Original · Unique Research · 2026-03-09
Editor's note: This is the complete English rendition of the cited edited roundtable recap, not an unabridged recording transcript. First-person narration and evaluative statements belong to the original author; “lobster” refers to OpenClaw. Portfolio survival, skill counts, productivity multiples, financing rounds, AI-adoption percentages, China's share of open-source output and the attributed Steve Jobs quotation are retained as source/speaker claims without independent datasets or verification. The source gives both “95% of companies have never used AI” and “95% of companies have AI penetration below 5%”; these are different statements and have not been harmonized. Forecasts about AGI, preinstalled PCs, populations and business opportunities reflect the March 2026 discussion, not present facts or investment recommendations. Zhu Xiaohu is rendered consistently in pinyin from the source's Chinese name; Xialiao is a romanization of the co-organizer's Chinese name.
Three Top Investors Say: The Lobster Has Arrived, and Entry Tickets Are Running Out
OpenClaw Hardcore Lobster Party · Shenzhen | UniqueResearch × Xialiao × MiniMax
“If you still haven't installed the lobster today, you are basically in serious trouble.”
This was not some entrepreneur peddling anxiety from the stage. It was said onstage by a top investor who manages a dual-currency fund and has invested in 80 companies without a single failure.
Then she added:
“The people who use the lobster well are also people with a foundation. A three-year gap cannot be closed overnight.”
You think you still have time. You may not.
On the afternoon of March 8, in Shenzhen's Qianhai district.
An “OpenClaw Hardcore Lobster Party” was under way—co-hosted by Xialiao, MiniMax, and UniqueResearch, with 300 hardcore developers, researchers, and founders crowded into the same space to discuss agent architecture, swarm intelligence, and edge AI.
At 4 p.m., three mystery investors made a surprise appearance.
One was Zhu Xiaohu, managing partner at GSR Ventures—one of the most active investors in China's AI sector, who has been present for every technological wave from Didi to the ChatGPT era.
Another was Cynthia Zhang, founder of FutureX Capital, who invested in 80 companies over the past three years without a single one shutting down.
The third was Bai Zongyi, founding partner of Glory Ventures—an engineer by training who focuses on hardware, semiconductors, and autonomous driving, and the calmest person in the room.
The moderator was Wu Wei, founder of UniqueResearch.
What follows distills the parts most worth reading repeatedly.
01 This Is Exactly Like the Moment the Mobile Internet Took Off
Zhu Xiaohu was the first to speak.
He said that what shook him about the lobster, or OpenClaw, was not the power of the product itself, but the growth rate of its ecosystem.
“Look at this past month: hundreds of thousands of new Skills, worldwide. The speed is insane.”
Hundreds of thousands. One month.
He went on to say that within the next one to two months, AI PCs would come with the lobster preinstalled. When that day arrives, the battle for the gateway will truly begin.
He used one phrase: “It is exactly like when the personal internet took off.”
Back then, projects went from development to launch in less than a month, and many were built by one or two people.
The same is true now.
The only difference is that the leverage back then was the PC, while today's leverage is AI. It is not merely a little greater; it is greater by more than an order of magnitude.
02 The Lobster Is Not Something New; It Is an “Amplifier”
Cynthia offered one judgment that was the sentence most worth chewing over repeatedly at the entire event:
“The lobster did not change anything. It accelerated every judgment we made about AI three years ago.”
Last year, she began looking for AI Agent opportunities across industries and examined many of them, but the friction was too great—every industry was an island, and an Agent could not be deployed in another industry.
After the lobster arrived, those islands began to connect.
“For teams that already had a strong understanding of AI, the arrival of the lobster increased their advantage over traditional teams from 100-fold to 10,000-fold.”
From 100-fold to 10,000-fold. This is not a metaphor; it is the reality she has observed in her actual portfolio.
One AI manufacturing company in which she invested has a team that uses the lobster very well. She used one word: “crushing.” It is crushing traditional manufacturing.
“It is like higher-dimensional beings fighting lower-dimensional ones in The Three-Body Problem.”
03 The People Selling Shovels Are More Clear-Headed Than the People Digging for Gold
Bai Zongyi was the calmest person in the room.
Trained as an engineer and having invested in semiconductors, autonomous driving, and SiliconFlow, he is accustomed to examining problems from the foundations.
His core judgment came down to one sentence: Token costs are still not low enough, and this is not over yet.
He mapped out two principal tracks:
Cloud: inference costs will continue to fall. New opportunities include dedicated browsers for Agents (“Today's webpages are made for people; future webpages will be made for machines”), high-quality data providers (because behind models that perform increasingly well on leaderboards are increasingly capable experts supplying high-quality data), and multi-model routing-optimization platforms.
Edge: in his judgment, today's greatest bottleneck is not compute but the bandwidth divide, or Bandwidth Gap. People buy Mac mini computers and install the lobster essentially because the chip is “just barely sufficient.” A year and a half ago, he incubated a company making 3D-stacked memory chips; it completed 5 financing rounds within three months. That is the logic.
Wu Wei summed up his logic: “While everyone else is digging for gold, you invest in the people selling water.”
Bai Zongyi nodded: “Water, shovels, and tools—those are the constants.”
04 The Way Forward for Software Companies: Build for Machines, Not for People
Zhu Xiaohu made a bold declaration:
“In the future, we will not need websites for people to see. Every website will be for machines and for AI.”
Software companies that design websites are already beginning to transform—not by making interfaces prettier, but by designing content structures that AI can understand and invoke more easily.
This is a completely different product philosophy.
He then discussed competition between startups and major companies and made a very “practical” observation:
“Domestically, there is basically no visible opportunity to compete head-on with ByteDance. ByteDance simply does not make mistakes and does not give you an opening. So you can only start by finding a small gap, enter it, grow very deep, and then see whether you can build your own ecosystem.”
His advice was to first ride the lobster's traffic, establish a moat in a particular vertical use case, and then plan what comes next.
This is not pessimism; it is clear-eyed thinking.
05 The Biggest Gold Mine Lies in the Traditional Industries You Look Down On
Cynthia shared an observation that deeply affected me.
She recently spoke with several entrepreneurs working on lobster-related ventures and found a common problem: they do not understand real businesses or industries. They wander around the virtual world, and all the opportunities they can see are very small ones.
“For example, helping someone install a lobster.” That is their ceiling.
But where is the real gold mine?
“The more traditional and more difficult an industry is, the friendlier it actually is to AI entrepreneurs.”
Healthcare, manufacturing, energy, and electric power—in these industries, AI penetration may still be below 5%. Fully 95% of companies have never used AI.
When you enter, you are bringing a higher dimension to fight a lower one.
She cited one data point: Last year, AI penetration in US healthcare surged from 4% to nearly 30%. Within a single year, how many startups secured more than US$100 million in financing?
That story is beginning to repeat itself in China's traditional industries.
06 Has AGI Arrived? That Is the Wrong Question
I asked a standard question: Has AGI arrived?
The three investors gave completely different answers.
Bai Zongyi said: “In an era of rapid change, do not use end-state thinking to judge whether a field is worth investing in. The money is made during the process.”
Zhu Xiaohu said: “You need the courage to enter and learn through the process.”
Cynthia gave the longest answer, and also the most interesting one.
She said that her friends at top Silicon Valley Labs—the people who helped build the world's leading large models—told her they believed they might lose their jobs within a few months, because even they could not control the pace of AI iteration.
In their eyes, AGI arrived long ago.
But across China's thousands of industries, 95% of companies still have AI penetration below 5%.
So has AGI arrived or not?
Cynthia's conclusion was that, as an investor, she never cares when AGI will arrive. What she cares about is this:
“After the lobster arrived, AI adoption across thousands of industries shifted from 'if' to 'when' and 'who will do it.' Most companies in traditional industries cannot implement AI well by themselves. Every one of these is therefore an opportunity for startups.”
07 What Is Most Irreplaceable in a Person in This Era?
Near the end, I asked what may have been the most important question of the entire event:
In the AI era, what quality in a founder cannot be replaced by AI?
The three answers struck me as worth recording.
Zhu Xiaohu said: “Carry no baggage; dare to launch and dare to make mistakes.”
In the AI era, the quality he values most in entrepreneurs is actually quite simple: carry no baggage. Teams with impressive backgrounds and large amounts of funding are often instead held back by “perfectionism,” afraid to launch their products for so long that they end up producing nothing. By contrast, he prefers entrepreneurs who dare to launch products quickly, continually make mistakes, and keep iterating. The AI era changes too quickly; if you wait until everything is ready before acting, the opportunity has often already passed.
Bai Zongyi said: rapid iteration. From the perspective of technology entrepreneurship, he emphasized that in the AI era, the most important ability for many application-layer companies is not getting it right in one attempt, but continuing to iterate. The market changes extremely quickly. Whether a founder can rapidly adjust direction, keep experimenting, and produce multiple versions within a few months often determines whether the company survives. For him, an excellent AI entrepreneur must be able to learn rapidly, experiment rapidly, and iterate rapidly; otherwise, it is difficult to keep pace with changes in technology and the market.
Cynthia gave a higher-dimensional answer: Wisdom.
She drew a distinction—Intelligence and Wisdom are not the same thing. AI intelligence has already surpassed human intelligence, but wisdom exists on another dimension.
She said wisdom includes how you manage your relationships with people and AI; what your sense of mission is; and whether you start a company to become famous, to make money, or because you genuinely want to leave something behind for the world.
“The lobster is an amplifier. If you originally score 10, the lobster turns you into 100. But if you already have problems, the lobster will amplify those problems too.”
Selected Q&A Transcript
Below are selected question-and-answer excerpts from the roundtable, preserving the granularity of the original conversation so that you can make your own judgment.
Q (Wu Wei): Has the lobster's arrival changed your investment logic in any way?
Zhu Xiaohu: I think it has given everyone a very good example, which is not to be impatient. Look at how explosive the PC era was. When the PC internet emerged, conferences were held every day; the same was true when the mobile internet emerged, with internet conferences held at Beijing's stadium. By today, very few have survived. The key is to enter at the right time, and now is an excellent time.
Cynthia: My investment logic has not changed in essence—I still invest in people with insight. The biggest change is that the lobster made me accelerate. It sped up all the prior Agent judgments. I previously looked for Manus opportunities in every walk of life and Agents in different industries, but always felt they could not be implemented because every industry was an island. With the lobster’s arrival, AI agents in these industries will be deployed rapidly. My investment pace therefore also needs to accelerate.
Q (Wu Wei): Will software companies, websites, and Apps still have value in the future? Will the interaction model change?
Zhu Xiaohu: The gateway is always changing. In another year, the lobster may be the terminal gateway—OpenAI itself does not even have one, and now it can be layered directly on top. Future websites will not need to be designed for people; every website will be for machines. Software companies that design websites are already transforming today, saying, “My design can be presented to AI and does not need to be seen by people.” This is a completely different form of business.
Bai Zongyi: Six months ago, I invested in a company building an AI Agent Browser. Today's browser is highly inefficient for people, resembling serial computation on a CPU: you can view only one webpage at a time. But Agent invocation is completely different—it does not view advertisements and it invokes multiple Agents simultaneously. How do we solve isolation and security in that environment? This is an entirely new Infra opportunity. When everyone opens a webpage today, do you look at the borders? Do you look at the ads? In the future, you will look at none of them.
Q (Wu Wei): How can startups compete with major companies within the lobster ecosystem?
Zhu Xiaohu: To be honest, it is difficult. I think you first need to position the product a little more narrowly and ride some of the lobster's traffic, then see whether you can build your own ecosystem after capturing that traffic. Unless you are exceptionally good and suddenly break out, there is almost no opportunity to compete head-on with a major company—ByteDance simply does not make mistakes and gives you no opening. The most reliable approach is therefore to find a small gap, enter it, and grow very deep.
Cynthia: By contrast, I am more optimistic about young companies applying AI in traditional industries. I do not mean a traditional company saying, “I am going to do AI,” but a team that came from a major company, has very deep industry knowledge, and is itself AI native, and it must be AI native. If I look at a founder today and that person has not been using AI seriously during the past three years, there is basically no reason to invest.
Q (Wu Wei): How do you view China's position in the AI competition?
Cynthia: I believe the real competition is not competition between China and the United States, but competition between open source and closed source. This is an even greater danger than competition between the two countries. Along the open-source track, China is leading—70% of open-source output now comes from China.
Zhu Xiaohu: The key is whether China can create an ecosystem like the lobster's. The lobster's success lies precisely in its ecosystem—no matter how capable one team may be, it cannot match an ecosystem. Hundreds of thousands of Skills can appear in one month; a single team cannot compete with that. If China can build such an ecosystem, it will be extremely important.
Q (Wu Wei): Is the ceiling high enough for a one-person company? Can VC invest in one?
Cynthia: I would absolutely invest in a one-person company, and one-person companies are extraordinarily valuable. Steve Jobs once said that he thought an excellent engineer produced 4-5 times as much as a mediocre engineer, but later discovered he was wrong: it was 20-30 times. Today, I believe the multiple is already in the thousands or even tens of thousands. One sufficiently capable person, accompanied by AI, can fully support an entire company.
Wu Wei: But funds have limited lifespans, so the ceiling is a problem.
Cynthia: I believe the lobster has changed that logic. For a one-person company, if that person is sufficiently capable, the ceiling is no longer a problem today.
Q (Wu Wei): Has AGI arrived? How far away is it?
Bai Zongyi: One relatively major mistake we have made internally over the past two years is using end-state thinking, in an era of rapid change, to judge whether a field is worth investing in. If you want to think about AGI's ultimate state today, I think you should simply stop. It truly is not that important; the money is made during the process.
Zhu Xiaohu: Learn through the process and dare to enter. That is what matters most. Many opportunities were missed because people did not dare to enter.
Cynthia: Friends working at top Labs in Silicon Valley feel that they may lose their jobs within a few months—in their eyes, AGI arrived long ago. But across China's thousands of industries, 95% of companies still have AI penetration below 5%. So has AGI arrived? As an investor, what I focus on is not when AGI will arrive, but what value it creates and how we seize opportunities amid this change. The lobster's emergence has greatly increased the certainty of AI adoption across thousands of industries. In the future there will be 10 billion humans and perhaps 100 billion AI agents; every system and interaction will be new, and the opportunity is enormous.
In Closing
One month ago, many people were still asking, “Is the lobster just a passing gust?”
Today's conversation confirmed one thing for me:
This is not a gust. It is a platform-level migration.
Like the migration from the PC to the mobile internet, it does not ask whether you are ready; it simply keeps moving forward.
The three investors emphasized different things, but were highly aligned on one point: the opportunities are real, and there are too many to finish examining in time.
There is only one real question: Are you someone who has used AI continuously for the past three years, who has taste and wisdom, and who carries no baggage?
If so, welcome to this era.
If not, you probably have only a little time left.
This article was compiled, with edits, from a recording of the “OpenClaw Hardcore Lobster Party · Shenzhen” roundtable.
The event was co-hosted by Xialiao, MiniMax, and UniqueResearch in Shenzhen's Qianhai district on March 8, 2026.
Roundtable guests: Zhu Xiaohu (managing partner at GSR Ventures), Cynthia Zhang (founder of FutureX Capital), and Bai Zongyi (founding partner at Glory Ventures)
Moderator: Wu Wei, founder of UniqueResearch