholder could go up… or down.
On one side, the GPU chipmaker is currently the world’s richest technology company based on market cap, thanks to the current boom in AI (and its extreme reliance on GPUs). On the other, if the bubble that is currently surrounding AI does burst… that could spell a different outcome for a lot of companies.
It will be interesting to see if PhysicsX is one of them. The startup — whose product is predicated on the idea of ensuring better predictability in the process of designing, manufacturing and testing new hardware — has built resilience into its very fabric, in a sense. The company first emerged from stealth in 2023, founded by Robin Tuluie and Jacomo Corbo, two physicists who gained fame for their contributions to the high-performance automotive world, specifically in racing.
Initially, PhysicsX targeted companies in materials design, automotive and manufacturing, where building and testing new products cost a fortune and didn’t work well.
More recently the startup has been expanding its messaging to the concepts of technological sovereignty and resilience for its customers, and it’s added one more critical sector to its roster of target customers: defence.
Here, it’s courting customers (and presumably signing them: we are asking for customer names) that are building high-precision drones and other next-generation devices for reconnaissance, munitions and more. Typically, that is work sometimes being done under time-constrained and budget-constrained conditions. The work also comes with a high degree of complexity and potential failure, as the recent events in Kenya demonstrated.
“What PhysicsX buys you is the ability to be able to predict the physics [of a system] with very, very high accuracy and fidelity, doing it, anywhere from 10,000 to a million times faster,” Corbo told me in an interview in 2023.
Indeed, defence companies are working at the cutting edges of autonomy and cyber warfare, and so companies that present more efficient and foolproof ways to test the next generation of kit have become hot properties.