The rise of OpenClaw is a global phenomenon. Initially built in Austria (its creator, Peter Steinberger, now works at an American AI lab), OpenClaw has received code contributions from developers worldwide. You can find meetups in a host of major cities around the clock. Everywhere you go in technology today, you are haunted by the lobster.
Especially in China. In the home country of MiniMax, DeepSeek, Z.ai, Moonshot, StepFun, and other AI labs, OpenClaw is exploding in popularity. OpenClaw-focused gatherings in the country are attracting hundreds of people, local governments are considering grants and other financial support to drive adoption of the tool, and today we learned that Chinese tech giant Tencent has released a version of OpenClaw that gives the agent access to WeChat and QQ, and even built out support for OpenClaw in its cloud products.
Sure, OpenClaw is popular in every city where tech folks gather, but what's driving its rabid adoption in China? Jason explained his thesis on the podcast recently:
Why does it matter if OpenClaw adoption in China seems as strong as what we're seeing here in the United States? The era of software products written for humans first and machines second is rapidly coming to a close.
Box CEO Aaron Levie recently argued that “agents will become the primary user of all software in the future.” China's OpenClaw fandom implies that the entire world is moving in the direction Levie forsees, and quickly.
There's risk in building for agents first and humans second in a world where humans still do most of the work. But the enthusiastic adoption of agents by corporations (Sierra didn't reach $150 million in ARR by accident) and individuals (OpenClaw is the top app on the OpenRouter inference charts) suggests we're in the twilight of humans doing the majority of typing.
Think back to Salesforce; it was radical to sell software as a managed service. Later, SaaS became the de facto business model for most of the technology world. Think back to Twilio; it was radical to sell technology on an on-demand basis. Later on-demand pricing became a common way to sell in the tech world.
There will be a startup that seems too early to build business software for agents first, and humans second, but in time it will become the norm. Call OpenClaw adoption an early-warning signal that everyone, everywhere, all at once wants their computers to do work for them. Not with them. And that means that every startup needs to welcome agents with open arms. Today, or at the latest, tomorrow.