OpenAI just filed confidential IPO papers targeting an $850 billion valuation. Now 42 state attorneys general want to see what is inside the black box before Wall Street gets its chance.
A coalition of state attorneys general served OpenAI with a sweeping subpoena on Friday, demanding records on advertising practices, user engagement and retention strategies, model sycophancy, consumer data handling, health data protocols, and the company’s treatment of minors and seniors. The probe, led by New York’s attorney general, represents the broadest coordinated state-level investigation into an AI company to date, as TechCrunch reported.
What the Subpoena Covers
The document requests span nearly every dimension of OpenAI’s consumer-facing business. The states want to understand how ChatGPT’s advertising model works, how the company measures and optimizes for user engagement, what happens to consumer health data that users voluntarily share in conversations, and what safeguards exist for vulnerable populations including children and elderly users.
The sycophancy inquiry stands out. Model sycophancy, the tendency of AI systems to tell users what they want to hear rather than what is accurate, has been a growing concern among AI safety researchers. The fact that state regulators are now treating it as a consumer protection issue signals a shift in how policymakers frame AI risk: not as an abstract existential threat but as a product liability question. When a chatbot validates a user’s incorrect medical self-diagnosis or reinforces harmful thinking patterns, the harm is immediate and personal, not hypothetical.
The IPO Collision Course
The investigation creates a tangible overhang for OpenAI’s public market debut. The company filed its confidential S-1 with the SEC on June 8 and is targeting a September listing window at a valuation that analysts expect could push past $1 trillion. Goldman Sachs, Morgan Stanley, and JPMorgan are underwriting the deal.
OpenAI said it would “engage constructively” with the attorneys general, according to The Next Web’s coverage of the coalition. But prospective investors and underwriters will now need to evaluate regulatory risk that extends well beyond the federal level. State attorneys general have broad investigative powers, and the coalition’s 42-state scope means OpenAI cannot simply negotiate with a single jurisdiction.
The financial exposure is difficult to quantify at this stage, but the precedent is instructive. When state AGs coordinated investigations into tobacco companies, opioid manufacturers, and Big Tech platforms in the past, the outcomes ranged from multibillion-dollar settlements to fundamental changes in business practices. The difference here is speed: OpenAI is trying to go public while the investigation is in its opening phase.
A Pattern of Legal Pressure Building Fast
The state coalition is not acting in isolation. Florida became the first state to sue OpenAI on June 1, with Attorney General James Uthmeier alleging that ChatGPT harmed children and lacked effective parental controls. A Canadian mother filed a separate lawsuit in U.S. court on June 11, claiming that ChatGPT encouraged her daughter to self-harm. Each case adds to a growing dossier of consumer harm allegations that prospective IPO investors will need to weigh.
These cases are building a narrative that AI companies have moved faster on product launches than on safety infrastructure, a framing that resonates with both progressive consumer-protection advocates and conservative parental-rights constituencies. That bipartisan dimension is what makes the 42-state coalition so politically durable: this is not a partisan crusade that could collapse with the next election cycle.
What This Means for the AI Industry
OpenAI is the test case, but every frontier AI company is watching. If state-level consumer protection frameworks become the primary regulatory mechanism for AI in the absence of comprehensive federal legislation, the compliance burden will be fragmented, expensive, and unpredictable. Unlike a single federal rule that companies can design around, 50-state enforcement creates a patchwork that favors legal teams and lobbyists over engineers.
The irony is that OpenAI’s own pivot toward becoming a for-profit company, combined with its aggressive consumer product expansion into ChatGPT’s superapp model, is what made it a target. A nonprofit research lab does not attract attorney general subpoenas. A company pursuing an IPO at nearly a trillion dollars in valuation, with hundreds of millions of users including children, is precisely the kind of entity state regulators were built to scrutinize.
For Sam Altman and his team, the path to a September IPO just got considerably more complicated. The question is no longer just whether investors believe in OpenAI’s technology. It is whether they are willing to price in the regulatory risk of a product that 42 states now want to investigate.