The shape of companies will still exist. But the underlying logic has already been completely reshaped by AI.
Author | Jinguanghao
Editor | Jingyu
In early 2024, OpenAI CEO Sam Altman made a bold prediction in an interview:
“In the AI era, a single person could build a unicorn company valued at $1 billion.”
By May 2025, at Sequoia Capital’s AI Summit, that idea returned to center stage:
The future may deliver the first true “one-person unicorn.”
Over the past two years, real-world trends have started to validate that direction.
Carta’s 2025 data marked a historic milestone: more than one-third of new companies are now founded by solo founders.
The share rose from 23.7% in 2019 to 36.3% in the first half of 2025—an increase of 53% over six years.
Terms like “super individual,” “independent developer,” “solopreneur,” “one-person company,” “OPC (One Person Company),” and “solo founder” are appearing more frequently than ever.
The rules of entrepreneurship are being rewritten.
1
In traditional thinking, starting a company meant building a team, finding a co-founder, and constructing an organization from day one.
Silicon Valley’s classic storyline is two brilliant people changing the world from a garage—Jobs and Wozniak, Page and Brin, Bill Gates and Paul Allen. These stories shaped how we imagine entrepreneurship.
Investors also tended to prefer co-founder teams, because they implied shared risk, complementary skills, and internal checks and balances.
But in the AI era, the rapid emergence of “one-person companies” is challenging that default script.
In practice, a one-person company does not literally mean a business that will always have only one human involved.
A more accurate definition is this: a company led by a single founder who makes the core decisions, while leveraging AI and external resources to expand the boundaries of what one person can achieve.
In other words, the founder doesn’t need to do everything alone. The real skill is orchestrating tools, partners, contractors, and platforms across the value chain.
The core advantage is powerful: a founder doesn’t have to lock themselves into a “good enough” co-founder at the very beginning. Instead, they can validate the business model first, then build a team gradually as the business proves itself.
As AI tools mature, this “start moving first, refine later” strategy has become genuinely viable.
One solo founder in AI music is a good example. She has no permanent employees, but through AI tools and a global collaboration network, she can integrate creation, production, and delivery end-to-end. When urgent projects come in, she can respond immediately by combining AI workflows with external specialists—often faster than traditional music companies.
That’s the core logic of a one-person company:
One person can operate like a full team.
2
Why has the one-person company model surged so sharply in the last two years?
There are a few key reasons.
First, AI dramatically increases what one person can do.
As tools like DeepSeek helped reduce AI usage costs, adoption in everyday work and life accelerated noticeably.
Several reports capture the shift. A Tencent Research Institute report in September found that 96% of people have used AIGC products, with nearly 70% using them daily.
Another September report from Soul showed that 99.2% of university students have used AIGC products, and 66% instinctively ask AI first. For many young people, AI has become the default first response when a problem appears.
Even more important: AI’s capability boundary is expanding.
2025 is widely seen in the industry as “the first year of Agents.”
AI is evolving from a helper tool into a “virtual partner” that can make decisions and execute workflows. Tasks that once required team collaboration—coding, design, copywriting—can now be handled by a founder who orchestrates a group of AI agents.
Second, the cost of starting a business keeps falling.
According to the “2024 Global One-Person Company Industry and Investment Ecosystem Observation,” 90% of one-person company founders started with less than $500.
Cloud computing, SaaS tools, and open-source models mean founders no longer need to build infrastructure themselves.
AI subscriptions are increasingly structured like mobile plans: a base monthly fee plus usage-based pricing. This flexible cost structure allows a one-person company to launch with close to zero fixed overhead.
In the past, founders had to raise money, form a team, and rent an office before they could even test an idea. Today, one person with a laptop can start validating a business immediately.
Policy support is also starting to follow. For example, Shanghai Lingang launched the “Super Individual 288 Action,” offering “zero-rent” startup space for OPCs. More recently, Shanghai Zhangjiang AI Town and Beijing Zhongguancun have introduced additional OPC-focused support measures.
These signals suggest that the one-person company model is becoming a real entrepreneurial trend—not just an internet concept.
Third, success stories are creating a strong demonstration effect.
Vercel’s Guillermo Rauch and Pieter Levels (founder of Nomad List and RemoteOK) both started solo, proved product-market direction, and only then expanded their teams.
As more independent founders raise funding and achieve meaningful exits, the market is validating this path. Investors are beginning to reconsider the assumption that a company must have co-founders to be credible.
Founders are also asking a more direct question:
Do I really need a co-founder on Day 1?
3
Traditional startup culture values co-founders because, in theory, the model offers major advantages:
Shared risk, complementary skills, balanced decision-making, and emotional support.
But theory and reality aren’t the same.
Misaligned vision, equity disputes, and decision conflicts between co-founders are among the most common reasons startups fail.
In The Founder’s Dilemmas, Noam Wasserman’s research shows that 65% of startup failures stem from conflict within the founding team. When founders fundamentally disagree on direction, resource allocation, or fundraising strategy, the company can slide into internal friction—or even break apart.
This risk is especially deadly early on.
An even more subtle trap is “settling.”
Many founders rush into finding a co-founder because “investors prefer teams,” “it’s lonely,” or “I need someone to share pressure.” But if the co-founder isn’t truly aligned or genuinely complementary, they can become a drag rather than a multiplier.
As one founder, Douglas, described it: “Finding a co-founder is like finding a spouse—rushing in often leads to a painful breakup.”
Of course, one-person companies also have challenges.
Loneliness is the most frequently mentioned. There’s no one to debate late-night decisions, no one to share anxiety with, and pressure concentrates on a single person.
But this isn’t unsolvable.
Founder communities, mentor networks, and peer groups can provide emotional support without sacrificing decision independence. For example, Shanghai’s SoloNest community has hosted more than 100 offline events and gathered over 4,000 one-person company founders who share learnings and encourage each other.
The key is simple: don’t search just to search. If the right co-founder doesn’t appear, starting independently may be the better move. Once the business reaches a certain stage, you can still attract core members through equity incentives.
Data supports this flexibility. Research from solofounders shows that the median equity incentives granted to early employees (the first five hires) in solo-founder companies are very close to those in co-founder companies.
A founder who starts with 100% equity can use that advantage strategically—offering compelling terms to recruit top talent who can later grow into true partners.
4
The rise of one-person companies points to a deeper question:
In the AI era, are traditional organizational structures still the best fit?
Modern company structures were built in the industrial age, designed to boost efficiency through division of labor.
The logic was straightforward: one person can’t do everything, so you hire people. Once you hire, you divide work. Once you divide work, you need coordination—departments, hierarchy, processes.
AI is changing this logic at the foundation.
When one person can use AI to complete work that once required multiple specialists, the necessity of “organization” weakens.
When AI can provide cross-domain expertise, the boundaries of specialization blur.
When AI agents can automate coordination, traditional management layers need to be rethought.
Manus is one of the fastest-growing general AI agents overseas this year. Just eight months after launch, it reportedly surpassed $100 million in ARR. Its defining feature is that it doesn’t just execute single tasks—it can plan, call tools, and complete complex multi-step workflows on its own.
If this capability continues to evolve, the combination of one person plus a set of AI agents could realistically accomplish work that previously required a full company.
Today, a typical one-person company stack might look like this:
Claude for coding, Gemini for front-end output, GPT for content, Notion AI for project management, and n8n for workflow automation.
One founder, a handful of subscriptions, and a monthly cost that may be under $500—yet the output can rival what a small team used to deliver.
None of this means companies will disappear.
Building chips, running factories, and operating heavy-asset businesses still require organizations.
But the form of organizations is likely to change.
McKinsey’s September 2025 report, The Agentic Organization, suggests that future organizations may shift away from pyramid hierarchies toward outcome-driven, dynamic networks made up of mixed human-and-AI micro-teams.
A “typical” team might include only 2–5 core human members, while managing 50–100 AI agents that execute full business workflows—from customer onboarding and product launches to finance and settlement.
5
Still, one-person companies are not a universal solution—and they’re not automatically the best solution.
The right path depends on what you’re good at, what you care about, and whether what you’re building creates real value.
One independent SaaS developer shared that he once worked 14 hours a day for three straight months because “everything is my responsibility.”
Later he realized the real essence of a one-person company isn’t “one person does all the work.”
It’s “one person decides what should be done—and what should not.”
According to public sharing by Karen, the organizer of the SoloNest community, among more than 2,000 offline samples she has seen, only about 20% earn money steadily.
AI can raise the odds of success—but entrepreneurship is still hard.
If you’re considering the one-person company route, it’s worth asking yourself three questions first:
Can I handle months where there’s no one to directly co-decide with?
Am I willing to learn and continuously iterate my AI toolchain?
Do I clearly understand my own capability boundaries?
If your answer is “yes” to all three, you may be more suited to this path than you think.
6
In recent years, the idea of the “super individual” has been widely discussed.
With AI maturing, it’s pushing this concept into a new stage: it’s not only that one person can do more work, but that one person can operate like a company.
That means more people can create business value independently—without relying on traditional organizations.
Of course, one-person companies won’t fit everyone or every industry.
They’re better suited to founders with strong self-motivation, a willingness to keep learning, and the ability to make decisions independently. They also tend to fit knowledge-intensive, creativity-driven, and highly digital fields.
For founders who have ideas and skills but have been waiting for the “right partner,” this might be the moment to reconsider the assumption.
Instead of waiting for perfect timing and perfect people, you can use AI to get the first version moving.
If you want to try building a one-person company, start with a weekend project.
Use AI to build a minimum viable product, publish it on Product Hunt or Xiaohongshu, and test whether anyone is willing to pay.
Thinking creates questions. Building creates answers.
