Why China is raising lobsters

You might’ve noticed pics of folks in China queuing up to install the ‘OpenClaw’ agentic AI on their devices, with users labelling it “raising lobsters” (the logo is a red lobster).

And because we can’t resist the pull of a) great lingo, and b) a big queue, we dove in.

Created by Austrian programmer Peter Steinberger, OpenClaw is a free, open-source AI agent that can complete tasks across several apps autonomously. For example, it could check and respond to your emails or reschedule your meetings. But much like a real assistant, OpenClaw needs access to your devices, apps, and personal data. 

Which brings us back to those folks queuing up in China. 

While OpenClaw is a global sensation, nowhere has this new lobster splashed down as dramatically as in China.

In the private sector, local tech giants like Tencent and JD.com have launched compatible apps and even organised events to teach folks how to use it, with the resulting buzz even triggering a brief but sizeable rally among China’s beleaguered tech stocks.

And local governments, already pumping AI to score brownie points with an AI-focused Beijing, have also got amongst it — at least four municipalities have now offered subsidies for companies building atop OpenClaw’s open-source structure.

Akin to last year’s DeepSeek craze (how is that already last year), it’s all fuelled another kind of AI gold rush for the ‘next big thing’.

Can you feel that ‘but’ coming? Of course you can, because this is China, where the ruling Communist Party always stands ready with a ‘but’, locked and loaded.

The first one dropped just last week, when authorities banned lobsters on official work devices, citing OpenClaw’s “extremely weak default security configuration”. What’s the issue? There are already some wild, real-life examples:

• A software engineer in North Carolina regretted granting his lobster iMessage access after it went rogue and started texting his wife and random contacts.

• Another lobster-owner apologised after his AI agent got angry at a forum moderator and autonomously published a hit piece to get the human fired!

So given those Black Mirror vibes, a bit of caution makes sense.

But there’s another ‘but’ in the chamber, because the Party’s security worries inevitably go beyond a keyboard-happy AI agent. Another wild example:

• During internal research, China’s Alibaba caught its own lobster-like agent going rogue and managing to escape the firm’s own firewall to start mining crypto!

And that’s where the fear kicks in: whether a rogue lobster is texting, defaming, or crypto-mining, the common thread is a loss of control — a nightmare for authoritarians. China guru Bill Bishop put it like this: “I’m surprised the Chinese authorities don’t think that this is maybe an NSA op”. The NSA has pulled off bigger stunts.

Anyway, it all raises a big question: why is Beijing only banning lobsters on official phones?

Three main reasons.

First, it’s early days. OpenClaw only dropped six months ago, and didn’t really gain traction in China until a few weeks ago. 

Second, the frenzy has brought local benefits (at least for now), while helping cement China’s image as an AI early-adopter.

But third and most importantly, Beijing has made clear (just last week) it sees widespread AI integration as the key to unlocking economic and technological supremacy.

So to achieve that goal, the Party relies on a well-tested modus operandi: welcome cutting-edge developments, study them, learn the correct lessons. Then… regulate the heck out of them to ensure they strengthen — not undermine — central authority.

By our calculations, we’re now at phase ~two. A phase-three crackdown can’t be far off.