What AI Physics Simulation Is and Why It Matters
AI physics simulation is the use of machine learning models trained on physical systems to replace or accelerate traditional numerical solvers, compressing tasks like structural analysis, fluid dynamics and thermal modeling from days or months into seconds while keeping engineering-grade accuracy. This new class of engineering software sits between general-purpose AI and classic computer-aided engineering, targeting the most time-consuming simulation steps in industrial design. Instead of running a single high-fidelity model at great computational expense, engineers can explore large design spaces on demand, test thousands of variants and integrate simulation into everyday decisions. The goal is not only speed, but also to change how hardware is conceived: shorter feedback loops, more automated design workflows and a tighter connection between simulation, manufacturing and operations across aerospace, automotive, energy and electronics.
PhysicsX Shows Investor Appetite for AI-Native Engineering
PhysicsX has become one of the clearest signals that AI physics simulation is now a core theme in engineering software funding. The company raised USD 300m (approx. RM1,380,000,000) in an oversubscribed Series C round, bringing its valuation to about USD 2.4bn (approx. RM11,040,000,000) and total funding to around USD 500m (approx. RM2,300,000,000). The round was led by Temasek with backing from Applied Materials, Nvidia, Atomico, General Catalyst, Siemens and others, underlining strong institutional confidence in AI-native industrial tools. Founded by former Formula 1 engineers, PhysicsX focuses on compressing design and simulation cycles for sectors such as aerospace, defence, automotive, semiconductors, materials and energy. “We are giving engineers the ability to explore thousands of designs where they once managed a handful, in seconds rather than weeks, across the most demanding industries in the world,” said co-founder and CEO Jacomo Corbo.
NP Company Bets on Transformer Physics Models for Industry
NP Company, a spin-out from Inria and Paris-Saclay research, represents the emerging wave of transformer physics models aimed at industrial design automation. The startup raised a €6 million pre-seed round led by Partech with participation from the Peugeot family office and angels including Mistral AI co-founders Guillaume Lample and Cédric O. NP Company trains transformer-based physics models on industrial data, adapting the architecture popularised by large language models to simulation tasks in aerospace, defence, energy, electronics, data centres and automotive. Unlike earlier AI simulators that needed extensive customer-specific training, its pre-trained foundational models are designed to deliver value on deployment. The company reports speedups of up to 1,000 times on industrial benchmarks, shrinking simulation runs from days or weeks to seconds while maintaining fidelity across entire assemblies. This approach positions NP Company to extend from fast simulation into automated design and real-time operational digital twins.
From Bottleneck to Accelerator in Industrial Design Workflows
Both PhysicsX and NP Company are targeting the same choke point in engineering: traditional simulation workflows that slow product cycles and inflate R&D costs. In aerospace, automotive and advanced manufacturing, teams currently queue for compute resources and specialist time to run a handful of high-fidelity models per design iteration. By replacing or augmenting solvers with AI physics simulation, these startups aim to turn that bottleneck into an always-on accelerator. Engineers can run thousands of what-if scenarios, optimise for weight, efficiency or thermal performance and integrate feedback earlier in the design process. PhysicsX speaks of deep physics AI enablement across the engineering lifecycle, while NP Company highlights transformer-native models that work across entire assemblies. The practical promise is shorter time-to-market, more ambitious designs and a closer alignment between digital models and real-world performance, especially in complex systems like aircraft, engines and power infrastructure.
Why Funding Is Flowing Beyond Language Models
The recent funding surge into AI physics simulation signals that investors see value in domain-specific AI, not only in general-purpose language models. PhysicsX’s USD 300m (approx. RM1,380,000,000) Series C and NP Company’s €6 million pre-seed round show that capital is moving toward tools that solve concrete engineering constraints: simulation speed, limited expert capacity and growing product complexity. These companies position themselves as AI-native infrastructure for the physical economy, rather than productivity add-ons. Their platforms promise direct impact on capital-intensive decisions such as aircraft design, chip cooling strategies or energy system layouts, where each iteration costs time and money. As they scale larger "large physics models" and transformer-based simulators, the competitive landscape in industrial design automation is likely to shift. Incumbent engineering software vendors may need to integrate or compete with these AI-first approaches to remain central to future hardware innovation.
