Open Claw concept stocks pull back, the "Lobster" three brothers decline, how will the AI investment landscape evolve?

Previously popular Open Claw concept stocks experienced a collective pullback, with A-shares and Hong Kong stocks related to these stocks all declining to varying degrees. Trading data from March 11 shows that in the A-share market, Boya Data once fell more than 12% intraday. Although it recovered somewhat by the close, it still ended down 8.75%. In the previous two trading days, the stock had risen 20% and 16.40%, respectively. Stocks such as Kunlun Wanwei, Shunwang Technology, and Meige Smart also fell in tandem. In the Hong Kong market, the so-called “Lobster” trio—Xunce, Mini Max, and Zhipu—each dropped over 6%.

It is noteworthy that these declines are not primarily in the main holdings of current public funds. According to public data, as of the end of Q4 last year, Boya Data was only held by one active equity fund; Kunlun Wanwei was heavily held by 36 funds, and Shunwang Technology by 7 funds. Market analysts point out that this correction mainly relates to a cooling of market sentiment rather than fundamental changes.

The trigger for this correction was the release of risk warnings by several listed companies. On the evening of March 10, Youkued announced that its lightweight cloud hosting products equipped with Open Claw images have not yet formed a scaled product system. The progress of technical iteration and commercialization may fall short of expectations. Currently, related business revenue accounts for a very small proportion, and future revenue scale and profitability are highly uncertain. Boya Data also stated that the company currently has no monitoring plans or products for Open Claw technology, and has not formed commercial applications or revenue, so it does not impact the company’s business.

Concerns about the security risks of new technologies are a key reason for the sentiment cooling. Risk alerts issued by relevant authorities show that recent downloads and usage of Open Claw have surged, but default security configurations are weak. Once attackers breach the system, they can gain full control. Four types of risks need to be closely monitored: prompt injection risks, misoperation risks, plugin poisoning risks, and security vulnerabilities. Cyberattackers can craft hidden malicious instructions to induce Open Claw to read web pages, potentially leading to leakage of user system keys.

From a technological innovation perspective, the new generation intelligent agent system represented by Open Claw is driving AI from “answering questions” to “performing tasks.” Guolian Fund analysts note that this shift means AI is beginning to understand instructions, call tools, connect with local systems and external applications, gradually gaining the ability to “do things for people.” Market focus is shifting from “model parameters and high scores” to “whether it can truly integrate into workflows, create real demand, and generate sustained calls.”

However, the transition from technological innovation to practical application still requires time for validation. A mid-sized public fund manager in Shenzhen said that although major vendors’ initiatives like opening cloud services show confidence in AI technological innovation, individual usage still has barriers. Investment should return to fundamentals, waiting for substantive technological breakthroughs and product deployment. The fund manager revealed that they will observe the fundamental changes of related concept stocks in the near future but are not rushing to adjust their positions.

Within the AI industry, clear differentiation has already emerged. Chengjin Hexin Fund’s chief economist pointed out that in AI hardware, companies in chips and computing power maintain high growth due to high margins, becoming core beneficiaries during the mature phase of the technology cycle. Meanwhile, traditional software is being disrupted by AI tools, with subscription models under pressure, compounded by unfulfilled AI investment returns, leading to valuation difficulties. This differentiation is part of the cycle’s process of selecting high-quality assets; only companies with core technological barriers and proven performance can continue to benefit during the technology cycle shift.

For investors, the lessons from Open Claw highlight a need to rethink AI investment frameworks. Guolian Fund believes there are at least three key points: first, application deployment is becoming more perceptible, and debates over “whether AI has real demand” may diminish as intelligent agents penetrate deeper; second, tokens and reasoning needs are expected to become important indicators, with companies that secure more calls, entry points, and workflows gaining a competitive edge; third, AI industry opportunities are no longer limited to model companies—areas like applications, cloud, computing power, edge hardware, security governance, and interface ecosystems could all benefit.

Regarding specific investment directions, Fuguo Fund recommends focusing on three major segments: first, cloud services and infrastructure—domestic leading cloud providers have announced support for one-click deployment of Open Claw, offering pre-configured images and computing resources, which will benefit cloud computing, data centers, and server infrastructure providers; second, computing chips and communications—rising call volumes for large AI models are exponentially increasing demand for computing power, making chips, GPUs, and communication devices critical; third, AI applications and software development—the open-source nature of Open Claw fosters opportunities in software applications, big data, and internet platforms.