The 2026 “Lobster Craze” serves as a stark reflection of the tech industry’s impatience and utilitarian tendencies, as well as the underlying anxiety and disorientation felt by small and medium-sized enterprises (SMEs).

In March 2026, Shenzhen’s Tencent Building North Plaza garnered more attention with its hundred-meter-long queue than with its post-Chinese New Year work resumption red envelopes. People clutched number tickets, not for a celebrity event, but to wait for engineers to install “Little Lobster”—the open-source AI agent OpenClaw—on their computers.

Meanwhile, in Beijing’s Zhongguancun cafes, programmer Xiao Li’s posters advertising “On-Site Lobster Installation, 500 RMB per Session” littered the tables. On a knowledge-payment platform, a course titled “Master OpenClaw in 7 Days, Earn 100,000 RMB Monthly” was priced at 998 RMB and sold over 1,000 copies in just three days. Kimi’s backend data revealed that its K2.5 version’s revenue in the past 20 days surpassed its entire 2025 annual total.

Dubbed “lobster farming,” this AI frenzy has given rise to three distinct wealth-creation groups: course sellers, installation service providers, and token traders. With GitHub stars surging to 260,000 in just four months and Apple’s Mac mini selling out as the “ultimate lobster tank,” questions loom: Is this digital lobster a productivity savior for SMEs, or merely another get-rich-quick scheme masquerading as technological innovation? Are the myths of “one-person companies” genuine AI dividends of the era, or deliberately amplified entrepreneurial illusions?

01 Gold Rushers and Shovel Sellers: The Three Types of Wealth Creators in the Lobster Craze

At its core, “lobster farming” involves deploying OpenClaw, an AI agent capable of performing “hands-on” tasks. Unlike conversational AI like ChatGPT, OpenClaw executes practical operations—such as data retrieval, report generation, and cross-platform management—via natural language instructions, earning it the moniker of a “digital employee.” However, before this technological revolution could fully materialize, get-rich-quick schemes had already taken flight, with three distinct player types rapidly carving up the profit pie.

The first group, the “course hawkers,” capitalizes on anxiety. They may lack AI expertise but excel in traffic generation tactics. On platforms like Douyin and Xiaohongshu, headlines such as “Learn OpenClaw from Scratch, Go from Unemployed to Successful” or “Leverage Lobster AI for Self-Media, Publish 10 Viral Posts Daily” abound. A tech blogger confessed that his course content merely translated GitHub’s open-source documents into Chinese, supplemented with basic demos. “Cost: under 50 RMB. Price: 998 RMB. Pure profit from information asymmetry.” More outrageously, some instructors launched “Enterprise-Level Lobster Application Courses” without understanding deployment steps, claiming to teach SMEs “cost reduction and efficiency gains” through AI.

These courses follow a predictable script: amplify anxiety and fabricate illusions. They repeatedly warn, “Not farming lobsters means obsolescence,” using fabricated success stories like “A student earns 50,000 RMB monthly via lobster-powered store operations” to spur enrollments. A traditional manufacturing boss who purchased a course for 3,999 RMB said he only learned to generate production reports with AI. “Useless for our assembly lines. The ‘customized solution’ was just a template with minor edits.”

The second group, the “box builders and installers,” earns their money the hard way. Due to OpenClaw’s technical barriers, demand surged for installation services and customized “lobster boxes” (hardware preloaded with OpenClaw). College student Han Hang earned nearly 6,000 RMB in a week from 40 orders at 150 RMB each, treasuring his first 100 RMB cash red envelope. In Shenzhen’s Huaqiangbei, merchants quickly launched “lobster-exclusive hosts,” marking up ordinary industrial computers from 2,000 RMB to 8,000 RMB with red lobster stickers.

Some even upgraded “installation services” to “managed operations.” A Beijing startup offered “Lobster Butler” services, promising 24/7 AI support for SMEs at 3,000 RMB monthly. In reality, they relied on fixed instruction templates. When complex issues arose, AI “failed,” and manual, perfunctory responses followed. The “260,000 RMB in days” legend benefited few; most installers were temporary gig workers. Once the hype faded, their businesses collapsed.

The third group, the “token traders,” profits effortlessly. As an open-source agent, OpenClaw itself generates no revenue but requires large language model (LLM) computing power, priced in tokens. This “lobster farming” boom essentially feeds traffic and revenue to LLM firms.

Tech giants like Tencent, Alibaba, and Baidu swiftly launched “one-click OpenClaw deployment” services, tying agents to their LLMs. Startups like Kimi and Zhipu raised token package prices amid the hype, with Zhipu AI’s GLM Coding Plan surging by 30%. Commercially, tokens are the “kilowatt-hours” of the AI era, with LLM firms acting as “power companies.” Regardless of how lobsters are “farmed,” users must pay “electricity bills” to them.

Notably, domestic and international token prices differ vastly. MiniMax M2.5 costs $0.3 per million tokens for input, while the U.S.’s Claude Opus 4.6 charges $5—a 17-fold gap. This gives Chinese LLM firms a cost advantage, further inflating the “lobster economy” bubble.

02 Overheated Frenzy: Can Lobsters Solve SMEs’ Real Problems?

As the “lobster farming” trend intensifies, many SME owners jump on the bandwagon, hoping this digital lobster will resolve their survival crises. Yet, most end up with “ineffective feeding.”

Visits to 10 SMEs in Beijing and Zhejiang (spanning manufacturing, retail, and services) revealed a stark gap between OpenClaw’s hype and reality. A Hangzhou clothing e-commerce boss spent 8,000 RMB on a “lobster operations box,” expecting AI to handle product selection, customer service, and copywriting. Instead, he faced constant setbacks. “AI-selected styles were outdated; customer service replies were irrelevant.”

The core issue lies in the mismatch between OpenClaw’s “capabilities” and SMEs’ “needs.” As an open-source agent, OpenClaw excels at standardized, routine tasks like data organization and document generation. However, SMEs often require personalized, non-standardized solutions.

Manufacturers need AI to integrate production lines and optimize supply chains, but OpenClaw cannot process real-time industrial equipment data. Restaurants need AI to adjust menus based on foot traffic and optimize inventory, but OpenClaw has near-zero offline scene awareness. Export firms need AI to navigate international market changes and mitigate trade risks, but OpenClaw’s geopolitical analysis lags far behind professional industry analysts.

More critically, SMEs’ core challenges are survival issues—not efficiency. Under economic pressure, shrinking orders, tight cash flows, and fierce competition are the “three mountains” crushing SMEs. A digital lobster handling standardized tasks cannot secure orders, resolve funding issues, or build core competitiveness.

“We don’t reject AI; we need AI that solves real problems,” said a Beijing machinery manufacturing boss. “I spent 50,000 RMB on OpenClaw deployment but only used it to generate meeting minutes. That money would’ve been better spent on worker bonuses.”

Beyond the mismatch, costs pose another hurdle. OpenClaw’s “open-source and free” facade hides hefty hidden expenses. Hardware costs: High-performance computers/servers needed to run OpenClaw cost thousands of RMB. Token costs: Monthly token fees for AI task processing can reach hundreds of RMB. Learning costs: Training employees to use OpenClaw requires time and effort.

For SMEs with slim margins, these costs exacerbate struggles. A small design firm founder calculated: 5,000 RMB for hardware, 800 RMB monthly for tokens, and 2,000 RMB for training—totaling nearly 15,000 RMB annually. “Our annual profit is around 100,000 RMB. Spending 15,000 RMB on a ‘useless lobster’ isn’t cost-effective.”

More alarmingly, the “lobster craze” fuels industry chaos. Scammers exploit it to peddle fake AI services to SMEs, extorting franchise and service fees. A Zhejiang furniture factory boss was duped by a “Tencent-authorized lobster service provider” into paying 20,000 RMB in franchise fees for promised “AI-powered production management,” which turned out to be an unusable software template.

As more SMEs realize “lobsters are useless,” the bubble bursts. Knowledge-payment platforms show a 70% drop in “OpenClaw course” sales by mid-March. Huaqiangbei’s “lobster boxes” went from scarce to gathering dust.

03 The Truth About One-Person Companies: AI Empowerment Myth or Exaggerated Illusion?

The “lobster craze” touts “one-person companies” as its most alluring promise. Slogans like “With OpenClaw, one person can run a company” or “AI as employee, boss as absentee landlord” fuel entrepreneurial dreams.

Yet, “one-person companies” are neither new nor AI-created myths. Under China’s 2024 revised Company Law, “one-person limited liability companies” are now treated equally to ordinary firms, retaining only the core rule that shareholders must prove company assets are separate from personal assets to avoid joint liability for debts. This means one-person companies have long been legally valid, and their rise’s link to AI is overstated.

Legally, their key advantage is “limited liability”—shareholders are liable only for their subscribed capital contributions, unlike sole proprietorships’ “unlimited liability.” However, this protection hinges on strict separation of company and personal assets. Many entrepreneurs lack financial discipline, commingling funds (e.g., using company accounts for personal expenses or vice versa). In debt disputes, shareholders risk joint liability, shattering the “limited liability” shield.

Operationally, AI can empower one-person companies, enabling “solo armies.” OpenClaw can generate copy, process financial reports, and handle customer inquiries, saving labor costs. But it cannot handle all work alone.

A successful company requires product R&D, marketing, client maintenance, and risk management—AI’s weaknesses. In marketing, AI can create promotions but cannot grasp consumer needs or handle PR crises. In risk management, AI can analyze market data but cannot respond to geopolitical shifts or policy changes.

An entrepreneur I know quit his job to start a one-person AI copywriting business amid the “lobster craze,” believing OpenClaw would let him “work by just talking.” Reality: AI-generated copy varied in quality, requiring manual edits; clients’ personalized needs demanded direct communication; and he handled finances, taxes, and legal work alone. “Working 16 hours daily, I’m more exhausted than as an employee. Earnings are lower than my salary,” he lamented.

More importantly, one-person companies succeed due to founders’ abilities, not AI tools. Bill Gates (Microsoft) and Mark Zuckerberg (Facebook) started as one-person firms, but their success stemmed from technical prowess, business acumen, and execution—not contemporary tools. AI is merely auxiliary, not a substitute for core entrepreneurial skills.

The “lobster craze” frames one-person companies as “low-barrier, high-return” shortcuts, but their actual barriers are high. Beyond personal skills, founders face rent, utilities, taxes, market competition, and client retention risks. For most, one-person companies are “risk traps,” not shortcuts.

Statistics show under 10% of China’s one-person companies survive, far below ordinary firms. Most fail due to lack of business acumen, risk awareness, and resilience—not AI tool shortages. The “lobster craze” merely cloaks one-person company illusions in technological garb.

04 Post-Frenzy: A Rational Return to AI Entrepreneurship

From the metaverse to ChatGPT and now OpenClaw, the tech industry never ceases chasing wealth frenzies. Each technological revolution spawns get-rich-quick groups and entrepreneurial illusions. But after the frenzy, rationality must prevail.

For course sellers and installation service providers, the “lobster craze” dividends will fade. As users gain awareness, fake courses and services will exit the market. When OpenClaw deployment becomes easier, installation services will lose relevance. Only those deeply invested in technology and quality services will survive industry shakeouts.

For LLM firms, token revenue cannot mask core competitiveness gaps. China’s LLM competition remains stuck in price and traffic wars, lacking technological breakthroughs. The “lobster craze” brought short-term gains but lured firms into “marketing-over-R&D” traps. Future LLM success hinges on solving real problems, not just selling tokens.

For small and medium-sized enterprises (SMEs), it is even more crucial to remain rational and resist following trends blindly. When selecting AI tools, they should not be swayed by promotional gimmicks but should instead evaluate their true value based on their actual needs. The digital transformation of SMEs is not as simple as 'raising a lobster'; rather, it is a systematic endeavor that requires a multifaceted approach, including strategic planning, technology selection, and talent development. Instead of blindly deploying OpenClaw, SMEs would be better off focusing on their core business and enhancing their competitiveness.

For entrepreneurs, the myth of the 'one-person company' needs to be dispelled. Entrepreneurship has never been an easy feat, and regardless of whether AI is involved, it requires entrepreneurs to possess solid skills, unwavering determination, and strong resilience. Rather than believing in the fallacy that 'AI can be your employee while you sit back and relax,' it is more prudent to stay grounded, leverage your strengths, and choose a business direction that suits you.

The “lobster craze” of 2026 serves as a mirror, reflecting the impatience and utilitarianism in the tech industry, as well as the anxiety and confusion among SMEs. When the tide of the frenzy recedes, we will find that what truly drives industry development and solves business problems is never a get-rich-quick scheme but rather down-to-earth technological innovation and rational, pragmatic business practices.

This digital lobster will ultimately transition from being the 'star of the frenzy' to just another 'tool in the toolbox.' Those who seek quick profits amid the frenzy will eventually be forgotten by the market; only those who truly delve into technology and create value will be able to stand firm in the tides of the times.