Delaware’s corporate policy committee has proposed legislation to create a testing ground for companies operating under Artificial Intelligence control, marking a shift in how the state approaches AI regulation at a moment when billions in technology investment are reshaping business models nationwide.
The proposed measure, drafted by corporate attorney John Mark Zeberkiewicz of Richards Layton and Finger, would establish a new legal entity called an Artificial Intelligence Company, or AIC. Under the framework, AI agents would autonomously handle business tasks traditionally performed by humans, from coding services to contract negotiation and legal defense. The legislation is expected to be introduced in Delaware’s General Assembly next year.
The proposal emerges as Delaware seeks to retain its dominance as the corporate incorporation capital of the United States, even as competitive pressures mount. Tech companies have invested billions in recent years to build the data center infrastructure powering AI systems, creating an economic imperative for states to establish clear legal pathways for the technology. Infrastructure investment is reshaping how companies monetize AI capabilities, making regulatory clarity a competitive advantage.
Protecting Owners From AI-Generated Liability
A core feature of the proposed legislation addresses legal responsibility. The framework seeks to shield owners of AICs from liability stemming from autonomous actions taken by their AI systems. This protection is critical because the scope of AI autonomy under the proposal is broad. Zeberkiewicz indicated that “anything that a company can do, the AIC can do as a company,” suggesting AI agents could theoretically operate across nearly any business sector.
That breadth introduces questions the drafters themselves acknowledge remain unsettled. When asked whether an AI company could operate a small nuclear power plant, Zeberkiewicz noted the theoretical possibility but expressed skepticism that companies would pursue such high-risk applications in practice. The comment highlights a gap between what the regulation permits and what market actors are likely to attempt-a distinction that may not hold indefinitely.
Supporters argue Delaware cannot afford to wait for legal uncertainty to resolve before companies begin deploying autonomous AI systems. Businesses are already moving forward with AI agents performing autonomous work in real environments. A regulatory sandbox allows Delaware to gather empirical data on how these systems perform, fail, and interact with existing corporate and commercial law before establishing permanent rules.
Timing Pressure From Market Momentum
The proposal’s timing reflects urgency within both the technology and corporate legal sectors. Delaware’s committee drafted the AIC framework after observing accelerating AI deployment across industries, signaling that regulatory lag poses a competitive risk.
Market momentum around AI has been sustained despite volatile sentiment in broader equity markets. Technology stocks have driven recent rallies on major exchanges, with heavyweight names like Alphabet and newly listed AI-adjacent companies capturing investor attention. This investor appetite for AI-related exposure creates financial incentive for states to position themselves as favorable jurisdictions for AI business formation and operation.
Delaware’s corporate franchise generates substantial state revenue through incorporation fees and related services. Losing AI-related business formation to other states or jurisdictions that clarify legal status earlier would represent a genuine economic loss. The proposed AIC framework is partly defensive positioning against that scenario.
Unresolved Questions About Operational Safety
The proposal does not settle fundamental questions about oversight and safety limits. How will Delaware regulators monitor AIC operations in real time? What triggers intervention or suspension of an AI system’s autonomy? How does the framework handle situations where an AI agent makes contractual commitments that later prove harmful to third parties?
The regulatory sandbox approach assumes that controlled testing will generate answers. But the breadth of the proposed AIC definition-allowing AI autonomy across most business activities-means the range of possible failure modes is equally broad. A sandbox only contains risk if its walls are clearly defined and monitored. The proposal sketches the concept but leaves enforcement mechanisms and operational boundaries for future rule-making.
Environmental implications also remain unexplored. AI system expansion carries significant energy and resource costs that scale with deployment. A state-level sandbox approving widespread AI autonomy without energy impact analysis or sustainability requirements may accelerate costs that extend beyond Delaware’s borders.
Next Steps And Competitive Landscape
If introduced and passed, the legislation would position Delaware as the first state offering a formal regulatory pathway for AI-operated companies. Other jurisdictions monitoring corporate law trends may follow, creating a patchwork of state-level AI regulations. That fragmentation could either accelerate a national standard or entrench Delaware’s advantage depending on how quickly and successfully the sandbox operates.
The committee’s drafting work signals that Delaware leadership views AI governance not as a distant future concern but as an immediate business priority. The proposed AIC framework is incremental-it does not authorize unrestrained AI autonomy, but rather creates a testing environment with eventual rulemaking. Yet even that limited step acknowledges that AI systems capable of autonomous business operation are no longer hypothetical. They are arriving faster than permanent regulation can be designed, forcing policymakers to choose between sandboxing risk or ceding market position to competitors.
The outcome will shape how other states and jurisdictions approach AI business autonomy in the coming years.
