NVIDIA physical AI lifts Asian partners
- NVIDIA’s latest physical AI push — spanning robotics, simulation and autonomous systems — is lifting Asian partners including LG Electronics, Nanya, Desay SV and Pateo. - The sharpest detail is supply-chain concentration: Asian suppliers now make up about 90% of NVIDIA’s production costs, up from roughly 65% in 2025. - That matters because NVIDIA is widening from chips into factory, vehicle and robot platforms — and Asia sits inside that expansion.
Robotics is the new layer in the NVIDIA story — and that changes who benefits. For the last two years, the easy trade was chips, memory and servers. Now NVIDIA is pushing harder into what it calls physical AI: robots, autonomous machines, digital twins and software that trains systems to act in the real world. That shift is helping lift a broader set of Asian partners, not just the usual semiconductor names. ### What does “physical AI” actually mean? Basically, it means AI that does not stop at text, images or code. It has to perceive a space, model physics, plan an action and then do something in the world — move a robot arm, navigate a warehouse, run a factory process or help a vehicle make decisions. NVIDIA’s stack for that includes Jetson running embodied systems. ### What changed this spring? The big moment came at GTC in March, when NVIDIA moved from talking about the idea to naming a much wider industrial ecosystem. It rolled out new Isaac simulation frameworks, Cosmos world models, GR00T open models and a Physical AI Data Factory Blueprint. It also named partners across industrial, surgical and humanoid robotics — including ABB, FANUC, KUKA, Yaskawa, Universal a lab demo and more like a supply chain. ### Why are Asian stocks reacting? Because investors are starting to map that software-and-robotics push onto real companies in Asia. Bloomberg’s latest tally says names including South Korea’s LG Electronics, Taiwan’s Nanya Technology, and China’s Desay SV and Pateo have rallied on tie-ups, component exposure or product collaboration link-level. ### Why is LG in the mix? LG is a good example of the broadened thesis. Reuters reported on April 29 that LG Electronics said it had been discussing cooperation with NVIDIA in robotics, AI data centers and mobility. That is not just “buying chips.” It points to consumer robots, industrial systems and smart-machine platforms where NVIDIA’s compute and simulation tools can sit underneath LG hardware and services. ### Why does the 90% number matter? Because it shows how deeply Asia is embedded in NVIDIA before this next wave even fully arrives. Bloomberg says Asian suppliers now account for about 90% of NVIDIA’s production costs, up from roughly 65% in 2025. So when NVIDIA expands from training clusters into robots, factories and vehicles, the region is not starting from zero — it is already the company’s manufacturing and component backbone. ### Is this just a chip story in disguise? Not quite. Chips still matter — a lot. But physical AI adds new layers: power systems, cooling, sensors, automotive electronics, factory software, simulation tools and robot hardware. That is why companies like Delta, Foxconn and Taiwan manufacturers keep showing up around Omniverse and hyperscale data centers. ### What is the catch? The catch is that physical AI is harder to monetize cleanly than selling GPUs into an obvious compute shortage. Robots and industrial systems have longer deployment cycles. Customers need integration, safety validation and proof that simulation really transfers to messy real environments. NVIDIA is trying to solve that bottleneck with world models, synthetic data and digital-twin tooling — but this is still earlier than the data-center boom. ### So what’s the bottom line? NVIDIA’s next act looks less like “more chips” and more like a platform grab across machines that move. If that works, Asian partners do not just assemble the boom — they become part of the product.
