Germany protects the factory floor, while China promotes one-person businesses with AI

AI makes micro-enterprises more powerful: Fewer staff, more output? Why China is focusing on AI-powered one-person businesses

Artificial intelligence is not only changing how companies operate. It could also change how large a company needs to be to offer sophisticated services. China is specifically promoting so-called AI one-person businesses: small organizations in which a single entrepreneurial leader uses AI systems, digital platforms, and shared infrastructure to accomplish tasks that previously often required entire teams. Cities like Shenzhen and Hangzhou are connecting workspaces, technical resources, financing, and contacts with industrial clients to facilitate this. This is not yet proof of a new economic superiority: political goals are not the same as sustainably successful companies, and many questions about demand, platform dependency, and working conditions remain unanswered. However, China's approach raises an important question for Germany: how can industrial strength be preserved while simultaneously creating an environment in which small, AI-powered companies can collaborate productively with large corporations?

China's one-person businesses: How AI is changing the industrial competitive order

The economic policy significance of artificial intelligence is not solely determined by which country develops the most powerful models or builds the largest data centers. Equally important is who uses this technology to create new companies, marketable services, and more productive industrial processes. It is precisely at this intersection that China is pursuing an approach that deserves considerably more attention in Germany: the systematic promotion of companies whose performance is not primarily based on the number of their employees, but rather on the combination of a responsible entrepreneur with AI systems, digital platforms, and shared infrastructure.

The term "One Person Company," or OPC for short, is emerging as an economic policy category that goes beyond the traditional image of the solo self-employed individual. It doesn't simply refer to a freelancer with a better word processing program. At its core is a business organization where one person makes key decisions, while AI applications and, increasingly, software-supported agents handle numerous operational tasks. Product development, programming, customer communication, marketing, and administrative preparation can thus be managed with a significantly smaller team.

This does not make large companies obsolete. However, it does change the conditions under which small businesses can emerge and collaborate with them. China's strategic approach is not to leave this development solely to individual ingenuity. Municipal infrastructure, technological resources, financing, and industrial demand are to be linked in such a way that a new entrepreneurial class can grow from individual startups.

For Germany, this raises a different question than the question of the next protective instrument for existing businesses: How can industrial substance be preserved while simultaneously enabling an economic structure in which significantly smaller organizations can deliver sophisticated services?

A new form of organization, not a miraculous abolition of work

The term "one-person business" is easily misleading. It can refer to a company with only one owner, a business without employees, or a particularly lean organization that works intensively with AI. These variations are not economically identical.

A company with a single shareholder can employ numerous people. A self-employed individual, on the other hand, can operate largely without AI. An AI-driven OPC, however, can have additional employees and still be fundamentally led by a single entrepreneur.

This distinction has been explicitly adopted in China. A terminology standard developed in Zhejiang, which came into effect on August 1, 2026, describes AI OPCs as companies under the significant control of a natural person, generally with no more than ten employees, and whose business focus is on the research, development, application, or provision of AI technologies. This is an industry or association standard, not a new, nationwide uniform legal form.

This acknowledges an important economic reality: the defining characteristic is not the literal absence of employees, but rather the unusually high leverage of a small organization. The founder is responsible for the business model, makes priority decisions, and monitors the results. Technical systems multiply his execution capacity.

The number of agents employed is also not a reliable performance indicator. Twelve agents can support a well-organized process or create twelve new potential sources of error. What matters is whether a company reliably sells services, covers its costs, and fulfills its obligations.

The economically relevant entity is therefore not simply human plus AI. It consists of expertise, entrepreneurial responsibility, controlled automation, access to customers, and a viable technical infrastructure. If any one of these components is missing, the model remains fragile despite impressive demonstrations.

From local startup experiment to national industrial policy

China's OPC development is not just a trend within the technology scene. It is increasingly being embedded in economic policy programs.

In June 2026, seven government agencies published an action plan for cooperation between large, medium, and small enterprises in the platform economy for the period 2026 to 2028. The participating institutions include the Ministry of Industry and Information Technology, the National Development and Reform Commission, the Ministry of Science, the Ministry of Commerce, and the agencies responsible for internet governance, market regulation, and data. The plan explicitly mentions the accelerated development of AI-powered one-person businesses.

What is remarkable is not so much the inclusion of a new buzzword, but the context in which it occurs. OPCs are not treated as an isolated exception, but as part of a collaborative system comprising platform companies, specialized service providers, and industrial users.

The national framework also includes specific development targets. By 2028, these include the publication of three batches of lists for opening platform resources, the selection of at least 100 pilot projects for relevant application scenarios, and the development of at least ten service platforms and 60 practically implementable smart service scenarios. These targets relate to the overarching platform economy program and should not be interpreted as standalone OPC funding objectives. However, they do indicate the infrastructure into which the new companies are to be integrated.

This is the real strategic difference compared to purely individual start-up support. A grant improves the financial situation of a single company. Accessible technical and commercial infrastructure, on the other hand, can change the conditions for starting many companies simultaneously.

However, it would be an overstatement to conclude from this that a fully realized national production architecture has already been established. Significant differences exist between political objectives, municipal implementation, and actual market success. Likewise, it cannot be stated categorically that all large companies are already legally obligated to award contracts to OPCs. What is demonstrable, however, is the political will to open up resources and organize cooperation more systematically.

Shenzhen is not just building offices, but economic hubs

The approach is particularly evident in Shenzhen. The city combines its existing industrial and technological infrastructure with targeted support for AI-powered micro-enterprises.

The action plan for 2026 and 2027 envisions more than ten leading OPC communities, each with at least 10,000 square meters of space, by the end of 2027. The total area of ​​OPC communities citywide is projected to increase to more than 500,000 square meters. Furthermore, the plan aims to support more than 1,000 high-growth AI startups and bring together more than 10,000 people involved in AI innovation and entrepreneurship. These figures represent policy objectives, not achievements to date.

At first glance, half a million square meters seems contradictory for companies without traditional office equipment. However, this can be explained economically. Such communities are not intended solely to rent out workspaces. Their value arises from shared services, technical support, contacts with investors, access to industrial applications, and proximity to other specialized companies.

Especially in Shenzhen, the connection to the physical economy is relevant. An AI application for an electronic device may require components, prototype production, testing facilities, and distribution partners. The software can be developed by a small organization, while the physical implementation relies on a dense network of existing companies.

OPC is therefore not outside of industry. It becomes a new, particularly streamlined interface within the industrial division of labor.

However, physical space alone does not create economic value. If communities are primarily evaluated based on their size, occupancy rate, or number of registered companies, a familiar problem of government location policy arises: visible infrastructure is mistaken for actual innovation. What matters is not how many desks are available, but how many paying customers, robust products, and sustainable business relationships are generated.

Hangzhou pools capital, computing power and market access

Hangzhou is pursuing a similar approach, but focuses more clearly on combining financing with digital operational infrastructure. The Shangcheng district, in particular, is positioning itself as a location for AI-powered micro-enterprises.

The measures implemented there include affordable workspaces, administrative support, resource allocation, and a shared AI service platform. This platform aims to consolidate models, computing power, and development tools within a membership system. In addition, an OPC fund with a target size of one billion yuan has been established. Furthermore, at least 100 million yuan has been announced for annual support. It's important to note that the fund volume, annual funding, and actual invested capital are distinct metrics and should not be added together or equated.

The economic logic of such a pooling of resources is more compelling than simply providing discounted office space. Otherwise, a single founder would have to procure numerous services separately: access to software, cloud services, development environments, data storage, payment processing, and potentially support with legal or accounting issues.

Every additional interface increases search effort, integration costs, and the potential for errors. A shared platform can reduce this burden. It transforms part of starting a business from an individual integration project into standardized access to existing services.

But therein lies a risk. The more comprehensively a founder organizes their technical infrastructure, customer acquisition, and billing through a single system, the greater their dependence on its operator becomes.

The platform can reduce start-up costs while simultaneously capturing a significant portion of subsequent value creation. Low entry prices are therefore not automatically proof of permanently favorable business conditions. Data portability, contract stability, alternative providers, and the ability to maintain existing customer relationships independently are also crucial.

Jiangsu demonstrates why easy access is economically important

In Jiangsu, and particularly in the industrialized region of Suzhou, OPC (Open Platform Commercial Enterprise) promotion is linked to the goal of more closely integrating digital startups into existing economic structures. For Suzhou, plans include the establishment of 30 OPC communities and the development of 1,000 related companies by 2028.

The underlying logic ranges from collaborative work opportunities and computing and modeling resources to the provision of concrete use cases. Low-threshold administrative services, which closely integrate access to a workplace, business registration, and support services, are intended to limit organizational effort.

However, such instruments should not be prematurely interpreted as universally identical or standardized rules. Local programs differ according to location, operator, and target group. A political announcement does not replace an examination of the actual conditions of access.

Nevertheless, the approach is economically significant. For a large company, additional administrative steps are often cumbersome but manageable. For a solo founder, every hour of administrative work directly competes with product development, customer meetings, and service delivery.

A week lost to unclear responsibilities, duplicate data entry, or technical registrations cannot be shared among multiple people in a one-person organization. This is precisely why relatively minor simplifications can yield significant economic benefits.

Even more important is the connection to industrial customers. A founder with access to a real problem faced by a manufacturer or logistics company often has a better starting position than a technically talented developer without verifiable demand.

The new operational size results from decreasing coordination costs

Why could AI increase the number of economically viable micro-enterprises? The answer lies less in spectacular machine intelligence than in the changing cost structures.

Companies consolidate activities because their internal organization offers advantages over constant coordination with external providers. Employees share knowledge, develop routines, and coordinate interconnected work processes. These advantages justify salaries, hierarchies, and fixed organizational costs.

When AI takes over some of the standardized knowledge work, this trade-off can shift. For certain products, it will become less necessary to establish dedicated departments early on. A small organization can initially cover additional functions technically and bring in external specialists as needed.

This includes, for example, the creation of initial software versions, the preparation of product information, the handling of recurring customer inquiries, or the preparation of standardized analyses. Not every one of these tasks will be fully automated. Even a significant reduction in the amount of human time required can change the minimum economic viability of a bid.

However, not all costs decrease simultaneously. New expenses arise for technical integration, monitoring, data security, and quality control. In addition, there are ongoing fees and the effort required to identify and correct erroneous results.

The relevant calculation, therefore, is not how much cheaper an AI expenditure is compared to an hour of human labor. What matters are the total costs of a reliably delivered service. A cheap automated process can become expensive if it results in extensive rework or damage for the customer.

A numerical example explains the potential and its limitations

A simplified scenario illustrates the business impact. The following figures are not measured averages, but a deliberately transparent calculation example.

Let's assume a small, specialized service unit incurs annual operating costs of €240,000 before considering project-related external services. With a contribution margin of 60 percent, it needs €400,000 in revenue to cover these costs.

A leaner, highly automated organization could operate with annual costs of €120,000 under a specific business model. This would have to include a reasonable entrepreneurial salary, technical services, insurance, external support, and administrative expenses. With the same contribution margin, the calculated revenue threshold would be €200,000.

That is a significant difference. A market segment that would be too small for a conventional structure could become economically attractive for such an organization.

However, a lower cost base does not guarantee higher profits. If numerous competitors use the same tools, prices can fall. If the contribution margin drops to 40 percent in the second scenario, the required revenue already rises to €300,000.

This example illustrates the key connection: AI can reduce the cost of market entry and service provision. Whether the entrepreneur retains this advantage depends on competition, customer benefits, and differentiation.

Those who merely resell generally available AI services often have little pricing power. In contrast, those who understand and reliably improve a complex industrial process can develop a substantial economic advantage, even with a small organization.

The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business - Image: Xpert.Digital

Xpert.Digital is a data-driven B2B industry hub led by Konrad Wolfenstein . The company acts as an external, quasi-in-house solution for industrial partners, closing operational gaps in marketing, content, and sales – without requiring additional resources on the client side.

More information here:

• The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business

The bottleneck is shifting from production to demand

Much of the public AI debate focuses on how quickly content, programs, or analyses can be created. However, for businesses, a different question is often crucial: Who pays for it?

When the production of digital services becomes easier, the supply initially increases. More providers can develop websites, produce advertising copy, process data, or publish simple applications. However, customer attention and willingness to pay do not automatically increase at the same rate.

This can shift the bottleneck. Success is not determined by the ability to manufacture a product, but by the ability to identify a relevant problem, build trust, and reach customers economically.

This explains why opening up to industrial use cases is so important. A real customer delivers more than just revenue. They help to refine requirements, reveal technical limitations, and verify willingness to pay.

A voucher for computing power facilitates development. A paying pilot customer tests the business model. For the sustainable emergence of an OPC economy, the second function is usually more valuable.

This also reveals a limit to government funding. It can provide infrastructure and create opportunities for contact. However, it cannot permanently replace the fact that a customer demands a service out of their own economic interest.

Industry and logistics offer more attractive niches than interchangeable mass-produced goods

For Germany, the OPC approach is particularly interesting where deep expertise meets recurring information problems.

In industry, for example, technical documentation, spare parts information, maintenance records, or sales data may be insufficiently accessible. A specialized provider could help make existing knowledge discoverable, structure information, or improve internal search processes.

In logistics, potential application areas include document verification, status communication, the preparation of deviation analyses, and the processing of customer requirements. Such possibilities do not constitute a blanket guarantee of technical feasibility. Their usefulness depends on data quality, system access, and specific process requirements.

This is precisely where a small, technically strong organization can have an advantage. It doesn't need to serve a global mass market. A narrowly defined problem with high economic significance can be sufficient as a foundation.

Therefore, the best opportunities do not necessarily lie in another general AI application. A solution that particularly understands a specific document type, a particular maintenance process, or a clearly defined customer group may be more attractive.

Liability remains crucial. A flawed summary for internal research purposes must be assessed differently than an incorrect safety-relevant instruction. The greater the potential consequences, the more stringent the expert review, traceability, and human approval must be.

A viable one-person organization therefore does not automate indiscriminately. It distinguishes between low-risk routines and decisions that require special responsibility.

The key figures for the start-up wave can easily be misleading

Caution is advised when evaluating Chinese OPC figures. Reports cite more than 16 million single-member limited liability companies by mid-2025. This figure is sometimes used in connection with the AI ​​trend, but it does not represent a proven total of AI-powered businesses. A company with one shareholder is not automatically an AI-powered OPC.

Anyone who conflates these categories overestimates the scope already achieved by the new organizational form. Likewise, received applications, registered companies, and economically active businesses should not be equated.

A location can attract many potential customers without generating corresponding revenue. A company can legally exist but conduct little business. A community can be fully utilized while its members primarily rely on subsidies.

More meaningful data would be on paying customers, recurring sales, survival time, productivity and economic development after the end of the funding.

The often-cited notion of a million AI-powered one-person companies should also be treated as a scaling vision, not as an already established economic reality. A large number of newly founded companies would, in itself, be neither proof of productivity nor evidence of successful industrial policy.

The more robust findings are narrower, but still noteworthy: China is developing a growing number of specialized support programs and increasingly linking them to national goals. How much additional value is generated as a result remains an empirical question.

Many failed companies still lack a successful strategy

The fact that a significant number of new businesses may fail is not an argument against supporting startups. Economic renewal also arises from ventures whose outcome is uncertain.

However, this should not lead to the convenient assertion that every failure is automatically productive. What matters is the cost incurred and what knowledge, customer relationships, or reusable technologies remain.

An OPC strategy can be economically attractive if the costs of individual experiments remain low and successful approaches can scale rapidly. It can become inefficient if many nearly identical offerings emerge, whose only common source of demand is the funding program itself.

Financing that combines private debt with poorly vetted business models would be particularly problematic. A politically desired startup boom could then mask individual risks.

Even municipal competition deserves a sober assessment. If several cities compete for the same startups with similar rent reductions, subsidies, and infrastructure offerings, some of the support may simply trigger relocations.

The economic benefit only arises when additional services, new market participants, or measurable productivity gains are added. A company that relocates because of a subsidy is not yet a newly created company.

Therefore, a serious evaluation requires comparison groups, temporal tracking, and a clear separation between windfall effects and actual triggered activity.

The independent company can become a dependent platform operator

The OPC narrative emphasizes individual freedom. A person can build a business with limited resources without having to finance a large organization early on.

At the same time, this very model can create new dependencies. Those who organize computing power, models, payment processing, customer contacts, and sales via external platforms may only own a limited portion of their own economic infrastructure.

Control over value creation does not then shift exclusively from the large corporation to the independent founder. It can also move to the providers of the technical infrastructure.

A large number of small businesses can thus lead to increasing concentration on a few platforms. The number of market participants grows, while key access conditions are determined by a small number of players.

For startups, this means that technical interchangeability becomes an economic safeguard. Established customer relationships, exportable data, documented processes, and alternative suppliers are not merely minor technical details. They limit dependence on price fluctuations or changes in contract terms.

This presents a challenging task for economic policy. It should facilitate start-ups without simultaneously subsidizing closed systems of dependency. A publicly supported platform is not automatically in the public interest simply because many small businesses use it.

Less staffing needs do not automatically mean less efficient work

The labor market consequences are contradictory. On the one hand, AI-supported startups can create new employment opportunities. People with expertise, customer contacts, or entrepreneurial experience gain the opportunity to offer services that previously would have required a larger organization.

On the other hand, these very companies can take over tasks that were previously performed by employed teams. New forms of self-employment and the displacement of certain activities can therefore occur simultaneously.

The crucial question is not simply whether new companies are founded. What matters is the income they generate, the stability of their demand, and the risks they pass on to individuals.

A high number of legally independent providers can also reflect weak demand for regular employment. Without relevant data, it is impossible to reliably distinguish whether a wave of new business start-ups primarily reflects new opportunities or a lack of alternatives.

Added to this is the personal burden. A founder can automate operational tasks, but often remains responsible for sales, quality control, liquidity, and customer relationships simultaneously. Technical support does not eliminate this responsibility.

The image of the effortlessly running one-person company is therefore misleading. Successful automation can reduce workload. However, it can also increase the number of simultaneously monitored processes and thus create new forms of overload.

The contrast between a factory floor and an AI company is too simplistic

The pointed juxtaposition between German preservation of the status quo and Chinese future orientation touches a raw nerve, but remains incomplete.

An industrial site requires energy supply, production facilities, technical training, logistics, and long-term investments. An AI-powered OPC (Operational Power Processing) system does not replace a steel mill or a semiconductor factory. Digital business models also rely on physical infrastructure.

Conversely, a modern factory hall needs more than just machines. It requires software, data integration, technical documentation, maintenance expertise, and the ability to quickly integrate new technologies into existing processes.

The economically productive relationship therefore lies in the complementarity. Small AI companies can improve the industrial base, while industrial customers offer them demanding use cases and solvent demand.

It would be wrong for Germany to pit the preservation of industrial capacity against new organizational forms. It would be equally wrong to equate the preservation of existing structures with a comprehensive future strategy.

The assertion that there is no term in this country for the combination of a human and AI is also too sweeping. Solo self-employment and one-person businesses are not unknown phenomena. The relevant question is rather whether the specific combination of small organization, high automation, and industrial networking is adequately taken into account.

A European response needs less symbolism and more functioning interfaces

Europe would not need to completely copy the Chinese model. A convincing response could build on existing industrial strengths while simultaneously taking its own requirements for data protection, competition, and entrepreneurial independence seriously.

The first lever would be a consistently simpler company organization. Information should not have to be submitted multiple times. Responsibilities must be clear, standard processes should function digitally, and ongoing obligations should be proportionate to the company's size.

The second lever would be reliable access to technical resources. Small businesses don't necessarily need free computing power, but rather predictable prices, understandable contract terms, and an infrastructure that can be used without disproportionate integration effort.

The third lever would be collaboration with industrial users. Large companies could provide access to clearly defined problems, secure testing environments, and suitable contacts. It would be important that small providers do not have to meet the same organizational requirements as a large international supplier, unless the nature and risk of the service justify it.

This does not mean lowering necessary quality standards. It means aligning requirements with actual risks. A limited-scale analytical prototype requires different treatment than a safety-critical production control system.

The fourth lever would be a better combination of expertise and technical support. Not every promising founder is an AI developer. An experienced engineer, logistics expert, or industry consultant can understand a relevant problem more accurately than a technically skilled generalist.

Funding should reward results, not the continued existence of small businesses

A good OPC policy should avoid making the one-person structure itself the purpose of funding.

A successful company is entitled to hire employees, bring in partners, or expand its organization. However, if creating new jobs leads to the loss of all benefits, perverse incentives are created. The company might then avoid growth altogether or artificially outsource activities to meet a specific definition of incentives.

Conversely, a company should not be favored simply because it automates a particularly large number of tasks. What matters is whether it operates productively, solves customer problems, and manages its risks.

Funding would therefore be better aligned with verifiable development steps: a functional solution, initial independent customers, proven process improvements, and a decreasing dependence on public support.

Several related metrics are suitable for evaluation. These include the proportion of companies with recurring revenues, their economic stability after the funding period ends, the productivity of their customers, and the quality of the services provided.

The public costs per sustainably viable company are also relevant. However, they should not be considered in isolation. A failed company can leave behind valuable knowledge; a company that formally survives can remain almost entirely dependent on subsidies.

The challenging task is to enable genuine learning processes while simultaneously preventing permanent financial support.

The real competition is decided by the distribution of productivity gains

Even if AI significantly increases the performance of small businesses, it does not yet decide who will benefit economically.

Part of the benefit can remain with the founder and manifest itself in higher incomes. Another part can be passed on to lower customer prices through competition. Further shares can be skimmed off by platform providers, investors, or particularly powerful clients.

This makes the OPC economy also a question of bargaining power. A specialist with their own customer base and irreplaceable industry knowledge holds a different position than an interchangeable provider who is entirely dependent on a digital marketplace.

For Germany and Europe, therefore, the number of new companies is not the only decisive factor. What matters is whether these companies develop independent economic substance.

These include proprietary products, strong customer relationships, protected expertise, and the ability to switch technical providers. Without these elements, a seemingly independent company can become an outsourced execution operation that bears the majority of the risk but receives only a small share of the added value.

The strategic perspective should therefore not end with the company's founding. It must extend to the question of what market structures will emerge after a few years and whether new companies will actually remain competitive within them.

China's advantage lies in the organized experiment, not in the proven end result

The Chinese OPC strategy is neither merely a buzzword nor proof of a new economic superiority. Its remarkable core lies in the early political recognition of an emerging organizational form and the systematic development of its prerequisites.

Shenzhen is planning spatial and industrial networking. Hangzhou is pooling capital and technical services. Other locations are combining startup support with shared infrastructure and concrete applications. The national framework places these activities within a broader collaboration between platforms, small businesses, and industrial players.

Whether this will result in a large number of sustainably productive companies remains to be seen. Equally uncertain are the long-term consequences for competition, income, and platform dependency. A dispassionate analysis must explicitly acknowledge this uncertainty.

The central lesson for Germany is nevertheless clear: industrial policy should not solely focus on which existing companies need protection. It should also examine what technical, administrative, and commercial prerequisites new companies require to become economically effective.

The factory floor and the AI-powered micro-enterprise are not adversaries. Their connection could become a significant source of industrial innovation. Competitiveness doesn't automatically depend on the country with the most agents or the most generous start-up grants. What matters is where new value is reliably created from expertise, automation, and real demand.

From visibility to trust: Your scalable path with Xpert.Digital - Image: Xpert.Digital

In industrial B2B, sustainable business relationships rarely emerge overnight. They develop step by step – through visibility, professional relevance, recurring touchpoints, and growing trust. Xpert.Digital's 4-stage model addresses precisely this: It offers a structured path that begins with a manageable entry point and can evolve into deeper collaboration in business development if needed.

Instead of relying on loud marketing promises, this model puts the relationship at the forefront. Companies start with clearly defined, easily calculable measures and then decide, based on their own experience, how far they want to expand the collaboration. A key factor for this undisturbed trust-building process: The platform completely avoids annoying advertising ads, so the editorial focus remains solely on the companies' expertise.

More information here:

• From visibility to trust: Your scalable path with Xpert.Digital

☑️ Our business language is English or German

☑️ NEW: Correspondence in your native language!

Konrad Wolfenstein

I and my team are happy to be available to you as your personal advisor.

You can contact me by filling out the contact form here [email protected]:or simply call me at +49 7348 4088 965. My email address is

I'm looking forward to our joint project.

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs