Stock Trading "Crayfish" Coming? 3 Major Financial Data Terminal Companies Officially Announce "Shrimp Farming"

21st Century Business Herald Reporter Liu Xiafei

After more than a dozen major internet companies like Tencent and Alibaba joined the “shrimp farming” battle, major financial data terminal providers such as Wind, Tonghuashun, and Eastmoney have also officially announced their involvement.

On March 11, Wind officially announced the launch of “WindClaw,” claiming it is a “small investment lobster that can research, evolve, and communicate data.”

Immediately after, on the early morning of March 12, Tonghuashun announced the launch of “iFinD Financial MCP,” emphasizing “providing professional financial data for OpenClaw.” At the same time, 21st Century Business Herald learned that Tonghuashun is also preparing to launch its own self-developed “iFinD Claw” product, similar in positioning to WindClaw.

That evening, Eastmoney released “Eastmoney Skills,” highlighting the installation of “investment decision-making assistance skills” for OpenClaw.

It is worth noting that after the explosive popularity of the DeekSeek large model in 2025, these three companies also launched their own self-developed large model products.

From “DeekSeek Moment” to “Lobster Moment,” the AI competition among major financial data terminal providers continues to heat up. Experts interviewed pointed out that future software competition may increasingly depend on who can be better “driven” by AI. The race among major financial data terminal providers to “farm shrimp” is not only a response to the trend of industry intelligence but also a necessary move to consolidate business barriers and extend service value.

However, while these providers are immersed in the “shrimp farming” craze, some of their main institutional clients—brokerages—have already begun to cool down the internal “shrimp farming” trend, imposing clear restrictions on the installation and use of OpenClaw.

Wind, Tonghuashun, and Eastmoney Announce “Shrimp Farming” Successively

Within two days, the three major financial data terminal providers each launched their own “weapons” to respond to the “shrimp farming” battle. However, their product strategies differ: some focus on “data,” some on “skills” to actively connect to OpenClaw, and others directly develop “professional versions of OpenClaw” based on the native OpenClaw.

A fintech expert explained to the reporter that, as an AI intelligent agent capable of full-process execution, OpenClaw’s capabilities largely depend on two supports: one is massive high-quality data, which is the “nutrient” for OpenClaw’s learning and evolution; the other is diversified skills, which are the “tentacles” for practical application and problem-solving. Therefore, many companies choose to connect to OpenClaw based on their own advantages, prioritizing one of these two paths.

Specifically, Wind has chosen to develop a “professional version of OpenClaw,” launching “WindClaw,” which is still in public testing.

According to Wind’s official introduction, the core features include access to Wind’s professional financial data, one-click local deployment, and continuous autonomous evolution based on user investment habits. The reporter found on the public test application page that its homepage features functions such as “Monitor the Market,” “Track Individual Stocks,” “Follow News,” “Stock Analysts,” “Macro Researchers,” and “Strategy Excavator.”

Image: WindWindClaw public test page

Meanwhile, Tonghuashun has prioritized data access, acting as a “professional financial data source,” launching “iFinD Financial MCP.”

Regarding MCP, Tonghuashun explained, “Without MCP data tools, large models would only search the internet and fail to meet the needs of research personnel for financial data.” iFinD MCP emphasizes seamless access to research-grade databases, pure natural language interaction mode, built-in professional data cleaning, and token optimization mechanisms.

According to Tonghuashun, the core modules currently open for iFinD MCP include A-share stock analysis, mutual fund analysis, macroeconomic and industry data, announcements, and news.

However, Tonghuashun has not given up on developing its own “professional version of OpenClaw.” The reporter learned that Tonghuashun is also planning to launch “iFinD Claw” soon, similar in positioning to WindClaw, aiming for “out-of-the-box” usability.

Image: Tonghuashun iFinD MCP access to OpenClaw page

Eastmoney, on the other hand, is using “skills” as the entry point, releasing “Eastmoney Skills,” which proposes installing “investment decision-making assistance skills” for OpenClaw.

“Skills are standardized folders that help large language model assistants acquire professional financial data service capabilities,” Eastmoney explained. After “installing” a skill into OpenClaw or other AI assistants, the AI gains the ability to call corresponding financial interfaces.

Eastmoney stated that after installing Skills, OpenClaw can achieve real-time market information retrieval, automated cleaning and structured processing, and can systematically screen and analyze thousands of targets based on fundamental and technical indicators, helping investors efficiently identify suitable targets.

The installation interface shows that Eastmoney Skills currently mainly include three skill packs: news search, financial data, and intelligent stock selection.

Image: Eastmoney Skills installation page

From “Large Models” to “Lobsters,” the AI competition escalates again

In fact, the current rush to connect or develop “small lobsters” is just a microcosm of the AI competition among leading players in the financial data terminal industry.

In 2025, after the “DeepSeek Moment” ignited the “large model craze,” several top financial data terminal companies released their own large models, including Wind’s “Wind Alice,” Eastmoney’s “Miaoxiang,” and Tonghuashun’s “WenCai HithinkGPT.”

According to a report by 21st Century Business Herald, the competition among financial data terminals is shifting from “selling water” to “selling shovels.” In the past, terminals gained competitive advantage through data resources; now, data itself is less of a barrier, and what is needed more are truly useful research and trading tools.

Now, with the arrival of the “Lobster Moment,” what does this mean for major financial data terminal providers?

On one hand, the transformative potential of OpenClaw as a phenomenon-level AI agent may herald a new model of software competition.

宋巍巍, a fund manager at China-Europe Fund, told the 21st Century Business Herald that OpenClaw surpasses simple dialogue and can help users execute tasks. Personal computers, smartphones, and personal cloud servers will become the main platforms for AI agents—“digital employees” or “personal assistants.”

Software with good APIs can be precisely and efficiently invoked by OpenClaw, becoming an “organ” within the agent ecosystem. As AI agents like OpenClaw become more widespread, operating systems may shift from “human-centered” to “agent-centered.”

“Future software competition may be about who can be better ‘driven’ by AI,” said Song Weiwei.

On the other hand, from the perspective of product iteration logic in financial data terminals, actively integrating OpenClaw is also an important extension of service value.

“These products essentially fill the ‘last mile’ from data to application, using AI agents to connect data supply with actual use,” said Zhang Ning, director of the China Fintech Research Center at Central University of Finance and Economics.

Zhang Ning analyzed that current customer demands are upgrading from “obtaining data” to “highly useful data,” pushing service providers to shift from data supply to intelligent tool empowerment. Data vendors, leveraging their underlying data barriers and scene understanding, can quickly deeply integrate intelligent agents with their own databases, building differentiated competitiveness.

“Under this two-way push, the deployment of such products by financial data terminal companies is both a response to industry intelligence trends and a necessary move to consolidate business barriers and extend service value,” Zhang Ning concluded.

Internal tightening of “shrimp farming” by brokerages

Although major financial data terminal providers continue to announce “shrimp farming” initiatives, some of their key institutional clients—brokerages—have already begun to cool the internal “shrimp farming” trend.

The reporter learned from multiple brokerages that many have issued internal compliance notices explicitly restricting the installation and use of OpenClaw, mainly concerning installation, deployment, and use on company devices and internal networks.

From the specific notices, most are reminders of related risks and safety assessments before use; some have issued outright bans, requiring a suspension of installation and use; others have implemented approval processes, requiring employees to apply for permission if business needs justify it.

Notably, on March 10, the National Internet Emergency Center issued a risk alert, warning that for critical industries like finance and energy, certain security vulnerabilities in OpenClaw could lead to leaks of core business data, trade secrets, and code repositories, or even cause entire business systems to crash, resulting in enormous losses.

Zhang Ning warned that “lobster”-type AI applications, which tightly integrate data retrieval, analysis, operation, and command execution, not only fill the gap from data to user but also introduce more covert security risks.

宋巍巍 also stated that once AI gains Full Disk Access, any security vulnerability could lead to systemic data leaks. Additionally, the third-party plugin ecosystem (ClawHub) for OpenClaw may also pose security risks.

Regarding financial practitioners, Zhang Ning believes that beyond traditional risks like data leakage, compliance, intellectual property, and reputation, new risks related to the “black box” extension should also be vigilant.

For these new “black box” extension risks, Zhang Ning explained that the integrated application of full-process data operations breaks traditional segmented supervision and audit trails. Internal network operations are highly concealed, making abnormal flows and external transmissions difficult to detect. Cross-system integration also provides covert channels for injection attacks and plugin risks, making risk propagation harder to trace and increasing the likelihood of systemic data security incidents. This is the core reason why many brokerages strictly prohibit such tools on their internal networks.

(This article also contributed by reporter Li Yuchen)