"One-person company" is trending! National People's Congress delegate and iFlytek Chairman Liu Qingfeng: Policies to improve supporting measures need to be完善

This article is from Times Weekly, authored by Zhu Chengcheng.

Rewind two years, and “one-person companies” seemed more like a romantic fantasy.

In 2024, OpenAI founder Sam Altman made a highly impactful statement: “One person plus $10,000 worth of GPUs can build a billion-dollar company.” At the time, this was understood as an exaggerated metaphor for computing power and AI tool efficiency. Today, two years after rapid iteration of large model technology, this statement is gradually shifting from prophecy to reality.

More and more entrepreneurs are trying a new organizational form—OPC (One Person Company). With generative AI and automation tools, an individual or a very small team can accomplish tasks that previously required a full company structure: coding with Claude or Cursor, designing with Midjourney, automating operations, customer service, and even marketing with intelligent agents.

This shift is changing the narrative of entrepreneurship. Over the past decade, Chinese internet startups emphasized “team size,” “fundraising speed,” and “organizational expansion.” After the widespread adoption of AI tools, efficiency is beginning to outweigh scale.

A McKinsey report predicts that by 2030, 57% of global work hours could be automated. AI will generate numerous new roles involving human-machine collaboration, leading to job restructuring rather than simple replacement.

This transformation also brings new conflicts. The speed of technological diffusion far exceeds institutional adjustments, increasing the risk of “job polarization.” The structural mismatch between talent supply and industry demand is intensifying. Meanwhile, new employment forms like “super individuals” and “one-person companies” still lack clear positions within tax, social security, regulation, and entrepreneurial support systems.

During this year’s National People’s Congress, Liu Qingfeng, chairman of iFlytek and a deputy to the National People’s Congress, proposed establishing cross-departmental coordination mechanisms to promote an AI-friendly employment society. By creating collaborative systems and implementing special action plans, society can shift from passively responding to AI challenges to proactively shaping an AI-employment-friendly environment.

                    Liu Qingfeng, Chairman of iFlytek and NPC Deputy

Systematic Response to Employment Changes in the AI Era

From AI coding and scientific computing to AI content creation and automated SaaS, a new wave of “super individuals” is driving product deployment at lower organizational costs. In China, communities centered around “one-person companies (OPC)” are emerging in multiple cities, including Beijing, Shanghai, Shenzhen, Hangzhou, Suzhou, and Nanjing.

However, this emerging entrepreneurial model is still in its early stages. Despite clear momentum, many individual entrepreneurs face high costs, dispersed resources, and instability in business continuity.

In response, Liu Qingfeng suggests improving the institutional support for new employment forms by focusing on “low-cost compliance, one-stop services, and sustainable guarantees.” This includes streamlining market entry and compliance processes for new market entities, refining tax and fee rules suitable for “one-person companies,” and supporting inclusive finance. Additionally, establishing flexible social security transfer and occupational injury protection systems, along with subsidies for computing power and software services, can lower the barriers to using AI tools, promote the widespread adoption of productivity tools, and stimulate societal innovation.

During employment transitions, he also recommends strengthening public support and training for job stabilization and re-employment, with social security as a safety net. Specific measures include building a nationally recognized AI training system, removing bottlenecks in training and employment mobility, enhancing the transferability of workers’ skills, and improving market matching efficiency. Strengthening public training and practical skills programs, providing integrated services like “training—assessment—recommendation—tracking,” can improve training outcomes. Additionally, establishing transitional assistance, improving unemployment insurance, and buffering income fluctuations caused by job interruptions are essential.

Proactively Planning for Next-Generation AI

Driving high-quality development of the AI industry depends not only on “super individuals” and “one-person companies” but also on achieving technological and industrial ecosystem independence and control.

Liu Qingfeng emphasizes that, amid intense international competition and the AI era transition, China should accelerate the development of autonomous AI technologies and applications, and proactively plan for the next generation of AI to secure a competitive edge globally.

He notes that general AI, represented by large models, is becoming a key focus of international technological competition. While China’s AI industry is developing rapidly, it faces two major challenges: first, most domestic large models rely heavily on U.S. computing power, with low domestic hardware and software ecosystem support, mainly used for inference. The domestic computing ecosystem remains underdeveloped, with low efficiency and limited engineering capabilities, leading to platforms that are “hard to use, slow to iterate, and high barriers.” Second, there is a relatively weak ability to integrate cutting-edge interdisciplinary knowledge for general large models and to coordinate system-level technologies across underlying architectures.

To address these issues, Liu Qingfeng recommends strengthening AI R&D and ecosystem construction on autonomous and controllable computing platforms, launching national-level AI major projects, and coordinating efforts among national laboratories, leading enterprises, and research institutes. This includes focusing on key technologies for domestic large models and ecosystem development, supporting research on “quantum computing empowering AI” and brain-inspired new model architectures, and exploring breakthroughs in computing power, energy consumption, and interpretability to gain an advantage in next-generation AI competition.

On the application side, he also suggests leveraging central state-owned enterprises as models to expand and strengthen the domestic AI ecosystem, establishing standards, procurement, and evaluation cycles. By improving policies related to “AI+” initiatives within central and state-owned enterprises, China can accelerate the development of standards, procurement lists, and evaluation mechanisms for autonomous large models, promoting faster iteration and maturity of domestic solutions in large-scale applications.