OpenClaw's Creator Joins OpenAI

Background

By late January 2026, OpenClaw had survived the rebrand saga and was accelerating. The project had crossed 100K stars by late January. But the real story was what happened next: an entire economy materialized around a single open-source CLI tool.

Over the following weeks, a wave of hosted services emerged. MyClaw.ai, ClawFast, LaunchClaw, and AgentClaw.now offered one-click cloud hosting. NanoClaw forked the project as a security-hardened alternative. Cloudflare shipped Moltworker for edge deployment. ClawHub became a growing skills registry for community-built integrations.

Within days, Chinese tech companies announced competing services. Alibaba Cloud, Tencent Cloud, and ByteDance all launched hosted OpenClaw services within the same week. Reports indicated Baidu integrated OpenClaw into its search app, putting it in front of an estimated 700 million monthly users.

Meanwhile, Steinberger was fielding acquisition offers. Meta and OpenAI made competing offers — not for the code (it was open source) but for the person who understood the architecture better than anyone.

Aftermath

On February 12, Steinberger appeared on Lex Fridman's podcast (Episode #491), where he revealed the competing offers and stated plainly: "The project must stay open source." Two days later, he announced he was joining OpenAI.

The decision carried a specific irony: OpenClaw's most popular configuration used Anthropic's Claude as its AI backend. The creator of the tool that made Claude more accessible to developers was now working for Claude's primary competitor.

The project transitioned to foundation governance. Steinberger's stated reasoning was characteristically direct: "I'm a builder. OpenAI is where I can build at the biggest scale." He drew parallels to Brendan Eich creating JavaScript and then building Firefox — the browser that ran everyone's JavaScript, not just one company's.

The day after the announcement, Moonshot AI launched Kimi Claw — a cloud-native implementation of the OpenClaw protocol integrated into kimi.com, with 5,000+ pre-installed skills and cloud storage. It was the clearest signal that the protocol had outgrown its creator.

By February 15, OpenClaw had 196,000 GitHub stars and 600 contributors. The foundation governance model was in place. The question was no longer whether the project would survive without Steinberger — it was whether the ecosystem he'd accidentally created was now too big to steer.

Industry Impact

Steinberger's departure crystallized several tensions in the AI tools landscape:

The founder paradox: Open-source projects derive legitimacy from individual vision, but the most successful ones must outlive that vision. Steinberger's exit was both a risk (loss of direction) and a validation (the project was bigger than any one person).

The platform play: OpenAI and Meta weren't bidding for code — they were bidding for protocol influence. Whoever hired Steinberger gained insight into how 196K developers preferred to interact with AI models.

The China factor: Multiple major Chinese cloud providers launched hosted OpenClaw services within the same week, demonstrating that open-source AI tooling had become a competitive front in cloud distribution strategy.

Protocol vs. product: Kimi Claw's launch proved that OpenClaw had become a protocol, not just a product. Multiple implementations could coexist, each optimized for different models and deployment contexts. This mirrors how HTTP clients proliferated once the protocol was established.