I’ve been bombarded recently by news of local governments across China’s tech hubs announcing aggressive measures to build an entire industry around OpenClaw. But it’s not just official policy; it’s a total social media wave. My feed is a non-stop stream of people deploying OpenClaw, entrepreneurs launching "Lobster" businesses, and a palpable "get-in-now" energy.
It’s a repeat of the 'exaggerated effect' I've experienced myself back in China - where a trend doesn't just grow, it explodes overnight. This OpenClaw wave feels like a direct evolution of that momentum.
The Scale of the "Deployment Boom"
To understand how significant this is, look at the sheer speed of this rollout over the last week:
The Tencent Queue (Friday, 6th March 2026): Nearly 1,000 people lined up outside Tencent’s Shenzhen headquarters just to get help installing OpenClaw. (ref: South China Morning Post)
The "Lobster" Subsidies: In Wuxi (9th March), officials published a draft supporting what developers call "raising the lobster" (or crayfish), offering rewards of 500,000 yuan (~$72,000). Shenzhen’s Longgang district (7th March) went even further, offering up to 2 million yuan (~$290,000).I know your tool automates workflows efficiently. But do you realise what your tool does might be someone's entire job? Have you considered the social impact you're creating? (ref: Business Insider)
A Viral Micro-Economy
This isn't just a corporate race; it’s a grassroots gold rush. On platforms like Xiaohongshu and Xianyu, "door-to-door OpenClaw installation" is a top search term.
The Side Hustle: One Beijing programmer received over 30 orders in just one week, earning upwards of 15,000 yuan (~£1,590). (ref: 36Kr)
Niche Markets: Prices range from 300 to 1,000 yuan, with some teams even marketing an "all-female engineer team" for a premium of 1,200 yuan per session.
In my own social circle on WeChat, I see people sharing their daily order logs. It is very real, and it is moving incredibly fast.
Product Reality: Is it just Claude Code?
From a product perspective, if you’ve used Claude Code or other agentic AI, OpenClaw feels familiar. It is a general AI tool that can act on your behalf. However, its power (and its danger) comes from its deep access to local files and system permissions.
The real game-changer here isn't just the AI itself, but how it’s delivered. Unlike Claude Code, which largely requires a CLI (Command Line Interface) and a developer's mindset, OpenClaw can be deployed through standard consumer apps. This means users can interact with a powerful AI agent through the apps they use every day. By removing the technical friction of the CLI, the barrier to entry has vanished, turning a high-tech tool into a mass-market commodity.
The Pendulum Swings: Security Risks
As quickly as the hype grew, the backlash has arrived. By 12th March, a new irony emerged: the same people who charged to install OpenClaw are now charging to uninstall it. (ref: Business Insider)
China’s cybersecurity authorities have issued warnings citing serious vulnerabilities like prompt injections and operational errors. I’ll be closely monitoring how the ‘conflict’ evolves:
Local Governments pushing adoption to fuel the economy.
Cybersecurity Authorities raising red flags about cyberattacks.
Big Tech (like Tencent with its "Work Buddy") rushing to capitalise on the popularity while navigating the risk.
Final Thoughts: Individual vs. Systemic Deployment
This "lobster" craze highlights a major difference in how AI is deployed. In China, individualist deployment is being forced forward by systemic pushes - a direct result of the socialist economic model. In the UK, while AI is a policy priority, it lacks this "exaggerated" explosive effect. The economic models simply don't trigger the same level of mass, coordinated individual action.
My Takeaways:
Culture vs. Tech: Even in a tech-driven world, human culture and political systems dictate how a tool is used more than the code itself.
The Information Gap: As decisions become more complex, information gaps will always occur. It is becoming crucial for individuals to make informed decisions, as the impact of these AI tools is harder to track once they are materialised in the digital world.
The Burden of Action: This places a heavy burden on us to understand both AI products and the shifting landscape of policymaking. It is hard to justify if these moves are always in the best interest of the nation, so the responsibility falls on the individual.
Being active and open to being reactive is more important than passively waiting for the best to come. I firmly believe that this is a skillset we all need to learn and embed.
A Note on Perspective:
This exploration isn’t about judging the speed of the "Lobster" boom as right or wrong. Instead, it’s a recognition that we are at a critical inflection point where technical convenience - like a one-click AI agent - is colliding with systemic security risks. As we build and adopt these tools, the ease of the "click" doesn't absolve us of the responsibility to understand the access we are granting. The world we inherit will be shaped by how deeply we consider these trade-offs today.