Sponsored by: NVIDIA
From Perception to Physical AI: What Jensen Huang’s GTC Paris Keynote Means for Design Engineers
At VivaTech 2025, NVIDIA CEO Jensen Huang delivered more than just a vision for the future of artificial intelligence, he unveiled a call to action for engineers worldwide. His keynote at GTC Paris emphasised a shift from digital-only AI toward intelligent systems that exist, move, and learn in the physical world.
This shift has direct implications for mechanical designers, embedded engineers, and innovators who work at the intersection of software, hardware, and the real world.
The Three Waves of AI
Huang introduced a clear framework for understanding AI’s evolution:
- Perception AI – Focused on recognition: images, speech, and patterns.
- Generative AI – The current wave: creating text, visuals, code, and simulations.
- Agentic AI – The emerging wave: AI that reasons, plans, and acts autonomously.
“The next big thing is Physical AI, AI with a body,” Huang declared. “Robots, autonomous machines, industrial systems… It’s all coming.”
For designers and engineers, this signals a shift from designing tools to designing intelligent, interactive agents, with embedded awareness of the world around them.
Physical AI and Robotics
In a highlight of the keynote, NVIDIA showcased “Grek”, a humanoid robot developed entirely in NVIDIA’s Omniverse simulation platform. Trained using reinforcement learning, Grek could walk, grasp, and manipulate objects with high fidelity, all before ever touching the real world.
Grek - Image credit: NVIDIA
This isn’t science fiction. It’s a glimpse into the next generation of engineering workflows where:
- Simulation precedes and reduces the need for physical prototyping.
- Embedded systems are AI-native from day one.
- Digital twins operate in real-time, across edge, cloud, and GPU environments.
For designers and engineers, this means building systems that learn, adapt, and physically interact, not just execute pre-programmed instructions.
Digital Twins, Real Factories
NVIDIA is investing heavily in AI Factories across Europe, including a facility in Germany with 10,000 Grace Blackwell GPUs. Partners include Bosch, ASML, Schneider Electric, and Siemens, all using AI to optimise chip manufacturing, factory robotics, and sustainable facility design.
- Siemens is integrating Omniverse APIs into its Xcelerator platform for real-time, multi-user digital twin simulation.
- Schneider Electric is deploying these tools to design and run energy-efficient facilities.
This is not just about cloud computing, it’s a redefinition of how real-world systems are designed, tested, and operated.
Expect real-time simulation, digital-physical prototyping, and AI co-pilots to become the new normal in engineering workflows.
Hybrid AI: Classical Meets Quantum
Huang also introduced CUDA-Q, an extension of CUDA designed for hybrid quantum/classical computing. While quantum computing is still in its early days, NVIDIA sees it accelerating fast.
Their new Grace Blackwell superchip enables developers to simulate quantum algorithms on GPUs today while laying the groundwork for tomorrow’s quantum processors.
For design engineers, this raises exciting questions:
- How will quantum workflows affect simulation and modelling?
- Will complex physics, materials science, or energy modelling benefit from early quantum integration?
While still early, forward-thinking engineers should keep an eye on this space.
Full-Stack AI for Industry
NVIDIA isn’t just building chips, they’re building ecosystems:
- NeMo for training large language models (LLMs).
- Retriever for managing multimodal AI memory.
- Lepton for deploying AI agents to real-world environments.
- Omniverse Cloud for industrial digital twins.
- DRIVE AV is rolling out to vehicles from Mercedes, Volvo, and Jaguar Land Rover.
- Collaborations with Siemens, Schneider Electric, and others point to a coming explosion in industrial AI applications.
This full-stack approach means tools used in mechanical design, control systems, and embedded development will increasingly need to interface with these AI workflows.
What This Means for Design Engineers
At RS, we believe that designing the future means being part of it early. Jensen Huang’s keynote underscores the need for hardware-savvy, systems-literate engineers who can bridge design with intelligent behaviour.
Here’s how you can prepare:
- Explore digital twin workflows using tools like DesignSpark Mechanical + simulation platforms.
- Start integrating AI into embedded projects.
- Think physical, design for motion, sensing, adaptation.
- Stay quantum-aware, especially for high-fidelity simulation or next-gen compute workloads.
- Design with integration in mind, your CAD designs may soon feed directly into real-time AI systems.
A Front-Row View from DesignSpark
Joydipto Choudhuri, Design Tools Experience Manager at DesignSpark, attended the event and shared his thoughts:
“At GTC Paris 2025, I had a front row seat to exciting breakthroughs in AI hardware and software from NVIDIA. What used to be science fiction is now reality with supercomputers for your desktop and quantum computing being used to find cancer treatments. As an electronics engineer, I see NVIDIA truly pushing the envelope with AI superchips that are outpacing Moore's law, a guiding principle for the semiconductor industry in the past 60 years. The future is surely now!”
Joy’s insight echoes what many engineers in the audience felt: this is not just an evolution in computing, it’s a fundamental shift in the role of the engineer.
The RS Team at GTC.
Final Thoughts
Jensen Huang didn’t just present a vision of AI, he presented a challenge to the global engineering community: build it.
From robotics to quantum workflows, and from chips to entire industrial stacks, the message is clear: the next generation of systems won’t just be smart, they’ll be alive in the real world.
For Design Engineers, this is your call to action.
Want to dive deeper into AI? Check out our AI Hub.
Are you looking for advice and support for AI projects? If so, get in touch; We may be able to help.
Comments