PhysicsX Closes $135M to Accelerate Engineering Simulation

The big picture: PhysicsX, a London-based startup, has closed a $135M Series B round led by Atomico. The funding will scale its AI software, which accelerates complex engineering simulations from weeks to minutes. The company’s total funding now stands at nearly $170M.

Why it matters:

• Traditional methods are slow: Designing complex components like aircraft engines or semiconductor chips relies on physics simulations that can take hours or days per design iteration, severely limiting the number of options engineers can explore.
• AI enables rapid iteration: PhysicsX uses deep learning to predict aerodynamics, heat behavior, and structural stress in seconds, allowing engineers to test thousands of design options in the time it once took to test one.
• Strategic validation: Major industrial players Siemens and Applied Materials have joined as strategic investors and customers, validating the technology’s potential despite their own tools sometimes competing with PhysicsX.

How it works:

• Physics surrogate models: PhysicsX employs deep learning models trained on high-quality simulation data to act as AI stand-ins for traditional, time-consuming simulation runs.
• Predictive accuracy: These models rapidly predict critical physical properties, such as heat behavior and structural stress, enabling near real-time design feedback.
• Seamless integration: The platform connects with existing engineering tools like Ansys, CATIA, OpenFOAM, and Siemens Star CCM+, allowing customers to integrate AI acceleration without rebuilding established workflows.

The catch: The market for simulation software is competitive, with established players like Ansys and Siemens, as well as emerging AI-powered competitors such as Quanscient and BeyondMath. While strategic investors Siemens and Applied Materials validate PhysicsX’s technology, their existing tools sometimes compete, posing a complex relationship. PhysicsX’s rapid revenue growth is noted, but without a disclosed base figure, its absolute scale remains unclear.