Earlier this year, an entrepreneur in China registered a digital content company in 12 days on her own. Her staff is a fleet of artificial intelligence agents. Her corporate wrapper is China’s newly liberalized one-person limited liability company, or OPC. And her ambition is global.
She isn’t an outlier. By the end of last year, nearly 7.32 million OPCs were newly registered in China — up 42.3% year over year — fueled by the country’s revised Company Law and the explosion of agentic AI tools available to solo founders. I have described this as the “Raphael model”: one master creator directing specialized AI agents across multiple concurrent ventures.
China’s revised Company Law — which took effect in July 2024 — arguably has built the world’s first corporate-law framework compatible with such agentic ventures. And Alibaba International’s Accio agents demonstrate the model at multinational scale. However, while China solved the founder-side problem, the unresolved question for every jurisdiction is the legal status of the AI agents in OPCs.
Meanwhile, this trend is manifesting globally. Powered by advanced agentic platforms such as Anthropic’s Claude, US solo founders are suddenly achieving the market footprint of mid-sized corporations from a single-shareholder laptop. Like China, the US market is witnessing the rapid transformation of the single-member LLC — the country’s equivalent of the one-person company. The US commercial agency law is due for an upgrade for autonomous agents.
China’s OPC Framework
China’s new Company Law introduced a vital structural pivot. It explicitly canceled earlier restrictions that limited a natural person to owning a single OPC. A solo founder can now incorporate multiple OPCs simultaneously — one per AI product line — with full legal separation between them.
Combined with streamlined, online registration and governance flexibility for sole directors, the framework provides the optimal corporate wrapper for a multi-threaded entrepreneur.
To prevent abuse, the law upgraded creditor protections, extending joint and several liability across entities if a shareholder uses multiple OPCs to evade debts.
Multinational Scaling
The legal issue emerges when these hyper-empowered solo entities interface with the global economy. The clearest proof is Alibaba’s Accio Work platform, which functions as a full-stack digital workforce powering more than 230,000 online stores globally.
According to Alibaba.com president Kuo Zhang, these autonomous agents partner with local AI models to independently navigate country-specific legal and financial requirements — processing customs paperwork, VAT filings, and compliance documentation across more than 100 markets.
This crystallizes the legal gap. An OPC founder in an emerging market directing Accio agents is operating a cross-border agentic enterprise from a single-shareholder wrapper. Every jurisdiction their agents touch has commercial laws built upon traditional principles of human representation and signatories.
While China has legalized the founder’s ability to scale, no jurisdiction has defined the agent’s legal status when executing transactions across borders.
Redefining Agency Law
Other jurisdictions are responding — but at the governance layer, not the corporate structure layer. Frameworks such as Singapore’s Model AI Governance Framework and the EU AI Act regulate software behavior and risk tiers. They don’t touch corporate entity forms, the attribution of intent, or the liability rules when an autonomous agent acts beyond its authorized scope. The intersection of agentic AI and single-member LLCs presents complex legal challenges that current frameworks are struggling to address.
As I noted in my Bloomberg Law analysis of the OpenClaw framework, we are in an “iPhone 1.0 moment”: transformative capabilities outpacing legal infrastructure. The “know your agent” dilemma is the sharpest symptom.
When an AI agent creates a compliance breach, who is responsible? Unlike a human employee, an AI agent lacks legal personality and can’t share liability. US and common-law courts have consistently affirmed that corporate duties of competence and supervision can’t be outsourced to a machine. If a solo founder fails to maintain strict corporate formalities and court-ready telemetry, plaintiffs will seek to pierce the corporate veil, holding the human founder personally liable.
This reality presents a profound structural advantage that common-law jurisdictions — including the US, UK, and Singapore — have yet to exploit. While civil-law frameworks require statutory revisions every time technological autonomy outpaces codified doctrine, common-law courts can adapt commercial agency doctrine iteratively, case by case, to calibrate liability and define AI-driven attribution of intent in real time.
Next Legal Frontier
China’s OPC framework answers half of the agentic enterprise question: What corporate structure governs the founder? The other half — what legal status governs the agent — remains unanswered. Who legally acts when an Accio Work agent negotiates a supplier contract, an OpenClaw agent executes a financial transaction, or a Claude-integrated workflow files a regulatory document?
China has engineered the corporate structure for the modern workshop. The same agentic shift has transformed the US single-member LLC into a formidable economic engine that can replace the labor of whole teams.
Washington, London, and Brussels possess the legal tradition to build the commercial agency law that governs the new digital workforces. The question is whether Western policymakers will move quickly before AI agents drive their founders to look elsewhere for it.
This article does not necessarily reflect the opinion of Bloomberg Industry Group Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.
Source:AI-Enabled One-Person Companies Present a Corporate Law Quandary