Particle.news

Nvidia Launches NemoClaw and Rubin–Groq Hardware to Secure and Speed OpenClaw Agents

Nvidia aims to turn the agent boom into safe, revenue‑grade deployments by baking guardrails into OpenClaw.

Overview

  • Nvidia introduced NemoClaw, a stack that installs OpenClaw with Nemotron models and the OpenShell runtime to enforce sandboxing, network policy and privacy routing for autonomous agents.
  • CEO Jensen Huang called agentic systems "the new computer" and said every company needs an OpenClaw strategy as agents move from demos to business workloads.
  • The Vera Rubin platform integrates Groq LPUs in new co‑processing racks and adds a Vera CPU tier to accelerate real‑time inference for multi‑agent tasks.
  • Huang forecast about $1 trillion in demand for Blackwell and Rubin chips through 2027, citing the shift from training to high‑throughput inference and continuous token use.
  • OpenAI hired OpenClaw creator Peter Steinberger, and reported Chinese state‑enterprise limits plus corporate bans over security risks are driving vendors to add enterprise guardrails and support.