---
title: "Four Investors Spent an Hour Discussing AI. What Worried Them Most?"
author: "Unique Research"
sourcePublication: "Unique Research Substack"
originalPublishedAt: "2026-04-10T10:18:14+00:00"
canonical: "https://ffcap.cn/en/research/src-20260410-01html"
source: "https://uniqueresearch.substack.com/p/src-20260410-01html"
language: "en"
---

# Four Investors Spent an Hour Discussing AI. What Worried Them Most?

_Original · Unique Research · 2026-04-10_

_Editor's note: This translation retains the original author's first-person account and the investors' statements, not the translator's firsthand experience. The source pairs a 2018 investment with “six years later,” despite its April 2026 publication date; that discrepancy is preserved, not silently corrected. Relative references such as “this year” and “last week” remain anchored to the original discussion. Revenue passages do not explicitly state a currency, so none has been added. Company rankings, investment histories, corporate relationships, official English names and the NVIDIA performance figures require verification before release. Qianli Technology in the narrative and Qianxun Positioning in the transcript are retained as distinct source references. Investment views, estimates and agent-market scenarios are attributed opinions, not independently verified outcomes or investment recommendations._

Unique Awards

Four Investors Talked for an Hour. These Are the Five Things I Remembered.

As AI moves from concept to practical use, capital is looking beyond the excitement to markets, teams, speed, and value that can actually be delivered.

"

There are two kinds of people: those who find AI frightening and worry that it will take over humanity, and those who use it boldly.

In the AI era, speed is the primary force of productivity.

At Hangzhou AI WEEK, I sat in on a roundtable Panel with four investors. To be honest, I had expected another session full of empty talk about how “AI will change the world.” But after listening, I felt there might actually be something to it.

I suddenly realized that the way these people, who had spent more than a decade working in the industry, understood AI was a world apart from the perspective of someone like me, constantly online following the excitement.

“Musk Is Doing It”

Let's start with Wang Aiwu, founder of Longqi Investment, which invested in Cao Cao Mobility. He told a story about Geely.

When they invested in Cao Cao Mobility in 2018, they did so according to an “internet investment logic.” What worried them then? That a traditional automaker like Geely lacked internet DNA.

Six years later, where is Geely now? The country's largest automaker, aiming for the global top five. It has moved into intelligent driving, Qianli Technology, Guangyu Mingdao, and even commercial spaceflight and Geespace.

Wang Aiwu said he had asked the chairman at the time: Why build satellites?

The chairman replied: Musk is doing it.

I laughed when I heard that. It was so true to life. Not some grand strategic blueprint, just benchmarking against someone else. Yet that benchmarking helped turn a traditional manufacturing giant into an integrated operation spanning ground, air, and space.

Wang Aiwu summed it up: In Zhejiang, traditional manufacturing has uniquely favorable conditions for embracing AI.

Thinking it over, I took his point to mean that Zhejiang has a manufacturing base, a world-class software company like Alibaba, and an active venture investment environment. Put those three together and you really do have a distinctive ecosystem.

“20 Million in Revenue, Loss-Making—but We Invested”

Next was Yuan Zhiyong, a partner at Saizhi Bole, which invested in Deep Robotics, one of Hangzhou's six AI “Little Dragons.”

He said candidly that when they invested in Deep Robotics, it had annual revenue of less than 20 million and was losing money.

“We rarely look at price-to-earnings ratios. At the early stage, there are no profits at all.”

What do they look at instead? Three things: the market's future size, its growth rate, and the team's characteristics.

I think this is worth considering carefully. Many voices in the market talk about an “AI valuation bubble” or say they “cannot understand AI companies' business models.” But Yuan Zhiyong's logic is that they were never playing by the price-to-earnings-ratio rulebook in the first place.

AI brings digital employees and greater efficiency, but the evaluation logic has not fundamentally changed. He focuses more on one point: the ability to deploy large models in industry-specific scenarios.

“General-purpose foundations are for Kimi and DeepSeek. We invest in vertical innovation.”

That woke me up. Large models are infrastructure, but the real opportunity lies in injecting industry know-how. A vertical To B version of OpenClaw could be the next battleground.

“The Era of Showing Off Technology Is Over”

Ni Min, executive president of Zheshang Venture Capital, was even more direct.

He said AI infrastructure is a relatively certain opportunity. “The wider the gap in a field, the more compelling the team. Where the gap is smaller, we assess profitability.”

He then mentioned a defining change.

AI investment used to focus on “showing off”—robots dancing and performing. Now it focuses on practicality: whether they can genuinely assist people.

I found that shift particularly interesting. It means AI is moving from a phase of “watching the spectacle” into a phase of “doing the work.” It is no longer about showing muscle, but about implementation.

“One Person Is a Listed Company”

Another of his judgments was more radical: In the future, going public may no longer be necessary.

“One person developing an OpenClaw can be an enormous company. That is already true of livestreamers, and it will be true of developers.”

Countless images flashed through my mind: people building products alone, running operations alone, and monetizing alone. AI really is breaking apart the concept of a “company.” One person's capabilities may match what once required a small team.

“Will There Be a WeChat for Agents?”

Finally, there was Wu Wei, founder of Unique Capital and one of the event's organizers.

He identified three entry points.

The first is data. Whether OpenClaw is used well depends on memory and context management. A vertical To B version of OpenClaw needs industry know-how injected into it.

The second is the medium. Web pages, WeChat, in-car systems, smart hardware—each medium is an entry point.

The third is interaction. Suppose future software is built not for people but for an Agent. Could communication and transactions between one Agent and another create a “WeChat for Agents” or an “Alipay for Agents”?

That point genuinely struck me. It was an entirely different perspective. Most people keep thinking about “how humans use AI.” But have we considered “how AI uses AI”?

Speed Is the Primary Force of Productivity

At the end of the Panel, Wu Wei said something I thought could serve as the discussion's conclusion.

“There are two kinds of people. One finds AI frightening and worries that it will take over humanity. The other uses it boldly. Capabilities, safety, and concerns all deserve attention, but what matters even more is embracing it actively.”

“Time favors the prepared. In the AI era, speed is the primary force of productivity. The faster you iterate, the more likely you are to compete with the big companies in particular niches.”

When I left the venue, the sky over Hangzhou was already dark.

Ni Min's words kept echoing in my head: “Sounds hard to understand, doesn't it?”

It really was hard to understand. But that is precisely what made it interesting.

More from the Conversation

Unique Awards · Hangzhou AI WEEK Trend Roundtable Panel: “Industry and Capital: Rethinking Investment Logic and Discovering Value in the AI Era”

Guests: Wu Wei, Founder of Unique Capital | Ni Min, Executive President of Zheshang Venture Capital | Wang Aiwu, Founder of Longqi Investment | Yuan Zhiyong, Partner at Saizhi Bole

Moderator: Wang Chunfeng, Secretary-General of the Hangzhou Science and Technology Innovation and Entrepreneurship Association Incubator

Wang Chunfeng: Our first question concerns the value of integrating industries. By industry integration and value, I mean how manufacturing and Hangzhou's software sector can integrate deeply with AI. At the same time, as AI becomes a digital employee, how will a company's value change? Let's invite Mr. Wang from Longqi Investment.

Wang Aiwu: Hello, everyone. I'm Wang Aiwu, founder of Longqi Investment. Our investment company is mainly made up of Alibaba partners, alumni of our Zhejiang College of Finance and Economics, and organizations of Zhejiang entrepreneurs. We began investing in the digital economy in ’16. I can address the moderator's question through one of our investments. As I recall, in ’18 we invested alongside Zheshang Venture Capital in Geely's Cao Cao Mobility, using an internet investment logic. At the time, we felt Geely might lack that internet way of thinking. But as everyone now knows, Geely is almost the country's largest automaker. This year it may challenge BYD or enter the world's top five automakers. The automotive industry is now the largest manufacturing industry. Looking back at our investment in Cao Cao Mobility, what we are now experiencing is that “the future has arrived”—intelligent driving, for example, and Geely's comprehensive embrace of AI. We can see a whole series of moves, from Geely's vehicle manufacturing to Qianxun Positioning, the currently very popular Xingji Meizu, and, in industrial internet, a project that has not yet been discussed much publicly—Guangyu Mingdao. There is now also commercial aerospace and Geespace.

Some major projects have also landed here in Zhejiang. Last year, I invested in a Geely robotics project and a new-energy methanol-powered vessel project. Through these projects, we can already feel how the traditional automotive manufacturing industry, like Musk's businesses, is becoming integrated across ground, air, and space. As I recall, Geely's Geespace and our Cao Cao Mobility were both in ’18 and had the same chairman. I asked the chairman then: Why did you think of doing satellites? He said: Musk is doing it. So I thought about how a traditional manufacturing giant—it is also a Fortune Global 500 company—could combine with Zhejiang, a province already ahead of the country in the digital economy, and with Alibaba, a world-class digital, software, and online company. We are now also receiving international recognition in AI. My son-in-law did his doctorate in the United States. I once asked him who was best at AI in the United States. He said Google, and said the best in China were ByteDance and Alibaba. So we bring those two sides together. Add Zhejiang's strong innovation and entrepreneurship environment, and I think our province has uniquely favorable conditions for traditional manufacturing to embrace AI.

Wang Chunfeng: Thank you. There is another question on industry integration and valuation: As AI becomes a digital employee, how should we assess a company's value? I think many people will consider and ask this. Our traditional investment logic might have involved price-to-earnings ratios, revenue, or other factors. What happens now with AI? The executives here have invested in many excellent projects. Mr. Yuan's firm, Saizhi Bole, invested in Deep Robotics, one of Hangzhou's six AI “Little Dragons.” Mr. Ni's Zheshang Venture Capital invested in DBAPPSecurity, and Zheshang Venture Capital is a long-established investment institution. Who would like to answer this question?

Yuan Zhiyong: Hello, everyone. I'm Yuan Zhiyong from Saizhi Bole. We are a Hangzhou-based firm that has been around for more than ten years. We essentially focus on early-stage technology. As the moderator said, we invested in Deep Robotics quite early, about six years ago. Returning to the question: Once AI empowers our employees, or even more digital employees start working for us, how should we assess value? For our firm, we rarely looked at price-to-earnings ratios even before, because we invest more at the early stage. At that stage, companies tend to have only a little revenue and no profit. When we invested in Deep Robotics, it had revenue of less than 20 million. When we invested in PingPong, it was at a similar scale. They were all loss-making. So we still follow our own logic: First, what will the future size of this market be, and what is its growth rate? Within that framework, we focus more on the team and the future market opportunity. That is the direction we follow.

Today, the first speaker also shared many companies working with OpenClaw—the “little lobster.” OpenClaw is a major topic now, and we will pay attention to it too. Our primary consideration remains what the company's future market niche looks like. Looking at today's agenda, many companies pitching focus on AI marketing, copywriting, and AI-generated short dramas. Beyond that, we return to this question, which also relates to Mr. Wang's point. We pay attention to opportunities that can generate profits in the short term, because the low barriers to starting a business can produce high investment returns, especially after the arrival of various AI tools. But domestically, within our investment focus, listing on the STAR Market or ChiNext is not something people can count on, or it is very challenging. So what else are we focusing on? How AI can combine with and further empower industrial manufacturing. Today's models are general-purpose large models. Whether it is Kimi, mentioned by the earlier speaker, China's DeepSeek, or ByteDance, these are foundational large-model capabilities. We focus more on industry scenarios and the ability to deploy large models within them: innovation in industrial companies or particular verticals, whether medical, legal, or scientific research. Within that innovation, our own employees can use AI tools to improve efficiency. That is our efficiency gain, and it can help companies move faster. So we still invest in this broad direction. Our approach does not change simply because a company becomes entirely an OpenClaw company. We still consider whether the market fits, its growth rate, and the team's characteristics. Those remain our considerations; the overall evaluation logic has not changed substantially. Thank you.

Wang Chunfeng: Thank you. I believe many people attending today, online or offline, want to start a business or already have a business that is doing fairly well. They will certainly be interested in where capital will invest next and whether the overall direction of AI investment will change. Everyone knows that companies' development depends on capital's support and assistance at every stage. From what I understand, the rapid development of Hangzhou's six “Little Dragons” has also been propelled by capital and government support. So people will want to know how our investment direction changes once AI provides that additional capability. I've organized three questions for the investors here. First, is capital moving from speculating on concepts to investing in infrastructure and applications? Second, do you favor computing power, hardware, and large-model algorithms, or applications that can make money quickly? Third, where do you think the next wave of emerging AI companies will come from? Let's first give the microphone to Mr. Ni, who has not yet answered.

Ni Min: I'm from Zheshang Venture Capital, a diversified investment company. As mentioned, Mr. Yuan's firm is particularly good at early-stage investing, with many star investments selected accurately and reliably. Mr. Wang's firm is good at deep integration with industry. We cover both early and middle stages, which relates to our company's origins. On AI, I think the whole country—and the whole world—is now experiencing “AI anxiety.” Government-led AI anxiety is more about competition between China and the United States. There is also AI anxiety among cities. Hangzhou has taken the lead in proposing that it become the leading AI city. What I would point out is that if this resonates in public discussion, it has one major benefit: From an investment perspective, a lot of capital will flow to the city. As long as capital flows here, it greatly helps both local companies and investment institutions. One significant change in AI companies over recent years, I feel, is the defining shift from the second half of last year to this year: from “showing off technology” to practical use. By showing off, I mean flashy performances—a robot dancing or performing a routine, for example. Ultimately, we want it to become a practical tool that assists people, and it needs an AI-related brain for that.

From an investment perspective, we see AI infrastructure as relatively certain. AI spans an extremely wide range: computing power, storage, optical modules, and much more. Why do we see it as highly certain? First, there is the gap between China and the United States. Having a benchmark gives us a direction in which to catch up, which is why we see certainty. NVIDIA held its GTC conference last week and announced several new products. The weakest product was for AI-assisted learning, with efficiency that could increase by 150%; the best delivered roughly a 10-fold increase. Under such rapid development, domestic AI companies will become even more anxious. I feel that most infrastructure projects—excluding the low and middle tiers, and looking at the middle-to-high end—are behind. Some are one or two years behind, others five to ten years. Catching up is not easy. That gives our investments a degree of certainty: The wider the gap, if your team is working on it, the more compelling we may find it and the easier the story is to tell. Where the gap is smaller, we are more likely to consider profitability and whether it makes money, assessing the various metrics. That covers computing infrastructure.

The other point is the team. As Mr. Yuan said, we too place great importance on teams, particularly AI teams now. Someone mentioned the young person from Google. In the past, when we looked for people returning after studying overseas, they had held positions in major foreign companies, perhaps as the head or deputy head of a department at some large US company. That is no longer the case. After the encirclement and sanctions, those head and deputy-head roles are no longer available. People returning now may be at a much lower level than before. They want to start businesses too. I think this year's defining change has created an investment opportunity for more AI enthusiasts: We may see companies with two or three people, or just one. One person can build a large company, so in the future going public may not be necessary. Think of livestreamers: One livestreamer is a large company. In the future, one person developing software or a little lobster—OpenClaw—can be an enormous company. If someone builds an enormous company alone, its capital or exit path may be relatively simple. We may see more and more companies like this, giving investment institutions an important direction: Find businesses with exceptional creativity that can deliver particularly surprising results through AI. I think there will be many. In particular, new faces now appear in the rankings every year. Today we may be talking about OpenClaw lobsters; next year we may be catching fish, and the year after something else. Things will change quickly. Wherever there is a use case, change will be rapid. We cannot imagine what next year will look like. Listening to young people, there is much we do not understand. Investors do not understand everything; some things really are difficult to grasp, and we have not learned them thoroughly. So we can only hold to our own logic: We learn what is clear and precise, and we need not learn everything that is not. That is my point: Our investment approach, particularly this year, has shifted noticeably from concepts toward practical use.

Wang Chunfeng: Thank you, Mr. Ni. I was embarrassed to admit that I could not understand some of it either. Now that you've said it, I can openly acknowledge that we also could not understand some of the very technical material shared by earlier speakers, so we must keep investing in learning. Because of time, I'll leave the third question—risk assessment and future predictions—to Mr. Wu. You have also been doing this for many years. I remember adding you on WeChat many, many years ago. You were already running Unique Capital then and had invested in quite a few projects. I'd like to leave this final question to you because iteration and change are extremely rapid in the AI era. Alongside choices about technology paths, there are ethical and regulatory considerations, which are very important, including the safety issues after OpenClaw—the little lobster—appeared. There are many risks, including the additional risk of tighter policy. Under these circumstances, how do you assess future investment risks, and what are your predictions?

Wu Wei: Thank you. First, I'd like to thank Secretary-General Chunfeng. Although we had been WeChat contacts for a long time, we only met in person recently. I feel you understand Hangzhou's entire startup and venture investment ecosystem very well, and it is an honor to exchange views on this panel. What Mr. Ni said made me happy. Why? The other guest, Mr. Wang, and I are of a similar age, which in that comparison makes us youngsters. In fact, I feel this is an especially good time for middle-aged people to start businesses. Why? They have already accumulated industry resources and experience. Combining those with AI's capabilities creates an excellent opportunity to empower traditional industries. I've met many AI entrepreneurs over the years. AI itself is a concept. If we talk about the recent OpenClaw phenomenon and the safety issues just mentioned, those have really emerged over the past month. Over that month, we have seen many companies building businesses around OpenClaw. Most are not new companies established within the last month. They had spent the past two or three years researching AI and building applications, tools, and various kinds of Agent. But when OpenClaw arrived, they could move into a new direction quickly. That was a very rapid response.

I see roughly three core opportunities here. First, the data entry point. Have you noticed that the key difference between people who use OpenClaw well and those who do not lies in memory and context management? In other words, your OpenClaw may be a professional while mine is an intern, so mine is naturally less professional. Those differences in data and memory mean we need to inject a great deal of industry know-how, experience, and knowledge into OpenClaw. A vertical To B version of OpenClaw could therefore be a very good future opportunity. That is the data entry point. Second, the medium entry point. We currently use it on web pages, and now in WeChat as well. In the future, we may use it in in-car systems or smart hardware, with an Agent and OpenClaw loaded onto those devices. Different media are also opportunities: Each medium is an entry point, whether communications software or any kind of smart hardware. Third, the interaction opportunity. Interaction is not just communication. We can change the premise: Suppose future software is not for people to use, but all software is for an Agent to use. Communication and transactions between one Agent and another would then emerge. For example, might there be a WeChat for Agents, an Alipay for Agents, or a Taobao for Agents? That could also be an enormous opportunity.

Of course, in this process there are two kinds of people. One finds this frightening and worrying: What if it surpasses humans and takes over humanity one day? The other uses it boldly. Whether we are talking about capabilities, safety, or concerns, we need to take them seriously and pay attention. But what matters more is actively embracing it and accepting the new opportunities that new things bring. Time always favors the prepared. In the Agent era and the AI era, I believe “speed is the primary force of productivity.” The faster you iterate, the more likely you are eventually to compete in particular niches with major companies such as Google and ByteDance, which we mentioned earlier. That is the only opportunity.

---

Original publication: https://uniqueresearch.substack.com/p/src-20260410-01html
On-site reading page: https://ffcap.cn/en/research/src-20260410-01html