Forty-two states. One subpoena. One lab heading into an IPO.
According to the Wall Street Journal, a coalition of 42 US state attorneys general has reportedly
launched a joint investigation into OpenAI. The New York AG reportedly served OpenAI with a subpoena
on June 12, 2026, demanding internal documents on advertising practices, user data collection,
treatment of minors and seniors, deep-learning algorithm design, and internal safety policies. OpenAI
reportedly stated it’s cooperating with the inquiry, though that statement hasn’t been independently
confirmed from fetched source content.
No official press release from any attorney general office was available at the time of publication. All specific claims about the investigation, the count of 42 states, New York’s leadership role,
and the subpoena’s scope, are sourced to the Wall Street Journal’s reporting and require that
qualification throughout.
The timing matters. OpenAI filed a confidential S-1 with the SEC, entering the IPO pipeline. A pending multi-state subpoena is now a material legal risk that must be disclosed or addressed in that filing. The investigation was opened shortly after the confidential S-1 was filed, the connection is factual, though the characterization of it as strategically timed by the AGs is an editorial inference, not a stated fact.
The consumer protection theory here is worth understanding. This isn’t a sectoral AI regulation
action. State AGs aren’t citing AI-specific statutes, they’re reportedly applying existing consumer
protection authority to OpenAI’s advertising practices, data collection, and treatment of vulnerable
populations. That’s the same legal toolkit used against tobacco manufacturers, opioid distributors,
and social media platforms. It doesn’t require new legislation. It requires an existing consumer harm
theory and enough states to coordinate.
Florida’s prior consumer protection suit
against OpenAI was a single-state action filed in early June 2026. The jump from one state to a
reported 42-state coalition in under two weeks is the pattern, not just the number.
What to Watch
Don’t expect the investigation to resolve quickly. Multi-state AG actions at the subpoena stage are
typically the beginning of a process that runs 18-36 months before any enforcement action. But the
subpoena itself is the compliance event. OpenAI must now preserve documents across the reported scope,
engage outside counsel for a multi-jurisdiction response, and disclose material legal risk in its S-1. Every other AI company operating a consumer-facing product should read the reported subpoena scope as
a preview of where state-level enforcement attention is heading.
The real question is whether the federal preemption provisions in pending AI legislation –
actively contested in Congress
– can reach state AG consumer protection authority, or whether that authority sits outside the
preemption window entirely. If it does, the 42-state model becomes the default AI accountability
mechanism in the absence of federal law. Compliance teams at every consumer AI company should be
building for that scenario now, not waiting for the OpenAI investigation to resolve.