NVIDIA GTC: Company Profile
NVIDIA positions itself as a full-stack AI systems integrator for robotics at GTC 2026, consolidating Isaac, Omniverse, and fleet networking around its Physical AI platform.
NVIDIA’s Physical AI Platform Consolidates Around Isaac, Omniverse, and Fleet-Scale Networking at GTC 2026
NVIDIA used its GTC 2026 conference in San Jose (March 16–19) to advance a coherent argument: that the company is no longer a GPU vendor but a full-stack AI systems integrator for physical environments. The dedicated “Physical AI and Robotics” track, Jensen Huang’s keynote on a “five-layer infrastructure stack,” and the presence of Agility Robotics’ CTO as a featured speaker collectively signal that NVIDIA is positioning its integrated toolchain — Isaac, Omniverse, CUDA-X, Jetson, BlueField DPUs, and InfiniBand/Spectrum-X networking — as the default development and deployment infrastructure for embodied AI. Whether that positioning translates into material robotics revenue in the near term remains an open question.
Business Overview
NVIDIA’s robotics and autonomous systems activity sits within its Compute & Networking segment, which is dominated by data center GPU revenue estimated at approximately 90% of total company revenue by third-party analysts (MODERATE CONFIDENCE — not separately disclosed in SEC filings). Robotics-specific revenue is not broken out, making it structurally difficult to track commercial traction versus strategic signaling.
The company’s robotics business model operates on two levels. At the silicon and systems layer, Jetson embedded compute modules serve as the primary on-robot inference platform for perception, SLAM, and policy execution. At the platform layer, Isaac and Omniverse provide simulation-to-real transfer tooling, while CUDA-X libraries supply optimized primitives for perception and planning. The BlueField DPU and InfiniBand/Spectrum-X networking stack addresses fleet-scale orchestration — telemetry, over-the-air updates, teleoperation fallback, and safety sandboxing.
The Inception startup program and Developer Program function as demand-generation infrastructure, seeding the ecosystem with NVIDIA-native engineering talent and reducing switching costs for early-stage robotics companies.