PhysicsX is a deeptech company with roots in numerical physics and Formula One, dedicated to accelerating hardware innovation at the speed of software. Pulse 2.0 interviewed PhysicsX Co-Founder and Director of Simulation Engineering, Nico Haag, to learn more about the company’s background and mission.
Nico Haag’s Background
What is Nico Haag’s background?
“My background, much like that of my co-founders, sits at the intersection of engineering, deep learning, and applied science. Before co-founding PhysicsX in 2019, I built my career in the automotive and motorsport industries, working with organizations like Bentley Motors, Audi Motorsport, and Mercedes-Benz.
Today, I lead the physical engineering at PhysicsX, pushing the boundaries of AI-driven simulation and design. I work closely with our Delivery, R&D, and Platform teams to develop next-generation CAE and AI-accelerated design optimization tools that help tackle some of the most complex engineering challenges of our time.”
Formation Of The Company
How did the idea for the company come together?
“The idea for PhysicsX came together from a need to tackle the challenges in modern engineering, where high-fidelity physics simulations are crucial for designing and optimizing complex systems — think jet engines, fusion reactors, or advanced semiconductor manufacturing. The problem with traditional methods, like computational fluid dynamics (CFD) and finite element analysis (FEA), is that they’re not only computationally intense but also slow and hard to scale. This leads to bottlenecks in industries that need quick design iterations and precise results to stay ahead.
PhysicsX was founded to solve these problems by leveraging AI-powered solutions that speed up physics simulations without sacrificing accuracy. Our Large Physics Models (LPMs) allow for real-time predictions of complex physical behavior, so engineers can quickly test and refine their designs. Ultimately, our goal is to break free from the constraints of traditional computing and time limitations, helping teams optimize performance, speed up development, and optimize whole systems in one go instead of optimizing single components one by one.”
Core Products
What are the company’s core products and features?
“Our product is the PhysicsX platform, designed to support the entire engineering process — from generating high-quality data and training models to deployment and continuous optimization.
Engineering teams face a wide range of complex problems, from early-stage design exploration to real-time performance adjustments. Our unified platform helps solve these challenges by seamlessly integrating with existing engineering tools like CAD and CAE. By applying AI, it addresses a variety of needs, whether it’s structural simulations, fluid dynamics, electromagnetics, chemistry, and so on. With our AI-powered ecosystem, companies can create, train, and deploy AI models that keep improving over time, cutting down on redundant work, speeding up knowledge transfer, and ensuring that innovations are shared across projects.
Challenges Faced
What challenges does PhysicsX face?
“One of the main challenges we face right now is making sure our customers feel confident that our AI models are maintaining the physical accuracy of traditional methods, while also being able to adapt to new design spaces. To tackle this, we are (1) developing broad foundation models that generalize very well and (2) use AI uncertainty quantification to automatically trigger a CAE feedback loop (i.e., active learning) to retrain the models in the areas that need the most attention. For a user that means that there is no need to think about model accuracy anymore, as it will just retrain itself in the background if uncertain.
Another challenge we’ve worked through is managing the complexity and variability of real-world geometries. To address this, we developed Large Geometry Models (LGMs), which help us efficiently encode 3D geometries and make AI predictions across a wide range of designs. By combining this with advanced optimization techniques like Bayesian optimization and genetic algorithms, our platform can explore over 100,000 geometries a day, optimizing key performance metrics for a variety of industries.
Differentiation From The Competition
What differentiates the company from its competition?
“What sets us apart from the competition is our unique blend of deep engineering expertise and cutting-edge AI. While many startups are working to bring AI-driven solutions for physics to market, most don’t have the same level of engineering know-how needed to address the toughest, most impactful challenges our customers face. We bring together a strong foundation in both science and engineering, combined with advanced AI technology for enterprise applications, which allows us to solve real-world problems in a way that delivers real, lasting value.”
Sector Developing This Year
How do you see your sector developing in 2025?
“I think we will continue to see an increasingly positive dynamic in shifting towards AI-driven digital engineering solutions. AI-powered simulations are becoming more widespread, allowing engineers to validate and refine designs much faster. Our LGMs and LPMs combined with out platform will play a key role, enabling engineers to leverage engineering AI, explore vast design spaces, and cut down on manual iterations.
From early-stage design to manufacturing and even in-field performance monitoring, AI models will continue to learn and adapt, optimizing designs not just during development but throughout their operational life. I believe that one day, engineering AI will become the norm, offering real-time insights while developing into a powerful tool for rapid co-engineering, similar to how AI has transformed areas like coding and content creation.”
