PhysicsX Ltd., a startup using artificial intelligence to help engineers design new products, today announced that it has closed a $32 million Series A funding round led by General Catalyst.

KKR co-founder Henry Kravis, Standard Investments, NGP and Radius Capital participated as well. The investment marks the first time PhysicsX has raised outside funding since it was launched in 2020 by its co-chief executive officers, theoretical physicists Robin Tuluie and Jacomo Corbo.

When engineers develop a new hardware product such as an industrial robot, they typically don’t build a prototype right away. Instead, they test the product’s design in a simulation to find potential flaws that should be fixed before a prototype is manufactured. The task is usually carried out with so-called numerical simulation applications, which use complex physics equations to estimate if a product can be expected to work as expected. 

The equations that engineers use to power their simulations often require a significant amount of computing infrastructure to run. Moreover, the calculations can take days or weeks to complete, which slows down product development. PhysicsX is working to speed up the process.

The London-based company offers AI tools that it claims can carry out simulations in a fraction of the time required by competing programs. According to TechCrunch, PhysicsX’s software is capable of predicting how a hardware system will behave 10,000 to 1 million times faster than traditional simulation applications. Moreover, it promises to do so with a high degree of accuracy.

AI models can outperform traditional simulation software because they take a statistical approach to processing data. Instead of running hardware-intensive physics equations to simulate a system, an AI model analyzes how the system behaved in the past and predicts how it’s likely to behave in the future. The AI also takes into account how similar systems behaved under similar circumstances.

“Engineering design processes were transformed by numerical simulation and the availability of high-performance compute infrastructure,” Tuluie said. “The move from numerical simulation to deep learning represents a similar leap and will unlock new levels of product performance and ways of practicing engineering itself.”

According to PhysicsX, its software can speed up not only simulations but also other parts of the engineering workflow.

The company promises to help engineers more quickly identify ways of enhancing an existing product to improve its usefulness and reliability. Additionally, PhysicsX says its software eases the prototyping phase of hardware projects. Using the platform, engineers can more quickly correlate test data collected from a product prototype with information that was generated during simulations of the product’s design.

PhysicsX will use the proceeds from its funding round to enhance its software. The company will also invest a portion of the capital in growing its customer base, which already includes multiple enterprises’ engineering teams. It says customers are using its platform to design wind turbines, aircraft engines, chips and a range of other products. 

Photo: Unsplash