The AI assistant named 'Lobster,' OpenClaw, has taken the world by storm. This open-source tool possesses low-level computer operation permissions and can act as a digital avatar, independently completing complex tasks such as reading news, sending emails, writing code, and even conducting online transactions. Within weeks of its launch, it became the fastest-growing open-source project in GitHub history and quickly prompted domestic large model manufacturers like Kimi and MiniMax to adapt. From assisting in conversations to autonomous execution, the boundaries of AI capabilities have achieved another leap forward.
On March 6, nearly a thousand developers and AI enthusiasts lined up outside the Tencent Building. Within an hour starting from 10 a.m., hundreds of reservation numbers were fully distributed. These developers and AI enthusiasts were there to complete the cloud installation of OpenClaw.
OpenClaw, an AI agent that can be installed on personal computers, has gained global popularity since the beginning of this year. On the code hosting platform GitHub, OpenClaw's star count (favorites) surpassed that of Linux, making it the most popular foundational software project on GitHub. This AI assistant also brought attention to the Mac mini desktop, which is deployable hardware. Due to OpenClaw’s need to leverage large language model capabilities, providers of such models and cloud service vendors actively worked on complementary features to attract users.
Ask Futubull AI:What would you use OpenClaw for?
How did OpenClaw come about?
The developer of OpenClaw is a retired programmer from Austria. After achieving financial independence and enjoying a peaceful life, he noticed there was no AI on the market capable of handling tasks via messaging. In his spare time, he created an AI assistant by hand and uploaded it to GitHub.
As a result, a red lobster named OpenClaw became an internet sensation.
Here’s how it works: First, you grant this lobster full operational permissions on your computer, then 'nurture' it so that it understands your preferences, behaviors, work patterns, and expands its learning. Subsequently, this lobster becomes your digital avatar—reading news for you, sending emails, writing code, posting rental listings, and even spending and earning money on your behalf.
This time, AI isn’t just assisting you—it’s acting on your behalf. 'The intensity of evolution is beyond imagination,' investors frankly admitted in response to the lobster phenomenon.
Often, a brand-new future arrives unexpectedly.
The trend of nurturing lobsters went viral, while Kimi and MiniMax saw surging traffic.
Two aquatic creatures have captivated the AI community: a blue whale and a red lobster.
DeepSeek, marked by the blue whale, once sparked a wave of democratized computing power, turning what was once an expensive high-end reasoning capability into an affordable commodity, providing everyone with a free yet powerful digital advisor.
OpenClaw, branded with a red lobster logo, represents another kind of disruption.
It is a digital executor with underlying computer permissions. After being trained on private data provided by 'shrimp farmers,' a simple command can allow this lobster to take over your mouse and keyboard, navigate across various app barriers, and get things done for you. It can even perform tangible commercial actions, such as making payments or generating income on your behalf.
The arrival of 'digital workhorses' has undoubtedly excited humanity.
How popular is OpenClaw? Within weeks of its launch, it swept through GitHub, becoming the fastest-growing open-source project in history. The Mac mini, an Apple device capable of running OpenClaw efficiently, saw a significant surge in sales as a result.
Major model developers are even more thrilled—whoever can deeply integrate with OpenClaw could become a global traffic gateway. Particularly in the Chinese market, companies like MoonShot and Minimax have launched Kimi Claw and MaxClaw, enabling quick access to OpenClaw-related functionalities. Meanwhile, Zhipu introduced AutoGLM-OpenClaw, which runs in the cloud.
As a result, during OpenClaw's explosive popularity, Chinese model developers surged to the forefront of global platform usage metrics.
Of course, this craze has also reached the investment community, where project enthusiasts exclaim, 'With all this buzz, we must investigate.' Not only AI investors but also those involved in embodied robotics and life sciences are studying how to 'raise shrimp.' Some investors have even posted publicly: 'Looking for someone who can help me raise lobsters.'
"Looking for someone who can help me raise lobsters."
Looking ahead, a crayfish farming industrial chain is quietly taking shape. The installation of OpenClaw is not straightforward for ordinary users, which has led to the emergence of paid on-site installation services ranging from hundreds to tens of thousands of yuan. Just today, Tencent officially announced: Forget 998 yuan, forget 448 yuan—today, free on-site installation of OpenClaw is available right at the entrance of Tencent's headquarters.
"Everyone is into crayfish farming; not participating makes me anxious," a friend frankly admitted. Legal professionals have also noted that previously, people in their circle weren’t particularly interested in many AI tools. However, after OpenClaw gained popularity, some individuals in the legal community began experimenting with it voluntarily. Fu Sheng, Chairman of Cheetah Mobile, publicly endorsed OpenClaw, claiming that within 14 days, he successfully built a super AI team capable of operating 24/7 autonomously.
Whether you are ready or not, whether you agree or not, the AI revolution has arrived in the era where everyone farms shrimp.
A financially independent engineer embarks on a surreal journey.
Tracing its origins, the creation of OpenClaw carries a strong sense of personal heroism.
The founder, Peter Steinberger, wrote the following in his personal profile on Github: "Came back from retirement to mess with AI. Previously: Founder of @PSPDFKit."
You’ve likely sensed it—this is the story of an entrepreneur who retired after achieving success but returned to the playground for AI.
Steinberger hails from a small town in Austria and developed an extraordinary interest in programming during his youth. He later enrolled in the computer science program at the Vienna University of Technology. Naturally, after graduation, he chose to venture into Silicon Valley startups.
At that time, shortly after the release of the first-generation iPad, Steinberger found PDF reading on the iPad unsatisfactory, so he created his own PDF integration tool. Riding the wave of digitalization, he resigned and founded PSPDFKit. This venture proved successful, attracting notable clients such as Apple, IBM, and Volkswagen.
"During those years, I was always operating at high speed, and running the company was exhausting," Steinberger once recalled about this entrepreneurial journey.
In 2021, he made a decisive exit — selling the majority stake of PSPDFKit, which he had operated for over a decade, for 100 million euros. In the public's eyes, this engineer has undoubtedly achieved financial independence.
However, it is likely that even he did not anticipate that a few years later, his inherent passion for programming would spark an even grander opportunity.
By 2025, outstanding open-source AI large models will have matured, and a series of code generation tools will be undergoing internal testing. For entrepreneurs in AI applications, this is an era with exceptionally abundant resources.
Steinberger also experiments with AI. A fundamental idea arose within him: could I interact with a computer via WhatsApp and have AI assist me with tasks? A familiar scenario unfolded once again. He built an AI assistant himself, receiving messages on WhatsApp, invoking the terminal AI programming assistant Claude Code, and then returning the results.
One person, one hour of manual work, completed.
Thus, this Agent was born: by providing this Agent with tools and full computer access, it can devise solutions independently, even if you have no programming experience. After updates and optimizations, at the end of November 2025, Steinberger, a proponent of open source, officially launched this project on Github: Clawdbot (renamed to OpenClaw).
The explosive moment arrived.
“The AI tsunami is just beginning.”
Amidst the celebration, a vision of productivity restructuring is unfolding.
A friend who is an investor remarked after using OpenClaw for a month: “I am more convinced that this AI tsunami is just getting started. The intensity of this wave, conservatively estimated, is five times that of the mobile internet, and its commercial impact may be tenfold or more.”
This is not difficult to understand. For many years, closing a business loop required initial financing, followed by a lengthy process of hiring and building a suitable team, along with the continuous payment of significant operational costs. However, under the disruptive influence of agents like OpenClaw, there now exists a group of tireless digital workers capable of handling data around the clock, managing content operations, and even controlling physical devices.
In other words, one person with a few computers constitutes a fully staffed digital workforce equivalent to a commercial fleet.
But when everyone can command a commercial fleet, how will the future landscape of competition and investment logic change?
For classical internet giants that once dominated user attention through isolated app ecosystems, the test of the old order looms large: When agents communicate directly with each other and humans interact seamlessly with agents, will traditional business models reliant on traffic entry points still be effective?
Before the arrival of the new world storm, underlying risks and hidden currents are equally surging.
While users excitedly hand over control of their mouse and keyboard, they are also surrendering the trump cards of the digital world—OpenClaw requires access to core permissions such as email, calendars, and system commands. Recently, top-tier investment institution Institutional Investor explicitly noted: OpenClaw demonstrates the capabilities of AI agents, but critical security vulnerabilities, a lack of governance frameworks, and unpredictability are the norm.
Nevertheless, Pandora’s box has already been opened.
The OpenClaw phenomenon foreshadows a signal: AI is transitioning from dialogue to execution, bringing us to a pivotal moment where the democratization of large model intelligence converges with the empowerment of agent-driven action.
The door has been slowly pushed open, and what awaits us behind it may not only be a liberation of productivity but also a brutal survival-of-the-fittest chapter in the history of human commerce.
Editor/Stephen
