The Commodity Futures Trading Commission’s chair has outlined two emerging challenges for finance: hedging the cost of computing power and deciding who is responsible when AI agents trade. Michael Selig’s remarks point to questions regulators will need to tackle, not new rules already in force.
Watch Desk analysis
What happened
Speaking at the agency’s Frontier Forum in Washington on 28 October, Selig said businesses investing in AI infrastructure could need financial tools to hedge compute costs. He compared the idea with farmers hedging crop prices and airlines managing fuel costs. More than 100 people responded to a CFTC request for public comment on compute derivatives earlier this year, according to PYMNTS.
Selig also raised questions about autonomous trading agents: whether developers could face registration requirements, who would be liable if an agent went rogue, and how regulators might determine whose interests an agent represents. He floated a “know-your-agent” approach as a question for examination, not an announced requirement. The report says he cautioned that his views did not necessarily reflect those of the commission.
Why it matters
AI infrastructure brings exposure to the price and availability of processors, electricity and data-centre capacity. Futures or other derivatives could give companies a way to manage that uncertainty, though Selig offered no contract details or approvals. For AI builders, this is a potential financial tool, not a ready-made market to shop in.
Agentic trading raises a different problem: existing financial rules assume identifiable people and firms carrying out defined activities. If software can identify opportunities, hold assets and execute transactions, regulators will have to work out where control and responsibility sit. That is a policy question with practical consequences for developers, users and financial firms, not yet a new compliance checklist.
Our read
The useful signal is that the debate is widening from what AI can do to who pays when the infrastructure costs swing, and who answers when an agent acts. Selig has put those questions on the table, but a speech is not a rulebook. Keep an eye on what the CFTC does next, rather than treating “know your agent” as a policy that has already arrived.
What to watch
- Whether the CFTC publishes a response to the compute-derivatives comments or takes steps towards a market framework.
- How any proposed compute contracts define the resource, price and risks being hedged.
- Whether future CFTC proposals address registration, liability or agent identity, and how the commission distinguishes exploration from formal rules.
Discussion spark: If an AI agent makes a trade, should responsibility sit primarily with its developer, the firm deploying it, or the person who authorised it?
Sources and evidence
- Source update (29 October 2026, 15:46 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.