Brokerage Strategy Conference Live Teaches "Shrimp Farming"! OpenClaw "Swims" into the Investment Research Circle

21st Century Business Herald Reporter Liu Xiafei

Recently, the topic of “Long lines outside Tencent Building to buy ‘Lobster’” has been widely circulated online. However, this queueing craze may now be spilling over into the securities firms’ strategy conference venues.

The 21st Century Business Herald reporter has noticed that OpenClaw, an open-source AI agent known for its red lobster logo and nicknamed “Little Lobster” by users, is moving from the geek community into the financial research and investment field.

Recently, many securities firms have been releasing “Lobster Raising Guides,” and some even hold dedicated forums at strategy meetings to teach “Lobster Raising”; fintech companies are competing to develop their own “Research Version of OpenClaw” to seize the opportunity.

However, moving from “spectating” to “actually working” involves several hurdles related to cost, security, and habits. How deep is OpenClaw’s “migration” into the research circle?

Image: Securities analyst teaching “Lobster Raising” on-site (Source: Founder Securities Research Institute)

Securities Firms Collectively Discuss “Little Lobster”

Recently, major securities firms have been holding their spring strategy meetings one after another. The reporter observed that, besides traditional macroeconomic assessments and industry analyses, this year’s strategy meetings also introduced the hottest “Little Lobster” into the scene.

The so-called “Little Lobster” is actually an open-source AI agent called OpenClaw, claimed to be capable of “executing tasks” and “working” like a human. As “Little Lobster” has become a viral sensation online, its application prospects in research and investment have attracted considerable attention from institutions.

For example, Guojin Securities scheduled a dedicated “OpenClaw Empowering Research and Index Investment Forum” during its spring strategy meeting on March 12-13. The agenda shows that this sub-forum covers topics from “The New Paradigm of Research Empowered by Large Models” to practical applications like “OpenClaw in Active and Quantitative Research,” “Building Personal Research Assistants with OpenClaw,” etc., offering comprehensive content. This is not Guojin Securities’ first attempt. Since late February, they have been touring Shanghai, Beijing, and other cities to hold “OpenClaw Empowering Intelligent Research” forums.

In fact, this “Lobster Raising” science popularization wave has recently swept through sell-side research institutes. According to incomplete statistics, by March 10, at least nine securities firms—including CITIC Securities, Huatai Securities, Orient Securities, Huachuang Securities, Eastmoney Securities, Dongwu Securities, Kaiyuan Securities, Founder Securities, Huafu Securities—have scheduled “OpenClaw” themed courses on roadshow days, introducing deployment methods and research application techniques to institutional and individual investors.

The reporter also noted that a session titled “OpenClaw: From Beginner to Expert,” held by OpenSource Securities on the evening of March 10, had nearly 1,000 views on the roadshow platform at the time of writing.

Image: Research institute OpenClaw-related roadshow schedule

From the content shared by various institutions, the training can be mainly divided into two categories: one is basic introductory content, including concepts of OpenClaw, quick deployment, information access, etc.; the other focuses on scenario applications, such as information retrieval, stock analysis, stock selection strategies, fully automated factor mining and backtesting, etc.

Additionally, several securities firms have published “Lobster Raising Guides” in the form of special research reports, evaluating OpenClaw’s research functions and providing practical breakdowns.

For example, OpenSource Securities prepared a 100-page PDF titled “OpenSource Financial Engineering OpenClaw Technical Documentation” for its online seminar, claiming that “no foundation is needed, and you can deploy your personal AI gateway in 5 minutes.” Some research reports, due to their detailed, step-by-step explanations, have gone viral on social media, such as Founder Securities’ “Empowering Financial Research with OpenClaw: 17 Efficient Application Cases” and Northeast Securities’ “Install these 20 Skill Packs on your OpenClaw to boost research efficiency by 10 times,” etc.

“Research Version of Lobster” Positioning Battle

Research analysts at securities firms are busy studying OpenClaw and teaching investors how to “raise lobsters.” Meanwhile, fintech companies with keen senses are focusing on the technical pain points of OpenClaw’s implementation, exploring deeper product development, and trying to upgrade it from a “geek toy” to a “professional research tool.”

Although OpenClaw claims to require “no coding experience,” as an AI agent, its practical application in research scenarios still faces two major hurdles: data authority and deployment complexity. The reporter observed that different companies are exploring their own optimal ways to connect with OpenClaw to overcome these challenges.

One approach is to focus on the data side, positioning themselves as “professional databases,” encouraging investors to connect their research data sources into OpenClaw. For example, Gangtise Information, a company known for aggregating analyst opinions, announcement summaries, and other research info, defines its research platform as an “AI research database + knowledge base,” and provides API interfaces for OpenClaw integration.

According to Gangtise’s research technology team, “Knowledge bases and databases are the two pillars of new research infrastructure.” Their way of “embracing” this “lobster” is to provide dedicated research data sources for OpenClaw.

Image: Research data access page (Source: Gangtise Research)

Another approach is to further productize and deploy in the cloud, aiming for “out-of-the-box” usability with lower technical barriers, allowing users to create their own “digital researcher” via natural language. For example, JMT (Jinmen Technology) has launched “Research Lobster,” which encapsulates and optimizes OpenClaw’s core capabilities. By integrating high-quality research data such as roadshows, research reports, industry maps, and EDB, and including the full set of Skills from the open-source community along with preloaded professional research Skills packages, they aim to reduce user learning and configuration costs.

JMT’s AI research team told the reporter that “Research Lobster” must have research-specific “data genes,” specially optimized operational capabilities, and a large intelligent research ecosystem.

In terms of data, it requires engineering processes such as unified data organization, standardization, and precise correlation to form a clear, reliable research knowledge system, while also ensuring data security in finance. On the ecosystem side, “Research Lobster” can connect with existing platform capabilities like AI conference hosting, AI transcription, AI translation, research brain, event signals, etc., to meet the full workflow needs, according to insiders.

Image: AI research tool interface (Source: JMT)

Notably, the reporter also found on social media that, besides professional fintech firms, some individual bloggers are selling their self-developed AI research systems based on OpenClaw, with core functions including news queries, data analysis, target tracking, priced at a few hundred yuan.

However, some users feedback that these “DIY” products are more suitable for “playing around,” and are not reliable enough for serious research—unstable data sources, outdated info, frequent errors, and other issues significantly impair work efficiency.

Why is the “Lobster Raising” phenomenon so popular in the research circle?

From securities firms to fintech companies, why has the “lobster” caused such a splash in research and investment?

From a product logic perspective, OpenClaw is regarded as a “productivity-level” AI tool. Several AI product managers told the reporter that compared to previous AI products, OpenClaw’s uniqueness lies in its evolution from “being able to talk” to “being able to do.”

Earlier AI products focused on Q&A scenarios, but OpenClaw can directly operate a computer to perform various tasks. This means it is no longer just a dialogue tool but an executor capable of completing specific tasks.

This capability directly addresses several pain points in current research work: information overload, repetitive tasks, and high technical barriers for quant strategies.

In research, OpenClaw simplifies tasks like data retrieval and coding into “demand expression,” allowing researchers to focus on analysis itself.

For example, as shown in Guojin Securities’ research report: simply “throw” a research paper to OpenClaw, and it can automatically parse logic, fetch data, write code for backtesting, and output net value charts and analysis results. In the past, this process would require analysts to manually clean data, write code, tune models, taking half a day or even several days.

Founder Securities’ financial engineering analyst Cao Xiaochun pointed out that OpenClaw can significantly reduce the difficulty of building various tools, data, and quantitative stock selection strategies, freeing investors from repetitive, rule-based work, and allowing more focus on complex decision-making and innovative strategy development.

“Lobster Raising” Business Insights in Research

Of course, beyond the product advantages, the phenomenon of institutions and companies rushing to “raise lobsters” points to deeper business logic.

For securities firms, the emergence of such a phenomenon-level industry hotspot as OpenClaw is a dual test of research capability and service ability.

A quantitative research analyst at a securities firm explained, “To truly understand the value and potential of such a phenomenon-level product, you need to try running it yourself for a while and complete several tasks. The process of writing reports is also a process of the team learning to use the tool.”

He added, “Learning it yourself” is just the first step; “teaching clients to learn” is equally important. Whether through “Lobster Raising Guides” or live tutorials, the core message from securities firms is: we not only understand research but also know how to use the latest tools to help you research. In the fierce competition among research institutes, this “capability output” is more scarce than just providing opinions.

For fintech companies, “Lobster Raising” also has special significance in the AI research business.

One is about positioning at the technical entry point. As user demand for these tools grows, whoever can make it easier for users to access OpenClaw will be able to embed into their daily research workflows and become an entry point into this “agent ecosystem.”

Another is that it introduces a new business model. A fintech insider explained that traditional financial terminals mostly operate on SaaS (Software as a Service), where users pay for tools and perform operations themselves. OpenClaw, however, brings a mode closer to AI as a Service (AIaaS), where users buy not just the tool but the “intelligent capability” to perform tasks. When AI can truly “do the work” for users, they will be willing to pay for the “results achieved,” not just for the “tool.”

Overall, whether it’s securities firms teaching “Lobster Raising” on-site or fintech companies developing “Research Lobster,” they are all leveraging the current focus of user attention to “seize” the next wave of research tools and service discourse.

However, whether “Lobster Raising” for research is a good business remains uncertain at this stage.

An industry insider told the reporter that building a truly professional and reliable research assistant using OpenClaw involves costs such as cloud deployment, token consumption, database procurement and setup, etc., and overall, it is still a “cost” phase rather than a “profit” phase in the short term.

Moreover, although institutions and companies are trying their best to make OpenClaw accessible to ordinary users, most users still adopt a “spectator” mentality. In real-world use, the time cost of tuning OpenClaw, the technical barriers, the risks of granting excessive permissions, and the potential for incurring exorbitant token fees all deter many users.

On March 10, the National Internet Emergency Center issued a risk alert, warning that security vulnerabilities in OpenClaw could lead to leaks of core business data, trade secrets, and code repositories, or even cause system paralysis, resulting in unpredictable losses in critical industries like finance and energy.

In summary, moving from “spectating” to “actually working” still faces several hurdles related to costs, security, and habits. Whether OpenClaw’s entry into the research circle is a “catalyst” stirring up a pond or just ripples passing over the water remains to be seen by time and market validation.