Written by: Ada, Deep Tide TechFlow

In late 2024, Israeli developer Maor Shlomo was discharged from the reserves, opened his laptop, and began working on a project. He had no funding, no team, and no Slack channel. Six months later, Wix acquired his company, Base44, for $80 million in cash. At that time, the product already had 250,000 users and a monthly profit of $189,000. In the three months leading up to the acquisition, he hadn't written a single line of front-end code.

Pieter Levels from the Netherlands takes it even further. One person, with zero employees, using only basic PHP and jQuery, he simultaneously runs three products: Nomad List, Remote OK, and Photo AI. In 2022, his total revenue reached $2.7 million. He has never held a regular job, never raised capital, and lives a digital nomadic life in over 40 countries and 150 cities.

These stories are so good that they give you the illusion that one person plus a computer can build a skyscraper from scratch.

Incubators for one-person companies are springing up everywhere in Shenzhen, Shanghai, Suzhou, and Hangzhou. Nanshan Moli Camp, with 100,000 square meters, has attracted 700 applicants; Lingang Zero Boundary Magic Cube offers free workstations, and all 300 workstations in Phase I were snapped up, with Phase II of 8,000 square meters already underway. Dan Shao, which runs a one-person company community in Chengdu, has seen its events fully booked in less than a month.

The trend has indeed exploded. According to the "China OPC Development Trend Report (2025-2030)," as of June 2025, the number of single-member limited liability companies nationwide has exceeded 16 million. In the first half of 2025, the number of newly registered OPC companies nationwide reached 2.86 million, a surge of 47% year-on-year, accounting for 23.8% of all newly registered enterprises.

But what lies beneath the surface of this trend?

We spoke with three people who are on this path. One is an observer who has been running a one-person company community for nearly two years and has more than 2,500 real samples; another is a Gen Z entrepreneur who has run two companies in Silicon Valley across China and the United States; and the third is an independent developer who transitioned from a primary market financial advisor to an AI agent product developer.

The stories they tell are quite different from those on social media.

The underlying logic of success

Dai Wenqian runs SoloNest, a one-person company community in Shanghai, for a very simple reason. In June 2024, she had just left the online education industry and wanted to know what a person could actually do. Unable to find the answer in books and videos, she decided to conduct her own field research: organizing meetings, examining samples, and conducting interviews.

In nearly two years, she accumulated more than 2,500 samples, 20% of which successfully completed the business cycle and achieved results. She also wrote a book about her observations of the samples: "One-Person Company".

A joint study by Qichacha and Xiaobaochao shows that the three-year survival rate of companies established in China in 2021 had dropped to 71%, with nearly a quarter failing within the first three years. Clearly, SoloNest's 20% success rate has outperformed the market.

But Dai Wenqian cares more about the 80%.

“There are two types of people who didn’t succeed. One type tried to make it work but failed, either because the traffic stopped or the model became unsustainable. The other type never even started,” she analyzed.

The number of people who haven't started is far greater than you might imagine.

“Because he’s not in enough pain, he has a way out. His brain is deceiving him into thinking he wants to start, but he’s really just afraid of missing out. Most of them are worried about their current jobs and feel that OPC entrepreneurship might be a solution, but fear-driven motivation can be a bad starting point,” said Dai Wenqian.

Leon, who has worked as a financial advisor in the primary market, has seen many more people like this. He attended offline events for one-person companies and found that many lacked direction, attending events and buying courses indiscriminately. "No one can help you find how to make money. The only right path is to do it, to learn from mistakes, and to accept losses," Leon said.

Of those who started, who survived?

The answer is very counterintuitive. Dai Wenqian saw a stable commonality in the effective sample: those who successfully completed the closed loop were almost never doing what they used to do in their original industry.

They don't choose their path based on "what I'm good at," but rather on "where there are unmet needs."

Dai Wenqian used herself as an example. Previously, she worked in branding at Himalaya FM, without experience in offline events or community building. However, she possessed a curiosity about people, an aesthetic sense for product development, and the ability to express herself in a structured manner—fundamental abilities that transcend industry limitations. Once she identified market demands, scenarios, and industry pain points, she could transfer these fundamental abilities.

There's a guy in the SoloNest community who makes tennis bags; he didn't always make bags. Because he loves playing tennis, he discovered a need that wasn't being met by existing products on the market, so he invested 100,000 yuan to create an original tennis bag. After two years, he now consistently sells over 300 a month. This has a precise counterpart in Pieter Levels' methodology. In 2014, Levels set himself a challenge: to create 12 products in 12 months and see which ones would resonate with the market. Nomad List was the seventh and only one to succeed.

The key is not just choosing the right track, but also quickly validating enough hypotheses.

Dai Wenqian breaks down this process into three watershed moments. The first is whether you dare to create something by hand and throw it into the market; unfortunately, many people don't even have the mindset to test it, they just think about it but don't act. The second is whether you can sell it after someone is interested; there's a huge gap between "someone thinks it's good" and "someone consistently pays for it." The third is whether you can free yourself from the delivery process.

The first two filters eliminate those who are inactive or do not move.

The third stage is the real battle.

1.2 million trap

Those who pass the first two hurdles will find that they are alive, but there is a ceiling above their heads, firmly welded there.

Dai Wenqian gave a precise figure: the annual revenue ceiling for an individual relying solely on delivery is approximately 1 million to 1.2 million RMB.

"No matter how hard you work, there's a limit to how much time you can sell," she said.

This is the most realistic dilemma facing one-person companies in China. Social media talks about Maor Shlomo's $80 million exit and Pieter Levels' $2.7 million annual revenue, but those are stories of Silicon Valley SaaS and global digital products. In China, one-person companies are more prevalent in the consumer market, the service industry, and the experience economy, with heavy delivery chains and close interpersonal relationships.

Barry, who has done business across China and the US, has seen a more fundamental difference. American entrepreneurs focus on B2B SaaS and AI agents. Chinese entrepreneurs focus on tangible industries like pets, elderly care, and food. This isn't about who's smarter, but rather a difference in the industrial structures and willingness to pay between the two countries. American companies have a strong willingness to pay; they can make a small SaaS tool work, while China's B2B ecosystem is completely different.

So how do we break through the 1.2 million ceiling?

The most intuitive path is automation, using AI to remove itself from the delivery chain.

But this path is far more difficult than the narrative describes.

There's a typical example in the SoloNest community. Jason's business is job hunting coaching, helping interns and recent graduates in internet operations with resume revisions, mock interviews, and job search assistance. He started by selling his time, taking on a dozen or so clients a month.

Many competitors get stuck here and fail. Jason's approach was to find a group of competitors who struggled to acquire customers consistently, train them, and refer clients to them. Payment was split per order, with no employer-employee relationship. Dai Wenqian called this "multiple one-person companies collaborating, not one multi-person company." Later, he expanded into B2B outsourced operations, transforming from a purely C-end business into a C+B model.

Jason is now working on the third step: building a knowledge base using consulting data from the past two years to create a semi-automated product delivery system. However, he has only completed 60% of it in two months.

Why is it so slow? Dai Wenqian provided a mathematical model: "Suppose your delivery chain has 5 key nodes, and each node, through automation, can only achieve 80% of the work done manually. Does that mean the overall pass rate after automation is 80%? No, it might be 0.8 x 0.8 x 0.8 x 0.8 x 0.8, only 33%. The longer the chain, the harder it is to automate. It's not an additive relationship, it's a multiplicative one."

This is why lobsters seem like they can be automated all at once, but once you actually build one, you realize that if any step in the process isn't done well, the result will be garbage. The prerequisite for using AI effectively is that you've already done a good job manually; otherwise, you won't be able to see where the problems are.

Leon has the strongest technical background among the three interviewees. He now develops his own AI Agent product without writing a single line of code; all development is handled by AI. AI penetration in his workflow is close to 100%.

However, he was very restrained in his assessment of automation: "To evaluate whether a task can be handed over to AI, we look at three things: whether the cost is low, whether the risk is high, and whether the effect is good. AI cannot be used for services for high-net-worth individuals. The way AI works is that you allow it to make mistakes, and it optimizes its strategy by making mistakes. But in the helpless service of high-net-worth individuals, mistakes are not allowed. If you make a mistake in communication, the whole business is over."

Some business processes simply cannot be replaced by AI.

Dai Wenqian herself admits that her AI penetration rate is only 30%. This is because her core delivery involves offline, person-to-person interaction, which cannot be automated. What she can do is partial automation, including content acquisition and knowledge base accumulation, but she cannot completely detach herself from the business.

She works more than 14 hours a day. Her tasks include creating content to attract new users, chatting with and filtering people, maintaining relationships with partners, designing products, and disassembling samples. She also has two regular offline events every weekend.

“Many founders of one-person companies wouldn’t post this kind of thing online. Nobody would see it. Everyone likes to see glamorous images: drinking coffee here, going to exhibitions there, making millions a year, a powerful businesswoman. But the reality is that entrepreneurship involves a lot of dirty and tiring work, constant repetition, and constant iteration,” she said.

One-person companies are not the end.

Automation is one path, but not the only one.

Dai Wenqian observed another way to break through: not by replacing oneself, but by piecing oneself together.

Jason's case illustrates this logic. He doesn't hire people; instead, he collaborates with other one-person companies. Each Lego piece is an independent entity, with its own capabilities and customers, and together they create added value.

If every one-person company can be enhanced by AI, then putting them together is like adding enhanced Lego bricks. Dai Wenqian believes this is the biggest source of imagination for one-person companies: "It's like Lego bricks. Not every Lego brick needs to be 100% AI-enabled, but every Lego brick is enhanced by AI. Putting three enhanced Lego bricks together is not 1+1+1, but 3×3×3."

Another approach is to share your experience and methodology with more people. Barry's practice validated this model. He is the founder of two one-man companies, exploring everything from scratch. Once the business cycle was established, he withdrew and handed it over to the team, allowing them to take over like a relay race, while he went on to run other businesses.

Maor Shlomo made a similar choice. Base44 grew to 250,000 users and nearly $200,000 in monthly profit within six months, but he still chose to sell it to Wix. His explanation was that although the growth was phenomenal, the scale and size we needed could not be organically achieved by one person. One person can take a product from 0 to 1, but going from 1 to 100 requires organization, resources, and distribution capabilities, which is something one person cannot do.

Three different paths: AI productization, collaborative integration, and partner expansion, but the underlying logic is the same: a one-person company is not the final state. It's a springboard. Once you've validated something at the lowest cost and it's working, you must find a way to move beyond that bottleneck. Otherwise, you'll be forever stuck on that 1.2 million line.

Before the door closed

The data for 2026 is impressive. Shenzhen issued an OPC startup ecosystem action plan, aiming to build more than 10 OPC communities of over 10,000 square meters each by 2027. Shanghai Pudong offers up to 300,000 yuan worth of free computing power to newly registered one-person companies. Suzhou attracted 300,000 university students in 2025, and its talent pool is rapidly expanding.

But Dai Wenqian said something that made people realize the truth.

"The barrier to entry has been greatly lowered. In the past, finding money, people, and venues was extremely costly to start up. Now you can validate anything at almost zero cost. But this benefits everyone equally. If it's easy for you, it's easy for others too. With more players, traffic becomes more expensive. It's an arms race."

Pieter Levels earns $2.7 million a year because he started in 2014 and built a decade-long SEO moat and built community trust. Maor Shlomo of Base44 was able to sell his company in six months because he had previously run a data company that raised $125 million; his network, judgment, and speed were not built from scratch.

These people are not the "ordinary people" portrayed in the one-person company narrative. They are among the brightest points of survivorship bias.

The real world of one-person companies is exemplified by the 2,500+ samples in the SoloNest community: 20% are consistently making money and moving to the next stage; 40% are stuck in various ways but are still iterating and trying to break through; and another 40% are still lost and searching for direction. Of the surviving 20%, most earn less than 1.2 million annually. They work until the early hours of the morning, their weekdays are packed, and they have no weekends.

In reality, the business of a one-person company profits from the time gap between when a niche demand is discovered but before it is captured by organized capital. This time gap has a name: shelf life.

The shelf life depends on two things: when you discover the need and how quickly you implement it.

Lowering the barrier to entry won't extend the shelf life. On the contrary, it will shorten it. Because what you can validate at zero cost, others can also validate at zero cost. The needs you see, others can also see. If you create an MVP by hand today and it survives for three months, tomorrow ten identical products will appear on the same user's phone.

This is why most people get "stuck." The essence of being stuck is not a matter of ability, but that the speed at which time is realized for them cannot keep up with the speed of market congestion.

The reason Maor Shlomo and Pieter Levels are advertisements rather than examples is precisely because they solved the shelf life problem in two opposite ways. Levels extended its shelf life to ten years through first-mover advantage and compound interest, while Shlomo compressed its shelf life to six months through speed and exit strategy.

The middle path is the most dangerous. For most founders of one-person companies in China, there's neither a decade to slowly build a flywheel, nor does Wix come and write checks. They work 14 hours a day, maintaining a ceiling of 1.2 million, thinking that as long as they hold on a little longer, they can break through. But the market won't wait for you to hold on. The next competitor with zero-cost validation can appear at any time, flattening your little moat.

A one-person company is never a state that can be maintained for a long time; it is a window with an expiration date.

When the window of opportunity opens, the barriers to entry are low, tools are inexpensive, and the needs are clear—it seems like the best of times for ordinary people. But the window won't stay open forever. It will be filled with people who come in later, crushed by more efficient tools, and finally closed completely by a startup that has secured funding or a business line from a large company that suddenly exits the market.

The only real challenge in this business is whether you can get yourself out of that bottleneck position before it shuts down.