With an annual income exceeding 800 million, a super hidden champion has emerged in Huzhou, Zhejiang: AI + community, on the verge of a major breakthrough.

"My 97-year-old grandfather suddenly lost feeling in his legs while taking a bath last winter. He struggled in the toilet all night without being discovered, and his life was almost in danger. Based on this experience, I invested in a device similar to a sweeping robot. It can move to the old person's side when it hears the call, and has one-key alarm and one-key recording functions. It doesn't involve privacy monitoring, but can help the old person save themselves.”

Zhang Runzhe is the founder of the listed company Lianzhang Portal (headquartered in Huzhou). When talking about the application opportunities of AI, he shared that compared with industries that require heavy investment such as large models, small scenarios + medium-sized tracks have more opportunities, such as elderly care robots.

The core business of Lianzhang Portal is smart communities and family life. Based on this, Zhang Runzhe can access many life scenarios. In a conversation with Pencil News, Zhang Runzhe mentioned multiple AI application opportunities.

1. A thousand-billion track is not as good as "the vertical user lifestyle in a vertical life scenario".

He said bluntly that he prefers the "small and beautiful" projects where a 20-person team can earn tens of millions a year. He believes that the real gold mines in the AI era are hidden in segmented scenarios such as daily life services.

2. Technology worship is not as good as scenario insight. In AI entrepreneurship, sensibility and rationality should coexist.

After covering 10 million families across the country through access control devices, he decided to abandon "technology worship-type" investments such as large models and instead bet on ultra-segmented scenarios. Under the extremely rational technology of AI, an extremely emotional entrepreneurial intention is needed to continuously serve users well.

3. AI will not have an impact on the O2O traffic entrance for the time being.

The traffic fortresses in offline scenarios need to be laid offline. This heavy physical inertia makes it difficult for even the most advanced AI technology to achieve a quick victory. In the AI era, the migration of traffic entrances will first occur online, and it will be slower offline, but it must not be missed.

4. There are no major opportunities found in AI + O2O for the time being. We need to wait for one year.

Now many AI technologies are still in the "integration period", especially in the TOC field, and no phenomenon-level products have emerged. In the short term (three to six months), I don't think there will be particularly big opportunities.

5. You can lose in entrepreneurship, but you can't go bankrupt.

In the smart community track, there was a battle among thousands of enterprises. Lianzhang Portal is a survivor who crawled out of the corpses. Its survival philosophy is that always getting 80 points is more important than getting 100 points once or twice.

One of Zhang Runzhe's current investment directions is to follow the ecological layout of Lianzhang Portal. The intelligent access control screens of the latter have been installed in more than 4,000 communities, covering more than 2.7 million households. Its revenue in 2024 was 114 million US dollars (about 818 million yuan).

Statement: The interviewee has confirmed that the information in the article is true and accurate. Pencil News is willing to endorse its content.

The traffic entrance in the AI era

Pencil News: Are you interested in AI investment based on strategic investment considerations or more on financial investment?

Zhang Runzhe: I'll talk about this from two levels. First, from the perspective of Lianzhang, we definitely consider our business ecosystem, that is, the community life scenario. Relying on the tens of millions of families we cover, we hope to become an important entrance for AI hardware to enter families. We don't want to be left behind in this wave of AI.

Second, from my personal perspective, I also paid attention to the technology field in the early days. In this wave of AI, I missed the early investment in the underlying framework technology. Now I value the application layer more and think there are still opportunities.

Pencil News: You just mentioned "missing out". Why do you say that?

Zhang Runzhe: In the past few years, Lianzhang mainly focused on the expansion of its own business. Although many people now say that smart communities are combined with AI, in my opinion, the big opportunities haven't come yet. It's just the right time for Lianzhang to enter the market.

Pencil News: So, have you seen any AI opportunities that make you "excited"? Or is it possible to grow a "new Lianzhang" in the AI field?

Zhang Runzhe: If I really have to say what makes me "excited", I'm more concerned about the application of AI hardware in family life scenarios. Lianzhang's current national market share is actually only 1%, serving 1% of the Chinese people. We originally had at least ten times more growth space. As Lianzhang's coverage continues to increase, our value as a family entrance will continue to expand.

I think this wave of AI application opportunities won't produce super big single products like Douyin or WeChat in the earlier days, but there will be many small and medium-sized applications. Especially for some small teams of about 20 people, with an annual revenue of over 100 million and a 20% profit margin, I think there will be many such companies.

Pencil News: Do you prefer small teams?

Zhang Runzhe: Yes, I don't pursue the "thousand-billion track". Instead, I prefer small teams in medium-sized markets with stable profits. Instead of betting on super unicorns, it's better to find some stable and high-quality application teams and empower them with Lianzhang's capabilities.

Pencil News: Which directions or tracks do you think are suitable for growing such teams?

Zhang Runzhe: I actually don't highly recommend "technological leadership". The core underlying technologies are in the hands of a few resources, such as foreign laboratories and academicians.

Now technology is iterating rapidly. What used to cost 100,000 yuan a month to do can now be done for 20,000 or 10,000 yuan.

Generally speaking, I value two aspects in investment:

First, the learning ability and innovation ability of the team, and whether they can quickly adjust the product logic in a rapidly changing environment.

I attach great importance to the entrepreneurial intention.

For example, we invested in a family elderly care robot because my grandfather suddenly lost feeling in his legs while taking a bath last winter and stayed in the toilet alone all night, almost endangering his life. Based on this scenario, we invested in a device similar to a sweeping robot, which has one-key alarm and one-key recording functions, doesn't involve privacy monitoring, but can help the elderly save themselves. Behind this is a "warm intention".

Pencil News: I really agree with this point. Many times, entrepreneurship can go further with those "soft" intentions. Do you not touch large models or projects that require heavy investment?

Zhang Runzhe: Yes, our company doesn't participate in this area. I'm more concerned about application implementation and more business-oriented. The company hopes that the investment projects have strong certainty and that we can help with them.

Pencil News: Then is it possible that these projects are more likely to produce a new "ByteDance"?

Zhang Runzhe: There is a possibility, but it depends on the time difference between the development speed of AI technology and the market explosion.

If the technology is updated quickly and the market suddenly explodes, and the giants can't react in time, it is indeed possible, such as DeepSeek.

But if the development speed is in sync with the market demand, then it's basically impossible.

Pencil News: Regarding investments around the Lianzhang ecosystem, such as smart communities and elderly care robots, do you think these directions are big enough?

Zhang Runzhe: I won't measure this by market value. Market value fluctuates. I value more how many families it can actually serve and whether it can develop second and third growth curves during the service process.

Especially, we should look at its entrance method and entrance population, rather than simply looking at "how big it can grow". The needs of a small number of people are also needs.

Pencil News: The entrance method is very crucial. Do you think the "entrance logic" of smart communities will change in the AI era?

Zhang Runzhe: I don't think so.

Lianzhang chose access control as the entry point because access control is a rigid need and belongs to the livelihood project. It won't be replaced by other technologies in the short term of ten years. A good entrance should meet several conditions:

1. It won't be replaced in the short term (like mobile phones and PCs);

2. It has a high enough frequency. Access control is a medium-high frequency scenario;

3. It has enough pain points. For example, elderly care has enough pain points in the scenario.

Pencil News: In the future AI era, can access control still maintain its position as an entrance?

Zhang Runzhe: Lianzhang has been working on this (the community entrance). Access control is just a tool-type entrance. Its commercial value is very different from that of consumer-type and service-type entrances. We polished our diversified revenue model through ten years of continuous iteration.

Now 60% of our main revenue comes from advertising, and 40% comes from user e-commerce revenue.

Essentially, the smart community market has changed from an incremental market to a stock market. So we must build an ecological chain. It won't work to focus on just one link.

Pencil News: I'd like to ask, in the past mobile Internet era, an important traffic entrance was the App Store, but later it was replaced by super apps like WeChat and Douyin. Do you think in the AI era, will there also be such a change in offline traffic?

Zhang Runzhe: This is inevitable. Traffic migration will definitely happen, just to different degrees. Now the mobile phone is the main hardware entrance, and these online giant platforms are still at the infrastructure level. But if a new interaction method appears in the future, such as smart glasses, a major migration may occur. But I can't see such an interaction form in the short term.

Offline is different from online. The biggest problem offline is "installation". It is much heavier than online. Just like when the online payment giants installed payment devices back then, the cost was high and the threshold was also high. So I think the traffic migration in the AI era will first occur online, and it will be slower offline, but it must not be missed. That's why companies like ours must participate in AI.

Pencil News: Then what do you think are the relatively big opportunities after the combination of AI and smart communities?

Zhang Runzhe: I don't think there will be particularly big opportunities in the short term (three to six months). I'm more optimistic about the situation after one year. Now many AI technologies are still in the "integrated" stage, especially in the TOC hardware field, and no phenomenon-level products have emerged. Online products like GPT and Kimi may explode within a year, but offline, especially killer products that rely on hardware, I don't think they have appeared yet.

The O2O opportunities of AI

Pencil News: Now the topic of AI + offline is very hot, such as robot projects. What do you think?

Zhang Runzhe: Now it's mostly in the TOB scenario, not widely popularized among the general TOC public. There are no real phenomenon-level products yet. The reason is that the underlying logic of AI is too rational, while TOC consumption is emotional. It takes time and market education for an emotional scenario to accept rational AI.

Pencil News: Then which opportunities do you think will explode first after one year?

Zhang Runzhe: I think it's small scenarios and small groups of people. The more specific and vertically segmented the problems are, the easier it is to use AI to improve. It's too difficult to handle big and comprehensive things. Now AI has significantly improved efficiency in the TOB scenario, but the TOC scenario hasn't reached that stage yet. The first step for AI to integrate into daily life is not to replace but to improve, especially in companion applications. Such applications will gradually appear in the future.

Pencil News: Many people also compare AI with the mobile Internet back then. The O2O at that time drove many changes in lifestyles. Will AI also have similar opportunities?

Zhang Runzhe: O2O solves the efficiency loss caused by information asymmetry. Now O2O itself is already using AI to improve efficiency, such as the route planning of deliverymen and warehouse management. But what I'm talking about is more about the real experience changes of TOC users. For example, a friend of mine invested in a massage robot project. I think the first step in such scenarios may be human-machine collaboration, such as a person massaging the upper body and the machine massaging the lower body. Essentially, users need interaction and a human touch, not pure intelligent replacement.

Pencil News: Then what do you think of directions like unmanned delivery and Robotaxi?

Zhang Runzhe: These directions are essentially a balance between policies, people's livelihoods, and technology. It's a rational process, similar to TOB. It's an inevitable trend, but the balancing time is undetermined. These are within the scope of personal investment, such as buying stocks of these listed companies. And the real opportunities suitable for mass entrepreneurs are what I mentioned - small scenarios and small groups of people.

Pencil News: What exactly do small scenarios refer to? Taking the family as an example, there are cooking, taking a bath, watching TV, elderly care, pets, etc.

Zhang Runzhe: I think we should first look at the population and then at the needs. An excellent team must be composed of people with strong emotional driving forces. They can dig deep into the real needs of specific groups of people. Moreover, a good team is more like a "connector" and a "combiner", connecting AI technology on one side and fitting real needs on the other side.

For example, in elderly care, one challenge is that the elderly don't like to be monitored, but they need someone to take care of them when something happens. This is a contradictory need, and technology should be used to solve this contradiction well. Another example is when the elderly take a bath. Putting a stool in the toilet is a real small need, but it requires enough sensibility to discover these details.

Pencil News: First look at the population, then at the needs. If we follow this idea, for men, women, the elderly, and pets, how would you rank them in terms of market size and opportunities?

Zhang Runzhe: (Smiling) I actually don't like to rank these markets. In my opinion, they are all very large, and there's really no need for ranking.

The key is to look at your intention and original aspiration. Can you really serve that scenario well? In fact, not to mention these four groups, just the "elderly" group can be further segmented into a huge market. For example, elderly people who can speak Mandarin, those who can't, or obese elderly people, frail elderly people... Real segmentation requires adding two or three adjectives. One adjective alone is not enough.

Pencil News: There used to be a saying in the industry that one should never do business targeting men, as men are not as profitable as pets. In the past new consumption era, this was a consensus. I wonder if this logic still holds in the AI era?

Zhang Runzhe: I think this view is too one-sided. For example, is massage a male consumption? What about going to KTV? Male consumption is actually not low, and there is a lot of entertainment consumption. So it still depends on the scenario. Each group has its own unique needs.

Pencil News