NVIDIA AI Updates: May 5, 2026

1. NVIDIA Wraps cuOpt as Agent Skills for Supply-Chain Decisioning

NVIDIA. A new developer-blog post walks through using cuOpt as a set of agent skills inside an LLM-driven supply-chain workflow. The pitch: classical OR teams typically need weeks to formulate, solve, and re-solve mathematical models when demand, costs, or capacity shift; exposing cuOpt as a tool an agent can call lets the system re-plan in the loop as conditions change. The post is positioned for teams already running agentic stacks and wanting GPU-accelerated optimization as one of their tools, rather than as a one-off OR project. Source

2. Jensen Huang: AI Is “Creating an Enormous Number of Jobs”

NVIDIA. Jensen Huang pushed back publicly on the AI-displaces-workers narrative, arguing that AI is producing jobs faster than it’s eliminating them. His framing is consistent with NVIDIA’s interest — every AI job ultimately runs on GPUs — but the comments are notable because Huang is increasingly the face the policy world reaches for when asking what happens to the labor market. Worth reading alongside Stuart Russell’s same-day testimony at the Musk-OpenAI trial, which took the opposite tack. Source