OpenClaw Crosses 368K GitHub Stars as Enterprise AI Agent Debate Heats Up

OpenClaw has crossed 368,000 GitHub stars and 12 million downloads, marking a new phase as both a community darling and an enterprise flashpoint.

The big story: Reports surfaced that Anthropic’s Claude Code has been detecting HERMES.md files (the OpenClaw agent configuration manifest) and OpenClaw-related commit messages — then either refusing to process requests or routing them to higher-cost billing tiers. Users reported cost increases of up to 50x. The Hacker News post accumulated 1,336 upvotes and 718 comments, becoming the top community story of late April.

Meanwhile, enterprise adoption is accelerating on different terms:

  • NVIDIA NemoClaw shipped alpha, integrating NeMo guardrails and OpenShell sandboxes — a hardened fork that risk-averse buyers can defend in architecture reviews
  • Tencent committed full-time maintainers to the project
  • v2026.5.4-beta.1 dropped as a milestone release, cementing OpenClaw’s transition from scrappy open-source project to legitimate platform target

The collision of hosted-AI subscription enforcement, sovereign AI procurement scrutiny, and the availability of enterprise-grade hardened variants has shifted the conversation from “let’s pilot a hosted agent” to “what’s our agent control plane?” — exactly the question self-hosted platforms like OpenClaw answer.

The full analysis is available at Big Hat Group.

← Back to News