“Have you raised lobsters?” Recently, this has been the greeting among Web3ers, and it’s likely to be the phrase on everyone’s lips.

As we stepped into 2026, following the explosion of the robot during the Chinese Spring Festival Gala, a new generation of AI Agents represented by OpenClaw became the new toy in the tech community. Some are using AI for customer service, some are writing code with AI, and others are even starting to try using Agents to simulate a whole set of “digital employees.” A concept that has been frequently mentioned on various internet platforms recently is “one-person company,” where a single individual can run what used to require a small team, all through an AI workflow.

Of course, the Web3 space has also been active. Lately, if you pay attention to industry media, you’ll find many projects starting to make a big deal about AI Agents. Some are researching how Agents can directly access on-chain assets or contracts, some are working on the payment, identity, or financial infrastructure for Agents, and some are discussing the “Agent economy,” enabling AI to participate in the network like users, and some even began to chant the new slogan of “Web4.0.”

At this point, you may feel a sense of familiarity.

It’s often said that the fashion circle operates in cycles, and it appears the tech (or crypto) circle does too. Remember back in 2022 during the bear market when ChatGPT exploded in popularity overnight, AI instantly became the hot topic of conversation. The Web3 community was certainly not idle either, quickly spawning a heap of new concepts like AI Agents, AI traders, and automated strategies, as if anything that touched AI could spin a new story. However, this excitement didn’t last long. Once the crypto market started to rise again, everyone’s attention quickly shifted back to Crypto itself.

Now, in the second half of 2025, as the crypto market trends bearish again, Web3 is on the hunt for new concepts to latch onto.

However, from the perspective of Portal Labs, the problem lies exactly here. When a narrative starts to trend, many Web3 startups aren't making technical and business judgments; they are making narrative judgments: Go with whichever concept is hot. That leads to a downturn—

Many teams only discover when they actually push projects forward that while a concept can be quickly put together, it’s much harder to bring a product to fruition. Where are the users? What are the specific scenarios? How will they sustain revenue? Can they attract investment? These questions often only surface after the project has been underway for a while.

When the hype dies down, the market is often left with a bunch of projects that haven’t been able to take off. Some products stall at the demo stage, others launch awkwardly but can’t find users, and some simply vanish along with the narrative. In the short term, it looks like a new track has opened up, but looking back after some time, there isn’t much of substance left behind.

This poses a dilemma: continue to delve deeper into Crypto or pivot to AI. If you choose the former, the market may not be good, and investment might not yield returns; if you choose the latter, there’s no certainty. The technical barriers, talent structure, and competitive environment of AI are quite different from Web3. The technical stacks, product experiences, and community resources accumulated by many teams over the past few years are built within the Crypto ecosystem, and a complete pivot to AI essentially means entering an entirely unfamiliar track. From model capabilities to data resources and engineering teams, almost everything needs to be rebuilt.

More realistically, the AI space is already very crowded. Whether it’s large model companies, traditional internet enterprises, or many startup teams, huge resources have been poured into this field. For a startup that originally focused on Web3, entering this market solely because of a narrative shift may quickly reveal that they lack both technical advantages and industry resources.

In fact, for many Web3 startups, there’s still a viable path forward. They don’t necessarily have to pivot to AI; instead, they can continue on their Web3 path while considering what capabilities Crypto can contribute to the AI ecosystem.

If you take a closer look at the current wave of AI development, you will find that many critical aspects haven’t been fully resolved.

The most typical issue is data. Models are getting stronger, but where do the training data come from, is the data trustworthy and compliant, and particularly, how can AI Agents achieve 1v1 customization? These problems have yet to be addressed by a good mechanism. For AI that depends on large-scale data training, this has been a long-standing foundational issue.

Another issue is identity and collaboration. When AI Agents start participating in task execution, automated trading, or even operational decision-making, they themselves also need identities, permissions, and collaboration rules. Who can call on a particular Agent? How do Agents divide tasks? How is settlement done after task execution? These questions essentially involve identity and value distribution within an open network.

There’s also the payment issue. Once AI Agents begin autonomously calling services, obtaining data, or executing tasks within the network, it means they require a small payment system that can facilitate automatic settlement. However, such a payment structure is challenging to realize within traditional internet systems.

These may all appear to be AI problems, but many solutions already exist within the Crypto technical framework. Whether it’s data incentive networks, on-chain identity systems, or open payment networks, these have been directions Web3 has been exploring over the past few years.

If Web3 startups truly intend to explore these directions, there are a few things they must clarify first.

The first consideration is the technical capabilities of the team. Different Web3 projects have vastly different technical accumulations. Some teams excel at developing on-chain protocols, others have long been involved in data networks, while some focus more on application-layer products. If a team has spent the past few years working on data-related infrastructure, such as data collection, data extraction, or data markets, then extending into the AI data layer will be relatively natural, such as creating data contribution networks, verifiable data sources, or providing incentivized data markets for models. If a team has traditionally focused more on on-chain protocols or infrastructure, they might consider working on the operational environment for AI Agents, such as on-chain identity, permission management, or task execution protocols, or providing automatic settlement and payment capabilities for Agents. For those teams that already work on application-layer products, such as trading tools, content platforms, community products, or consumer applications, AI is better suited as a capability layer embedded within their existing product framework. For instance, using AI to enhance data analysis capabilities, automate operational processes, or utilize Agents for functions that previously required manual handling.

Next, it’s important to examine whether there are real business scenarios. Many AI projects fade quickly not due to a lack of technology but because they lack clear use cases from the outset. The concepts might be compelling, but where are the actual users for this product, why would they use it, and why would they be willing to pay for it? These questions often go unanswered. Some concepts get a lot of buzz within the industry, such as “AI + Web3,” “Agent economy,” or “AI trader,” sounding grand, but when you dig deeper, the stable user groups are not that numerous. In contrast, demands that may seem less “sexy,” such as data processing, automated operations, information filtering, or task execution, have long existed in real business scenarios. Because of this, when determining whether to enter a certain AI direction, rather than simply checking if the concept is trendy, it’s more prudent to examine the scenario itself: Is this scenario a long-standing business problem? Has anyone been willing to pay for it, and can AI genuinely enhance efficiency in this area? If these conditions are met, then that direction is much more likely to transition from narrative to product.

Further down the line, it’s essential to look at whether the Web3 startup has resources that can truly engage in these areas.

The previously mentioned areas concerning data, identity, and payment are fundamentally not merely technical issues but also issues of network resources.

For instance, regarding data networks, if the team lacks a stable data source or a user group capable of continuously contributing data, then even if they develop the technology, it will be challenging to create genuine network effects. Similarly, if they want to establish an identity system or collaboration network for AI Agents, there needs to be real developers, applications, or Agents participating; otherwise, the protocol itself will struggle to form an ecosystem. The logic for payment and settlement systems follows in a similar vein. Once AI Agents start calling services, obtaining data, or executing tasks within the network, small payments will happen frequently. However, this payment network will only be meaningful when a large number of Agents and services exist simultaneously; otherwise, it remains merely a technical module.

Therefore, for many Web3 teams, what truly needs to be assessed isn’t “is there technical room for this direction,” but “can we become a part of this network?” Whether the team already has data sources, developer ecosystems, or application scenarios often dictates whether a project can genuinely engage with the foundational layer of AI instead of remaining at the conceptual level.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。