PhysicsX Raises $135M to Bring AI to Engineering

PhysicsX Secures $135 Million in Series B Funding to Reshape AI-Driven Industrial Engineering

There is a quiet but seismic shift happening inside the world's most consequential industries — aerospace, automotive, semiconductor manufacturing, and defence — and it is being driven not by a Silicon Valley giant but by a relatively young London-based startup that has spent the better part of six years solving one of engineering's most stubborn bottlenecks. PhysicsX, a deep tech company built at the crossroads of artificial intelligence and industrial physics, has just closed a $135 million Series B funding round led by prominent European venture capital firm Atomico. The round brings the company's total capital raised to nearly $170 million and positions it as one of the most well-funded AI-for-engineering startups in the world today. For those of us tracking the intersection of frontier AI and physical industry here at The AI World, this funding round is far more than a financial headline — it is a signal that the market has firmly accepted AI as the next foundational layer of industrial design.

The significance of this raise cannot be overstated. Unlike many AI startups that build tools for knowledge workers, software developers, or content creators, PhysicsX is targeting the physical world — the domain of turbine blades, semiconductor wafers, electric vehicle components, and aircraft engine casings. These are environments where the consequences of a design flaw are not just financial; they can be catastrophic. The fact that a company operating in this space has attracted $135 million in a single round, with strategic backing from industry giants like Siemens and Applied Materials, tells you everything you need to know about where enterprise confidence in AI is heading in 2026.

The Problem PhysicsX Was Built to Solve

To understand why PhysicsX matters, you first need to appreciate how fundamentally broken the traditional engineering simulation process has always been. For decades, engineers across industries have relied on tools like Ansys, Siemens NX, CATIA, and OpenFOAM to run computational simulations of how physical objects behave under stress, heat, pressure, and fluid dynamics. These tools are extraordinarily powerful, but they are also extraordinarily slow. A single simulation run for a complex component — say, a turbine blade for a jet engine or a heat sink for a next-generation chip — can take anywhere from several hours to several days to complete. In practice, this means that an engineering team designing a new component might only be able to test and evaluate three or four design options before deadlines force them to commit to a direction. The best possible design often goes undiscovered simply because there was not enough time or computing power to test for it.

This is not a minor inefficiency. In industries where marginal performance gains translate directly into billions of dollars in fuel savings, competitive differentiation, or national security outcomes, the inability to explore design space exhaustively is one of the most expensive limitations in modern industry. Engineers know their simulations are constrained. They know they are leaving better designs on the table. And until recently, they had no good alternative.

PhysicsX was founded in 2019 to change this entirely. The company builds what are known as physics surrogate models — deep learning systems trained on high-fidelity simulation data that can predict the physical behaviour of a component in a matter of seconds rather than hours or days. Instead of running a full computational fluid dynamics or finite element analysis simulation every time an engineer wants to test a new design configuration, the PhysicsX platform essentially acts as an AI stand-in for the simulation itself. A team that used to test four turbine blade geometries in a week can now explore thousands of configurations in the same amount of time, at a fraction of the cost, with a level of accuracy that is validated against the same simulation tools they have always trusted.

The company's platform is built to integrate directly with the tools engineers already use in their daily workflows — Ansys, CATIA, OpenFOAM, Siemens Star CCM+, and others — so there is no requirement to abandon existing processes or rebuild infrastructure. This design choice has been central to the company's commercial traction because it removes one of the largest barriers to enterprise adoption of AI tools: the disruption cost of switching systems.

The Founding Team and the Formula That Has Driven Growth

One of the more remarkable things about PhysicsX is where it came from. This is not a company founded by software engineers or machine learning researchers working from a university lab. Its origins lie deep inside professional motorsport, one of the most technically demanding environments on earth.

Robin Tuluie, who serves as chairman of PhysicsX, spent years as head of research and development at the Renault Alpine Formula One team and later at the Mercedes F1 team before going on to serve as vehicle technology director at Bentley Motors. Jacomo Corbo, the company's CEO, co-founded QuantumBlack, an AI and data science consultancy that was later acquired by McKinsey & Company, and worked as chief race strategist at Renault F1. Nicolas Haag, the third co-founder, serves as director of simulation engineering. Together, they bring a combination of elite engineering intuition, practical AI experience, and deep commercial understanding that is genuinely rare in the deep tech world.

The background in motorsport is not merely an interesting biographical footnote. Formula One is one of the few commercial environments in the world where engineering simulation is run at absolute maximum intensity — where teams run thousands of aerodynamic simulations per week and where the ability to explore design space faster than a competitor translates directly into race wins and championship points. The founders understood the simulation bottleneck in a visceral, firsthand way before they ever set out to solve it for the broader industrial world.

Since closing a $32 million Series A in November 2023, the company has grown from a team of just over 50 people to more than 150, and has more than quadrupled its revenue, which is a rate of growth that few enterprise deep tech companies achieve in any 30-month window. While the company has not disclosed an absolute revenue figure, the trajectory has clearly been compelling enough to attract one of Europe's most respected venture capital firms to lead a round of this size. General Catalyst, which led the Series A, returned for the Series B alongside a strong new group of investors including Temasek, July Fund, NGP, Radius Capital, Standard Investments, and Allen & Co.

Strategic Investors Signal a Deeper Industry Shift

Perhaps the most telling element of this entire funding round is not the size of the cheque or the identity of the lead investor, but the composition of the strategic investor roster. Siemens and Applied Materials — two of the most consequential companies in the global industrial and semiconductor supply chains — have both taken stakes in PhysicsX. This is notable for several reasons.

Siemens, through its Xcelerator platform, is itself a major provider of engineering simulation software. In some product areas, its tools directly compete with what PhysicsX is building. The fact that Siemens has chosen to invest rather than simply watch from the sidelines speaks to a growing recognition inside established industrial technology companies that AI-native simulation is not going to be a marginal improvement on top of existing tools — it is going to be a fundamental replacement of certain categories of work, and the companies that get ahead of that shift will have an enormous advantage over those that do not.

Applied Materials, the world's largest semiconductor equipment company by revenue, faces its own simulation-intensive challenges in chip manufacturing. The complexity of designing and validating new chip fabrication processes is immense, and the potential value of being able to run faster, more exploratory simulations in that domain is enormous given the scale of capital expenditure involved. Their participation here suggests that PhysicsX's technology roadmap extends meaningfully into the semiconductor sector, which is one of the most strategically important industries on the planet right now given the global competition over chip manufacturing capacity.

Laura Connell, partner at Atomico, captured the investor thesis clearly when she described PhysicsX as unlocking a genuinely new engineering paradigm — one built on the fusion of frontier AI research and deep domain expertise in the industries that form the backbone of the global economy. That framing matters because it positions PhysicsX not as a productivity tool bolted onto existing workflows, but as a foundational technology that changes what engineering teams are capable of achieving in the first place.

Jacomo Corbo, in articulating the company's vision, pointed to the broader geopolitical and economic context in which PhysicsX operates. Industrial manufacturing is being reshaped by questions of supply chain sovereignty, geopolitical realignment, and an urgent need for innovation in the physical industries that underpin national economies. The urgency to design better aircraft, more efficient chips, and more resilient industrial components has never been higher, and the traditional simulation tools that engineers have relied upon for decades are simply not fast enough to keep pace with that urgency. PhysicsX is building into exactly that gap.

Where the $135 Million Goes and What Comes Next

The company has been transparent about its priorities for deploying this new capital. North America is the primary geographic target for expansion, and that focus makes strategic sense given that the United States accounts for a disproportionate share of global aerospace, defence, and semiconductor spending. Building a meaningful commercial presence in that market — with dedicated sales, engineering support, and customer success infrastructure — is a natural next step for a company that has built its initial customer base primarily from a European base.

Beyond geographic expansion, the other major use of the funding is investment in what PhysicsX calls physics foundation models. This is where the company's longer-term vision becomes especially interesting from a technical standpoint. Today, building a surrogate model for a new physical domain or a new class of engineering problem requires significant investment in training data and model development. The goal with foundation models is to build a general-purpose AI architecture that captures enough fundamental physical understanding that it can be adapted to serve new industries and new problem types without needing to be rebuilt from scratch each time. If successful, this approach would transform PhysicsX from a company with strong but specialised capabilities in specific domains into a platform that can serve as the AI simulation backbone for virtually any industry that involves the physical behaviour of materials and components.

The market opportunity that PhysicsX is chasing is substantial and growing quickly. The global simulation software market was valued at nearly $20 billion in 2024 and is projected to exceed $36 billion by 2030, growing at a compound annual rate of over 10 percent according to market research from MarketsandMarkets. That growth is being driven by the same forces that are propelling PhysicsX — increasing complexity in manufactured products, shortened design cycles, and the growing availability of AI tools capable of handling simulation workloads that were previously intractable.

PhysicsX is not operating in a vacuum. Other well-funded companies are chasing adjacent parts of the same opportunity. Rescale raised $115 million in 2025 to modernise high-performance computing infrastructure for engineering simulation. BeyondMath secured $8.5 million for an AI-powered engineering simulation platform. Quanscient raised €10 million to build multiphysics simulation software that runs significantly faster than conventional on-premises alternatives. The competitive landscape is developing quickly, but PhysicsX's combination of deep domain expertise, an established enterprise customer base, integration with existing engineering tools, and now nearly $170 million in total funding gives it a meaningful head start in a race that is only just beginning.

What This Means for the Future of AI in Industrial Engineering

At The AI World, we spend a significant amount of time examining how artificial intelligence is moving beyond the digital domain — beyond language models, recommendation engines, and software automation — into the physical world where products are designed, manufactured, and deployed. The story of PhysicsX is one of the clearest illustrations of that transition we have seen.

What this company has built represents a genuinely new category of AI capability: one that does not just process information or generate content, but that models the physical laws of the universe well enough to predict how matter will behave under complex conditions. The implications of that capability, if it continues to mature, are profound. Engineers who have spent their careers working around the limitations of slow simulation tools are beginning to operate in an environment where the design space is effectively unlimited — where the question is no longer "how many configurations can we test?" but "which of the thousands of configurations we tested is actually the best one?"

That shift in what is possible is already happening at PhysicsX customers across aerospace, automotive, and semiconductor manufacturing. As the company expands into North America, grows its team, and invests in larger foundation models, it is building the infrastructure for that shift to happen at scale, across every sector of heavy industry. In a world where the race to manufacture better, more efficient, and more resilient products has become a matter of both economic and national strategic importance, the technology that PhysicsX is building is not peripheral to that race — it sits squarely at the centre of it. The $135 million raised in this round is not just a vote of confidence in one company. It is a marker of where the most serious investors in the world believe the next decade of industrial transformation is going to be built.