My partner is an AI.

In a residential building on Yuyuan Road in Changning District, Shanghai, entrepreneur Ye Chunjun runs a technology company with millions of users worldwide.

There are no busy employees here, only flashing screens and Ye Chunjun himself. However, in his words, this company has never “fought alone” because Ye Chunjun's team has a group of special partners – AI.

Over the past year or so, Ye Chunjun has gradually built a new set of corporate operating logic using AI. His company focuses on providing AI-driven interactive content and emotional companionship products. Since its operation, this company with millions of users has maintained a stable monthly income between $20,000 and $30,000.

Before the end of the summer of 2025, Chen Zhiyue, a graduate of Communication University of China, was busy looking for a place to stay in Hangzhou. The house she needed was not large, but she required that “there must be a study room” to house her “one-person company.”

She developed a single-person live streaming software called Mirrorfy. Users can use this software to start a virtual live stream anytime and anywhere. The software will use AI to generate virtual audiences and interact with users through bullet comments.

Mirrorfy was born out of a “72-hour AI survival challenge” that Chen Zhiyue participated in. At that time, she needed to survive for 72 hours in a closed room in a homestay in Pudong, Shanghai, relying on AI tools. During these 72 hours, Chen Zhiyue, who did not have mature programming skills, completed most of the code for the product with the assistance of AI. During the process, AI also designed the logo for Mirrorfy.

Mirrorfy interface, Image source: Provided by the interviewee

In recent months, “Jiazi Guangnian” has contacted several young entrepreneurs, and they have a common feature: their partner is AI.

In the Internet era, with the emergence of cloud computing and distributed digital infrastructure, some “one-person companies” were born. However, in this wave of the AI boom, as the capabilities of AI models continue to strengthen and the implementation of AI Agents, the meaning and possibilities of “one-person companies” have been newly expanded.

“Jiazi Guangnian” believes that a “one-person company” in the AI era is often led by a single or a small number of founders. With the help of AI models or AI Agents, most business processes are automated, and with minimal human resources, it realizes a business form of sustainable operation and growth that previously required the collaboration of multiple people.

There is also a bolder conjecture: Can a “unicorn” emerge from a “one-person company” in the AI era?

1. Everything has changed

Last year, when talking with Alexis Ohanian, the co-founder of Reddit, Sam Altman, the CEO of OpenAI, shared an interesting story.

“In our small group of tech company CEO friends, we set up a bet on when a one-person-operated billion-dollar company would appear,” Altman said. “This would be unimaginable without AI – and now it's about to become a reality.”

Generally speaking, an unlisted startup company that has been established for no more than 10 years and has a valuation of over $1 billion can be considered a “unicorn” company.

In May this year, Dario Amodei, the CEO of Anthropic, made a bold prediction about “one-person companies.” He expects that new AI models will enable “one-person companies” to thrive like never before. When asked if it was possible for “a person to create a company valued at $1 billion using AI,” Amodei's answer was: “This will definitely happen, possibly as early as next year.”

Regardless of when the “one-person unicorn company” will appear, the entrepreneurial logic in the AI era has changed.

“When I transformed AI from a tool into a ‘smart partner,’ everything changed.” Ye Chunjun told “Jiazi Guangnian.”

Ye Chunjun's desk, Image source: Provided by the interviewee

In Ye Chunjun's view, the motivation for the emergence of “one-person companies” in the AI era is different from that in the Internet era. It does not originate from a temporary measure to save costs but from entrepreneurial ideas that were once difficult to realize because AI has commodified the key production factors (such as engineer teams, computing power, and data) for creating and operating a technology company.

This means that entrepreneurs can bypass the complex form of the traditional “formation and management of human teams” and build a “human-machine collaboration” business system with efficient capital and operational efficiency by directly purchasing computing power and AI capabilities.

AI Agents can embed human work processes into software, freeing up human time to do more things in a shorter period. Entrepreneurs can assign tasks to AI Agents, and they can make decisions with varying degrees of autonomy. Multiple AI Agents can even collaborate on complementary tasks to complete real work.

For many entrepreneurs, AI partners can not only provide technical value but also emotional support.

A Hua (pseudonym), an entrepreneur of a “one-person company,” frankly told “Jiazi Guangnian” that AI is her most loyal entrepreneurial partner.

When she is confused about product direction, pricing, or even roadshow work, she will voice all her inner thoughts to AI and let AI sort out her chaotic thoughts. “It is always clear, always caring, and won't ask for equity, so I can pour those messy troubles that even I dislike into the dialog box without any embellishment.”

A Hua said that AI will relieve her decision-making pressure and has a charming quality that makes people rely on it: “AI won't take credit, nor will it deny you. It will only keep providing options. This ‘purity’ makes people dare to throw in their most original and naive ideas.”

Yuan Liuwei, an entrepreneur of a “one-person company” in Wuhan, Hubei, told “Jiazi Guangnian” a story of being “saved” by his AI partner.

Yuan Liuwei runs a knowledge payment company. His business revolves around how to help customers make good use of AI. Yuan Liuwei doesn't hire employees, and the whole company is like a modern “digital workshop.”

In the second half of 2024, Yuan Liuwei's AI introductory course business was on the right track. At the same time, Yuan Liuwei smelled a crisis. The business data indicated that the company's commercial orders began to show a decrease in quantity and stagnant growth.

At the critical moment of making a decision, Yuan Liuwei turned to AI for help.

“How can I iterate my business model?” Yuan Liuwei asked Claude of Anthropic, which he regarded as his “AI partner.” After just a few exchanges, Claude helped him summarize the countermeasure: develop B2B business.

“There are a continuous stream of commercial orders in B2B training. If you do a good job in building your reputation and brand, it will naturally attract more B2B business opportunities and also drive C2C students,” Yuan Liuwei summarized the result of his discussion with AI. This has become his decision-making idea that he has been following ever since.

After the AI partner identified the key direction, Yuan Liuwei began to accept corporate commercial orders at a low or even free price to promote cooperation with enterprises. After building his reputation, he found that he could stably receive corporate training orders, and the company's business situation also improved.

Such partners are not easy to find in reality. For young entrepreneurs, it is even more difficult to form a team and make the organization operate efficiently.

2. From a “bureaucratic machine” to a “smart life form”

In the early 20th century, Max Weber, a German sociologist, proposed the concept of “Bureaucracy” during his research on the rise of modern states and capitalism. He believed that to achieve rational and efficient governance in modern society, it was necessary to rely on an administrative organizational form with strict rules, clear division of labor, and clear rights and responsibilities – “Bureaucracy.”

The advantage of Bureaucracy is reliability and order, but the cost is that the organization loses its flexibility, and people are trapped in the “iron cage” of processes and indicators. When AI models and AI Agents enter the organizational structure, the main body for executing rules is no longer a large number of human resources but an autonomous and collaborative intelligent system.

Ye Chunjun, an entrepreneur of a “one-person company,” is using AI to deconstruct the traditional functional division of company departments. With the help of AI, Ye Chunjun can handle the work of product, marketing, customer service, and strategic departments alone.

Every day, Ye Chunjun needs to process information from user feedback. This information is scattered on social media platforms such as X (formerly Twitter) and Discord, being complex and fragmented. Collecting and organizing it purely by human resources would require a lot of energy and time.

Ye Chunjun introduced an AI assistant, let AI do the initial digestion and precipitation, and maintain it as a complete strategic document to assist him in product thinking and final decision-making. Occasionally, he will have brainstorming sessions with AI.

In the early days of his entrepreneurship, he once co-wrote the first draft of a product requirement document with Google's Gemini. When he saw the output result from Gemini, he was shocked by the completeness of the document. Thinking about it now, Ye Chunjun can only praise it: “Now, an AI-written PRD (Product Requirement Document) for a product may be better than those written by 95% of product managers.”

Ye Chunjun also entrusted the core part of technology R & D to AI. “Now, 90% of the coding in our technology R & D is completed by AI.” This has also changed Ye Chunjun's role: from writing code himself to architecture design and code review. AI is responsible for execution, while he is responsible for controlling the direction and quality.

The promotion materials for Ye Chunjun's company on social media platforms, from creativity to implementation, are all realized using AI. Many of the videos produced in this way have reached millions or even tens of millions of views, and Ye Chunjun doesn't need to spend a lot of energy on implementation. He only needs to focus on decision-making and result acceptance.

So far, most of the monthly revenue of Ye Chunjun's “one-person company” has been used to purchase computing power and technology to continuously improve the product experience. He plans to drive the long-term growth of the company through the ‘flywheel effect’ of intelligent resources rather than the accumulation of human resources.

“This is the charm of a ‘one-person company’ in the AI era,” Ye Chunjun said, taking the team formation of traditional companies as an example. He believes that entrepreneurship is now entering a stage where “fewer people can achieve more.” “In the past, if my team had 50 people, I had to consider the livelihoods of these 50 people when making any decision. Now, I only need to consider how much cost is needed to purchase computing power. The company's operation and products rely on computing power, and I just need to pay the bill.”

Xu Ting, the founder of Xiaolu Guangnian in Shanghai, also has deep feelings about this. She reviewed her experience of involving AI in the company's product development.

Xu Ting launched an AI memoir product. By connecting to an open-source large model, this AI memoir integrates functions such as automatic interviews, voice recognition and content analysis, and automatic writing. It can automatically complete interviews with people and write memoirs based on the collected materials.

In Xu Ting's vision, this memoir can help many elderly people write a personal biography that previously required hiring a professional team and spending tens of thousands of yuan at a very low cost. It is a project for the popularization of technology.

Xiaolu Guangnian memoir interface, Image source: Provided by the interviewee

Xu Ting's core team currently has five people. In addition, AI is also an important member of this team.

AI participated in the programming work of the Xiaolu Guangnian team. To make AI generate code more efficiently, Xu Ting and her colleagues developed a method: first, let the colleague in charge of the backend write a standardized interface file, and then write a prompt for AI to write the code for the remaining interfaces based on the interface file. After writing the first standardized interface file, the colleague in charge of the backend only needs to be responsible for the acceptance of subsequent interface files.

“Suppose we have 20 interfaces to write interface files. It would take 10 to 20 minutes for a backend developer to write each file manually, and they would need to think anew for each file. Even at the fastest speed, it would take most of the day,” Xu Ting calculated the time cost. “If we hand it over to AI, we can get the result within five minutes. Combining subsequent testing and debugging, it ‘won't take more than half an hour in total,’ Xu Ting estimated.

The Xiaolu Guangnian team at work, Image source: Provided by the interviewee

There are also enterprises providing services for these “one-person companies.” On November 4th, Kingdee, an enterprise management cloud SaaS company, announced that “Kingdee Cloud” had been comprehensively upgraded to “Kingdee AI” and launched a new AI product, “Xiao K.” Currently, “Xiao K” has integrated nearly 20 intelligent agents, covering multiple fields such as marketing, supply chain, human resources, finance, and ESG.

Xu Shaochun, the Chairman and CEO of the Kingdee Group, sighed at the press conference: “AI has transformed organizations from ‘bureaucratic machines’ into ‘smart life forms.’”

It is worth noting that the attitude of investors towards “one-person companies” is also quietly changing.

3. The two-way choice between “one-person company” entrepreneurs and investors

Ye Chunjun has had two financing experiences.

During his first financing, there was still the figure of a co-founder behind him, which conformed to the traditional “team narrative” in Silicon Valley. However, during his second financing in early 2024, he was alone. The company had shrunk from an initial team of five or six people to just himself. What he was promoting was a team organizational concept that seemed almost paradoxical at that time: the core value of the company is inversely proportional to the number of the team.

The sharpest question from investors was: “With the team downsized to this extent, how can you guarantee the company's sustainability? What will happen to the company if something happens to you?”

In the traditional investment logic, a structure supported by only one person is undoubtedly synonymous with high risk. The successful model of traditional investment logic is based on decades of “teamwork,” while Ye Chunjun's “one-person company” seems like a beautiful illusion.

Facing the doubts, Ye Chunjun's answer at that time was still a bit inexperienced as he was still exploring. He tried to prove the feasibility of the model with data and growth trends. Although AI Agents were not as widely implemented as they are now, Ye Chunjun had already been proficient in using a number of AI tools. The development of AI technology in the past year has further strengthened his choice.

In his view