The Maze: ChatGPT Ads just changed what “daily budget” means. From July 27, the number is an average across seven days, not a hard same-day ceiling. OpenAI can now spend up to twice the selected amount on one day, provided billed media spend stays within seven times that amount across the week. A $100 daily budget can therefore become a $200 day, but not more than $700 over seven days. The budget survived. The calendar certainty did not.
OpenAI has taken more control over timing, not total weekly spend. Existing campaigns move to the new model automatically, with no need to recreate, pause or edit them. The entered budget and other campaign settings remain unchanged. Ads Manager will show the associated maximum daily and seven-day figures when advertisers create or edit a campaign. That disclosure matters because the interface still calls the input a daily budget, while the actual control is now two linked limits: 2x for one day and 7x for the week.
The mechanism is ordinary ad-platform pacing. Delivery opportunities do not arrive evenly. Some days produce more eligible conversations, stronger matches or better-priced inventory than others. A rigid same-day cap can leave useful demand unbought. A weekly average lets OpenAI front-load spend and compensate later. Google Ads uses a similar 2x daily ceiling, although its usual reconciliation window is monthly at 30.4 times the average daily budget. ChatGPT’s seven-day window is shorter, but the transfer of timing control is the same.
Advertisers still have a stricter total-spend option. OpenAI’s campaign guide lets buyers choose a daily budget or a campaign-total budget. The latter is the harder ceiling across the campaign, but OpenAI warns that it is not a pacing control: spend can accumulate quickly rather than distribute evenly across the scheduled dates. So there is no setting that promises both exact daily smoothness and automated opportunity chasing. Teams must choose which risk matters more—unused inventory or lumpy cash flow.
The change arrives while ChatGPT Ads is still learning how to be an ad platform. OpenAI began rolling out self-serve Ads Manager on May 5, adding CPC buying and conversion measurement alongside CPM. Advertisers set budgets, bids and pacing, but OpenAI’s system controls delivery. The current campaign guide also supports conversion-optimized CPC, where the system seeks clicks more likely to produce a tracked action. More optimization creates more reasons for the platform to move money toward stronger moments—and more need for the buyer to audit what it moved.
Commerce teams should manage the week, but watch the day. A launch, flash sale or short promotional window can now absorb up to twice the number entered as the daily budget. That may be useful when ChatGPT finds high-intent product conversations. It may also front-load spend before a retailer has enough conversion data to distinguish signal from novelty. New-channel tests need daily alerts at the explicit 2x ceiling, weekly reconciliation against the 7x cap, and performance checks by objective. “Average” is doing a lot of work in the label.
Why it matters: ChatGPT Ads is moving from beta novelty toward familiar platform economics: the advertiser sets the budget, while the platform decides when opportunity deserves more of it. The new limits are not reckless—there is a clear 2x daily ceiling and a 7x weekly ceiling—but they shift operational control toward OpenAI. Ecommerce teams testing the channel should stop treating the daily field as a cash-flow promise. It is now a pacing instruction. The weekly cap protects the wallet; daily monitoring protects the plan.
