In the space of 48 hours last week, OpenAI did three things that, taken together, sketch the strategy of a company preparing to go public. It endorsed a California bill strengthening the state’s AI safety law, reversing its earlier opposition. It cut API prices on its flagship model by up to a third. And it turned on advertising in 31 European markets. On Aug. 21 and 22, the company that talks about safety, competes on price and monetizes attention was doing all three at once.
The regulatory move was the sharpest reversal. OpenAI publicly backed strengthening SB 53, California’s frontier AI transparency law, according to people familiar with the company’s thinking, a shift from its previous stance of opposing the bill. The about-face, reported on Aug. 21, came as the company faces scrutiny over safety incidents involving its systems and as it prepares for a public listing, where regulators and investors will both be watching.
SB 53 is already the first enforceable frontier AI law in the United States. Signed by Gov. Gavin Newsom in September 2025 and in effect since Jan. 1, it requires developers of the largest models to publish transparency reports, document risk assessments, report critical safety incidents and protect whistleblowers, with civil penalties of up to $1 million per violation. The new push aims to strengthen those requirements, and OpenAI’s endorsement gives the effort a prominent corporate backer it previously lacked.
The company’s own documents show how far it has come. In May, OpenAI published a Frontier Governance Framework that mapped its practices to SB 53 compliance, an acknowledgement of the law’s reach even before the company voiced support for tightening it. The sequence — opposition, compliance paperwork, endorsement — tracks a broader pattern of AI labs positioning themselves for regulation they once resisted, as the political center of gravity shifts from whether to regulate frontier AI to how much.
The price cut was aimed at a different audience: developers. OpenAI reduced API pricing for GPT-5.6 Sol, cutting input costs by 20% and output costs by 33%, and the discount runs through Nov. 21. The company framed the move as making its most capable model cheaper to build on, and it arrives as open-weight models from Chinese and other labs increasingly undercut closed-model APIs on cost. The AI price war, which began with smaller models, has now reached the flagship tier.
The advertising launch completed the week. ChatGPT ads began appearing this week in 31 European markets, OpenAI’s largest ad rollout to date, extending a business line the company has been building since it introduced advertising to the assistant in select markets. Advertising gives OpenAI a revenue stream that does not scale linearly with computing cost, and it diversifies a business that still depends heavily on API usage and consumer subscriptions.
Read together, the three moves are the playbook of a company trying to be everything to everyone at the moment before its stock trades publicly. To regulators, it presents itself as a responsible actor accepting oversight. To developers, it is the cost-efficient choice at a moment when rivals are cutting prices. To advertisers, it is a distribution platform reaching tens of millions of users. Each audience gets a tailored message, and the company gets the option value of all three.
The IPO context is hard to miss. OpenAI has been reported to be preparing a public listing that would be among the largest in technology history, and companies in the quiet period before a filing tend to avoid regulatory fights. Endorsing the strengthening of SB 53 removes a potential conflict with California, its home state, at a moment when the attorney general is the bill’s enforcer. Analysts said the move should be read as de-risking for a public-market debut.
The timing also reflects a real change in the regulatory climate. Other states have followed California’s lead — New York passed a frontier AI law in December, Connecticut followed in May — and the direction of travel is toward more oversight, not less. OpenAI’s endorsement positions it inside that trend rather than against it, and it gives the company influence over how the rules are written.
There are costs to the strategy. The price cut pressures margins that investors will scrutinize in a public filing, and the advertising business brings its own complications, including questions about how ads are labeled inside an AI assistant and whether free users’ data is used for ad targeting. The regulatory endorsement may also alienate libertarian-leaning users and researchers who see SB 53’s transparency regime as the first step toward heavier controls.
For the industry, OpenAI’s week is a preview of what AI competition looks like when the frontier companies mature into public corporations: price competition for developers, advertising for consumers, and regulatory engagement for governments. The three moves reinforce each other. A cheaper model grows usage, which makes the ad inventory more valuable, and a cooperative posture with regulators protects both businesses from legal risk.
Whether the strategy holds will be tested in the months ahead. The Nov. 21 price deadline gives developers a reason to build on GPT-5.6 Sol now and a reason to evaluate alternatives later. The European ad rollout will show whether advertising inside an AI assistant works at scale. And the SB 53 endorsement will be tested the first time California asks OpenAI to disclose something it would rather keep private. Last week, the company chose its sides on all three fronts.
