Why is this wave of "lobster craze" led by counties (cities, districts)?

“Lobster” goes viral, and the authorities step in.

On March 7, Shenzhen Longgang released the “Several Measures to Support the Development of OpenClaw & OPC (Draft for Comments)” (hereinafter referred to as the “Lobster Ten Measures”). Subsequently, more news was announced, including support policies related to “lobster” from Wuxi High-tech Zone, Changshu City in Suzhou, Hefei High-tech Zone, Qixia District in Nanjing, and Xiaoshan District in Hangzhou.

Throughout the process, both the number of policies and subsidy amounts have continued to increase. It’s clear that from the Greater Bay Area to the Yangtze River Delta, a rapid industry competition under the banner of “raising lobsters” has begun.

The so-called “lobster” is a nickname for the open-source AI agent OpenClaw, named after its red lobster icon. The popularity of OpenClaw signifies a crucial shift in artificial intelligence from “dialogue” to “execution.” For cities, this will inevitably trigger a new wave of industrial transformation.

“In the era of multi-agent systems, whoever first completes the closed loop will secure their position,” said Zhao Bingbing, Director of the Artificial Intelligence (Robotics) Department in Longgang District, Shenzhen, in an interview with Daily Economic News·Urban Evolution. Cities that deploy “lobster” early are expected to take the lead in industry transformation and become pioneers of new paradigms.

With OpenClaw’s rapid rise, related concepts like OPC (One Person Company) have attracted renewed attention. From OpenClaw to OPC, leading cities such as Shenzhen, Shanghai, Beijing, and Hangzhou are vying to be the first to scale deployment.

Another noteworthy detail is that this round of industry deployment is mostly led by county (city, district) levels. What new development trends might be hidden behind this?

Race for the Trend

From the Greater Bay Area to the Yangtze River Delta, “lobster” has become a new industry hotspot that cities are competing to develop.

First, Shenzhen Longgang launched the “Lobster Ten Measures,” offering subsidies and computing power support, with a maximum subsidy of 4 million yuan, aiming to create a “zero-cost startup” environment for AI agent entrepreneurs.

After Shenzhen’s initial move, policy support among cities has visibly intensified: Wuxi High-tech Zone, Changshu City, Hefei High-tech Zone, Qixia District in Nanjing, and Xiaoshan District in Hangzhou have all announced plans to “raise lobsters.” Policies are becoming more detailed, and subsidy caps are gradually rising.

From the breadth of policy coverage to the continuous enhancement of support intensity, it’s evident that local governments are fiercely competing on this new track, eager to secure their positions.

To understand this policy race, one must first recognize the impact OpenClaw has on the entire AI industry—its disruptive nature lies in breaking down technical barriers and pushing the industry toward a new stage of “inclusive and practical” AI.

Shen Hao, Chief Engineer of the Shanghai Artificial Intelligence Institute, pointed out to Daily Economic News·Urban Evolution that OpenClaw, as a phenomenon-level open-source AI agent, lowers the public’s understanding threshold of AI, allowing society to directly perceive the tangible value of AI carriers, and promoting technology from “big tech exclusive” to “accessible to all.”

This inclusive revolution is fostering new industry forms, not only driving the rapid rise of “AI+” sectors like virtual hosts and smart peripherals but also pushing AI technology toward more realistic and interactive capabilities.

More importantly, OpenClaw has pioneered a new technological paradigm of multi-AI collaborative invocation.

Economist and member of the Ministry of Industry and Information Technology’s Information and Communications Economy Expert Committee, Pan Helin, emphasized to Daily Economic News·Urban Evolution that OpenClaw has opened a new mode of multi-AI collaboration, enabling the invocation of other AI tools, cloud computing, software, and search engines to complete complex tasks for users. Over time, such applications will mature and become the trend.

Zeng Gang, Dean of the Urban Development Research Institute at East China Normal University, noted that the government work report in 2026 first introduced the concept of “intelligent agents,” emphasizing “cultivating new forms of intelligent economy” and “supporting the development of entrepreneurial models suited to the AI era.” The continuous issuance of policies supporting OpenClaw is, to some extent, a response to national strategic deployment.

“OpenClaw’s open-source nature and cross-platform capabilities make it a ‘new operating system’ connecting cloud computing and end-user hardware, creating opportunities to ‘rebuild’ all hardware,” Zeng Gang further explained. The fierce competition among major cities to deploy OpenClaw aims to seize strategic high ground in “AI + manufacturing” and “smart terminals.”

Many share the view that the emergence of OpenClaw marks another “DeepSeek” moment in city AI industry planning. In a sense, its impact at the application and industry level could be even more profound.

“As the government is willing to promote technological progress, especially in building new forms of smart economy, to better play its role and accompany everyone in trial and error,” Zhao Bingbing said. The rapid rollout of support policies by cities also signals to AI entrepreneurs that “once here, you can stay and develop well.”

However, attention must also be paid to potential risks. Pan Helin pointed out that most “lobsters” are deployed locally, and addressing security risks is the most feasible way to cope with the “lobster craze.” Additionally, encouraging large AI enterprises to develop secure, OpenClaw-like products and continuously enriching their functions through open-source communities is essential.

OpenClaw remains in its early stages; raising lobsters requires caution, and safety cannot be overlooked. Shen Hao also mentioned the need to focus on the information security risks of agents, and to quickly establish safeguards through normative rules.

Regional Ecosystem

Local “raising lobsters” policies often come with simultaneous OPC support.

This is understandable: from foundational open-source capabilities to individual entrepreneurship, there is a clear pathway driven by local governments racing to enter the AI track, aiming to lower barriers and incubate the next unicorn.

Currently, for AI entrepreneurs interested in “raising lobsters,” subsidies are a major attraction. Although OpenClaw is free and open-source, the process of “raising lobsters” consumes a large number of tokens, which must be paid for by users. Media reports indicate some users spend about 30,000 yuan per month on token consumption.

Government financial backing can help startups get through the initial “token tuition” phase but cannot build a long-term cost moat. Therefore, for local governments, policy support is not just about initial subsidies but also about creating a sustainable ecosystem.

At present, leading domestic cities are actively developing OPC ecosystems. Beijing, leveraging Zhongguancun’s AI Beiwang community, has launched an AI OPC service plan, coordinating capital, universities, and industry resources to form the country’s first systematic, full-cycle OPC cultivation system.

Shenzhen, through policy innovation, issued the “Action Plan for Building an AI OPC Startup Ecosystem (2026–2027),” aiming to establish over 10 OPC communities and cultivate more than a thousand high-growth enterprises by 2027.

Similar policies are widespread.

The OPC potential of various cities can be ranked as a reference: in February this year, Tsinghua University’s School of Journalism and Communication’s New Media Research Center and others released the “2026 China OPC Startup City Development Index and Rankings,” with Suzhou and Shanghai in the top tier, Shenzhen, Beijing, Wenzhou, and Nanjing in the second, and Hangzhou, Wuxi, Guangzhou, and Fuzhou in the third.

Beyond rankings, the specific development paths of cities are more noteworthy. Zeng Gang believes that among China’s top cities, Shenzhen and Shanghai show the strongest potential for large-scale OPC deployment.

Shenzhen’s approach is “vertical industry penetration.” Relying on its hardware supply chain advantages, Shenzhen can quickly realize large-scale diffusion of OpenClaw from digital to physical worlds. Whether it’s embodied intelligent robots, smart wearables, or industrial inspection systems, Shenzhen can complete the “technology-product-market” closed loop in the shortest time.

Shanghai’s path is “horizontal ecosystem replication.” Zeng Gang pointed out that the exploration of the OPC super-individual community on Fuxing Island in Yangpu District demonstrates that Shanghai’s “university + large factory” industry-university-research collaboration, combined with agile government services, has created a conducive micro-environment for innovation. Once this model proves successful, it is expected to be rapidly replicated in other parts of Shanghai and the Yangtze River Delta, forming multiple high-density innovation nodes.

As Pan Helin said, the development of OpenClaw in various regions should rely more on existing advantages. For example, Shenzhen’s cloud computing capacity, Hangzhou’s Alibaba and “Six Little Dragons” enterprise clusters are core capabilities supporting large-scale deployment.

At the same time, common bottlenecks in OPC development are gradually emerging.

Rong Weihong, Vice Chairman of the Standing Committee of Hangzhou Municipal People’s Congress, recently told the media that China’s OPC industry still faces issues such as traditional institutional constraints, isolated entrepreneurial ecosystems, and insufficient innovation elements support. This means that while many policies are being introduced, they must also improve precision and coverage.

A key breakthrough is opening government AI application scenarios and public data resources. Rong Weihong suggested that, under the premise of ensuring security and privacy, public data and scientific research data should be opened to OPCs in a graded and classified manner.

Key Platforms

A clear trend in the development of new industries is that more and more industry policies are led by county (city, district) levels.

For example, in the recent “Lobster Regulations,” districts and counties such as Shenzhen Longgang, Wuxi High-tech Zone, Hefei High-tech Zone, Changshu City, and Suzhou Changshu are the main policy issuers.

Shen Hao pointed out that compared to provincial and municipal governments, district-level governments are closer to enterprise needs, can respond quickly to technological iterations, and have faster policy response speeds.

Taking the issuance of the “Lobster Ten Measures” in Longgang as an example, Zhao Bingbing explained that Longgang’s ability to lead with a dedicated policy stems from establishing a specialized agency that consolidates industry-related functions previously scattered across departments.

As early as 2025, Longgang established the first government-affiliated agency in China dedicated to AI and robotics—the Human-Machine Department. Its clear official positioning covers industry planning, ecosystem building, enterprise services, scene promotion, and safety management, providing a “bottom line” for AI and robotics industries.

For local AI companies, the Human-Machine Department is the “first interface” for government services. For policy formulation, this flat organizational structure also eliminates the cumbersome process of multi-department approval, greatly improving decision-making and execution efficiency.

Therefore, even before OpenClaw became popular, the Human-Machine Department had already led several “pre-embedded” policy projects, laying out strategic directions and reserving scene resources, which provided a foundation for Longgang’s rapid response to industry hotspots.

“OpenClaw exploded in early March, and Longgang District released a special policy on March 7. Such speed can only be achieved by grassroots departments familiar with the industry,” Zeng Gang said. This “Longgang speed” is precisely the advantage of grassroots government-led industry policies. It also marks a shift from traditional “step-by-step” management to more agile “responsive” approaches.

Zeng Gang further stated that these newly implemented “lobster” policies indicate that China’s industry policy-making is moving from top-down “macro design” to more flexible, locally tailored “bottom-up” strategies—at district or bureau levels.

He believes the benefit is a shift from “broad water irrigation” to “precise drip irrigation.” Historically, most industry policies were issued by national and provincial governments, but future policies will increasingly come from grassroots levels, responding to the growing demand for more detailed industry support from enterprises and society.

Additionally, the logic of industry policies is changing from “competing for land and incentives” to “building ecosystems and fostering innovation.” The core of the “Lobster Regulations” is no longer land and tax breaks but constructing ecosystems around data, computing power, scenes, and talent—aligning with local economic development needs for new drivers.

Furthermore, district governments often serve as “test beds” for policy innovation. Small-scale experiments can accumulate experience for larger-scale implementation. Many of the “raising lobsters” policies are currently in consultation, reflecting a cautious exploratory attitude.

Ultimately, the development and application of “lobsters” and other intelligent agents are not overnight achievements, nor are they mature industries. Large-scale productivity improvements cannot be expected in the short term, requiring long-term planning. How to ensure that funding subsidies truly encourage innovation depends on continuous local government adjustments and clearer, more rational institutional design.

It is also worth noting that, according to CCTV News, recent monitoring by the Ministry of Industry and Information Technology’s cybersecurity threat and vulnerability information sharing platform found that some instances of OpenClaw open-source AI agents, when misconfigured or left in default settings, pose high security risks, easily leading to cyberattacks and information leaks. It is recommended that relevant units and users thoroughly check exposure to public networks, permissions, and credential management when deploying and using OpenClaw, disable unnecessary public access, and improve security mechanisms such as identity authentication, access control, data encryption, and security audits. Continuous attention to official security notices and reinforcement suggestions is essential to prevent potential cybersecurity risks.