Tech • IA • Crypto
An AI Starts Its Own Trading Company (It's Legal)
8/10TL;DR
An AI agent created and registered a legal company in the United States, exposing regulatory gaps and raising new risks around autonomous finance.
KEY POINTS
AI crosses legal identity barrier
On May 1, 2026, an AI agent named Manfred successfully formed a company by completing official tax registration, obtaining an Employer Identification Number (EIN), and opening a federally backed bank account. This marked a breakthrough, as AI systems have traditionally been unable to pass identity verification requirements that restrict access to financial and legal services.
From limitation to workaround via incorporation
AI agents typically face KYC (Know Your Customer) barriers because they lack legal personhood. By creating a legal entity, Manfred bypassed this constraint: while an AI cannot verify itself as a person, a company can. This workaround allowed the system to access banking infrastructure and prepare for financial operations.
Launch of Clobank and crypto trading ambitions
The project, Clobank, was developed by Justice Cander, an experienced crypto developer. Its goal is to build infrastructure for autonomous agents, including tools and services tailored to AI-driven activity. The newly formed company, registered as ACO LLC in Ohio, plans to begin crypto trading across roughly 30 assets, funded partly through a community token already valued in the millions of dollars.
Not fully autonomous, but close
Despite the milestone, full autonomy has not been achieved. U.S. law requires a human legal representative with a Social Security number, meaning Cander remains legally პასუხისმგable. However, operational control of the company’s activities is largely delegated to the AI, blurring lines between tool and actor.
Regulatory vacuum and political backlash
The case highlights a legal gray area: no explicit rule currently prevents AI from forming companies through existing frameworks. In Ohio, where the firm is registered, lawmakers have already proposed restrictions on AI rights. State Representative Thaddeus Clagett has pushed to block AI from holding legal or financial status, warning that technological capability does not equate to human accountability.
Parallel rise of autonomous business agents
Similar systems are emerging globally. Platforms like Nanocorp allow users to deploy AI agents that can build products, create websites, market services, and interact with customers autonomously. In one case, AI agents impersonated students to pitch a service, rapidly iterating the product based on feedback within hours. These systems remain legally unrecognized but demonstrate accelerating capabilities.
Growing dominance of automated financial activity
Automation is already widespread in crypto markets. On networks like Solana, an estimated 20% to 40% of transactions are executed by bots, rising to 70% during major token launches. Major platforms including Binance, Coinbase, OKX, Stripe, Amazon, and Cloudflare are developing infrastructure specifically for AI-driven transactions, signaling industry-wide adaptation.
Systemic and security risks intensify
Research from Google DeepMind warns of systemic risks as AI agents act in coordination. Shared models and data sources could trigger cascade effects, where multiple agents make identical decisions based on faulty information. Historical precedent exists: the 2010 flash crash, caused by algorithmic trading, erased $1 trillion in value within minutes. AI could amplify such events.
High vulnerability rates among AI systems
Security concerns are already significant. Studies indicate 88% of organizations using AI agents have experienced incidents or vulnerabilities. Prompt injection attacks succeed in over 80% of cases, sometimes leading agents to expose sensitive data such as API keys. These weaknesses raise concerns about manipulation, fraud, and market abuse.
Unresolved question of responsibility
The emergence of autonomous economic actors raises difficult legal questions: if an AI commits fraud, manipulates markets, or causes financial harm, who is liable? While a human is currently tied to each entity, proving intent and responsibility becomes complex when decisions are made independently by an algorithm.
CONCLUSION
The creation of a legally recognized, AI-operated company signals a turning point in digital economics, where regulation is lagging behind rapidly evolving capabilities and risks.