.jpg)
San Jose, CA — 16 March 2026 — At NVIDIA GTC 2026 today, PhysicsX announced a collaboration with NVIDIA to advance open standards for physics AI architecture, aiming to bring greater consistency, interoperability, and faster adoption to engineering and manufacturing workflows across advanced industries.
Why Now
Engineering teams globally are increasingly embedding physics AI open models into their daily workflows to simulate complex physical systems. But fragmented implementations, inconsistent architectural definitions, and non-standard pipelines are creating avoidable complexity — making it harder for the ecosystem to build and deploy production-grade engineering systems at scale.
PhysicsX is among the growing number of teams and startups that have adopted NVIDIA's open models and platforms, including NVIDIA PhysicsNeMo, to power their physics AI development. We believe that establishing shared architectural standards is the catalyst that will move physics AI from fragmented innovation to industrial-scale impact.
The Shared Vision
As part of the collaboration, PhysicsX will work with NVIDIA to define and validate a common architectural “language” for physics AI — including shared evaluation approaches and benchmarks — so that models can be expressed and compared more consistently across domains, workflows and infrastructure.
PhysicsX’s modular modeling framework, Opora, will contribute reference patterns and building blocks to support this effort. Opora provides a composable library of rigorously tested modeling primitives — spanning neural operators, graph neural networks, integral transform–based architectures, and multi-resolution models — that teams can assemble into bespoke architectures within a unified, reproducible environment.
Opora also supports a configuration-driven approach, enabling model architectures to be defined, shared, and reproduced more consistently — a key requirement for any industry standard to be adopted at scale.
Looking ahead, PhysicsX has an ambition to make these standards broadly available through open-source, ecosystem-oriented tooling —including in and around NVIDIA PhysicsNeMo and other NVIDIA physics AI open models and frameworks.
“We’re entering the era of Large Physics Models,” said Jacomo Corbo, CEO and Co-Founder of PhysicsX. “The capabilities of physics AI are advancing quickly. Models are growing bigger and more capable, covering larger engineering domains. But without shared architectural conventions for how these models are built and composed, the ecosystem risks fragmenting before it can scale. By collaborating with NVIDIA, including through PhysicsNeMo, and contributing modular frameworks like Opora, we’re working toward a common foundation for physics AI, setting open standards for industry and academia to build, evaluate, and deploy models with greater consistency and speed. This is ultimately a building block for faster innovation, and is just one of the ways through which we’re helping hardware development move at the speed of software.”
"AI-physics is critical for addressing rapidly increasing engineering and design complexity across manufacturing industries," said Tim Costa, VP and GM of computational engineering at NVIDIA. "NVIDIA PhysicsNeMo and PhysicsX's Opora frameworks give engineers a direct path from physics-grounded AI models to real-time design and engineering, compressing weeks or months of simulation into instant insights."
The Road Ahead: Large Physics Models
This collaboration represents the ambition to build the critical infrastructure for the era of Large Physics Models (LPMs) — AI models and systems that generalize across components, physics domains, and system scales.
Developing, fine-tuning, and deploying LPMs reliably across global industrial organizations requires more than capable models. It demands shared architectural conventions, interoperable tooling, and scalable pipelines. Establishing these shared foundations is a central objective of this collaboration.
An Open Effort
PhysicsX is shaping the software that will determine how bold, transformative industrial projects are designed and delivered in the decades ahead, powered by NVIDIA technologies.
Achieving that future demands shared standards and collective action. PhysicsX is advancing this as an open initiative and invites ecosystem partners leveraging NVIDIA technologies — from software developers and researchers to industrial practitioners — to help define, refine, and accelerate the next generation of physics AI standards.
About PhysicsX
PhysicsX is a physics AI company on a mission to accelerate innovation and overhaul what engineering and manufacturing look like today. The company is building a new software stack to deliver deep AI enablement across the entire engineering lifecycle. PhysicsX partners with leading organizations in aerospace & defense, automotive, semiconductors, materials, and energy & renewables, supporting them on some of their most critical and complex challenges. PhysicsX is headquartered in the United Kingdom, with offices in London and New York. For press inquiries, contact press@physicsx.ai.