OpenAI has replaced the fixed daily spending cap in ChatGPT Ads Manager with a seven-day average daily budget, according to a report published by PPC Land on 4 August 2026. OpenAI told beta advertisers about the change on 24 July 2026 and scheduled the switch for the week of 27 July 2026, roughly nine weeks after daily budgets first appeared on 22 May 2026 as a hard per-calendar-day limit.
The sourcing needs stating before anything else. The document underneath this story is OpenAI's weekly product-update email to ChatGPT Ads Manager beta advertisers, not a public announcement. PPC Land is reporting on that email, and there is no public OpenAI post to check the wording against. Everything below should be read as what OpenAI told a beta list, not as something OpenAI announced publicly.
What OpenAI told beta advertisers on 24 July 2026
OpenAI's wording, as quoted by PPC Land, is that "Daily budgets will change from fixed daily limits to average daily budgets: the average you want to spend per day over seven days." OpenAI added that "No action is required, and your total budget is not changing," and that spend "may vary by day while staying within daily and seven-day limits."
That is a change in behaviour rather than a relabelling. Under the version OpenAI shipped on 22 May 2026, the daily budget was a ceiling and the campaign stopped at the number. From the week of 27 July 2026, the same field is a target the system averages across a rolling week, so an individual day can run above the figure typed into the box as long as the week reconciles.
The single-day ceiling OpenAI asserted but never quantified
This is the part worth slowing down on. OpenAI's email refers to "daily and seven-day limits" but publishes no maximum daily figure, and PPC Land said so directly in its 4 August 2026 report: "The seven-day limit is calculable: seven times the average daily budget. The daily limit is asserted but never quantified." An advertiser can therefore work out the weekly ceiling with arithmetic and cannot work out the daily one at all, which leaves anyone reconciling spend to an approved media plan holding a known weekly bound and an unknown daily one.
Google publishes its multipliers, OpenAI has not
PPC Land drew the contrast with Google, and the figures in this paragraph are Google's published behaviour for Google Ads rather than anything OpenAI has stated. Google publishes that spend can reach up to twice the average daily budget on high-traffic days, and that monthly billing is capped at 30.4 times the average daily budget. No equivalent numbers have been published for ChatGPT Ads. Teams running Google Ads campaigns will recognise the mechanic immediately; the difference is that Google tells you the worst case for a single day before it happens.
What is published and what is not
The table below separates the limits an advertiser can actually rely on from the ones that do not exist in public yet, with the party that states each figure.
| Limit | Value | Who states it |
|---|---|---|
| ChatGPT Ads seven-day ceiling | Seven times the average daily budget | Calculable from OpenAI's wording; the arithmetic is PPC Land's |
| ChatGPT Ads single-day ceiling | No figure published | Asserted by OpenAI in the beta email, never quantified |
| Google Ads daily overspend | Up to 2x the average daily budget on high-traffic days | Google, published |
| Google Ads monthly billing cap | 30.4x the average daily budget | Google, published |
What this means for Thai marketers
The source does not address Thailand, and it does not state whether ChatGPT Ads Manager is available to advertisers in Thailand, so what follows is the practical read rather than reported fact. If a Thai team does have beta access, the reconciliation unit changes: checking yesterday's spend against yesterday's budget will now produce variances that are not errors. The honest check is the seven-day total.
For accounts on tight monthly retainers, the unquantified single-day overshoot is a cash-flow question more than a performance one. A day that runs hot is fine if the week lands, but it can still breach an internal daily approval threshold or a client's spend alert. Setting those alerts at the weekly level rather than the daily one avoids a fortnight of false alarms.
Budget mechanics inside an AI assistant are still being written in beta emails rather than documentation. Teams working on visibility inside AI answers and on how brands surface in ChatGPT should expect that pattern to hold: features arriving before the reference material that explains them.
Questions advertisers are asking
Do I need to change anything in my ChatGPT Ads campaigns?
No, according to OpenAI's own wording. OpenAI told beta advertisers that "No action is required, and your total budget is not changing." The source does not state whether existing campaigns were migrated automatically beyond that sentence, so there is no confirmation of the mechanics of the switch itself.
How much can ChatGPT Ads spend in a single day now?
The source does not state it. OpenAI refers to a daily limit but publishes no figure for it, which is the central gap PPC Land identified. The seven-day ceiling is calculable as seven times the average daily budget; the daily one is not calculable from anything OpenAI has released.
Did OpenAI announce this publicly?
No. The underlying document is OpenAI's weekly product-update email to ChatGPT Ads Manager beta advertisers, and PPC Land is reporting on that email. There is no public OpenAI post to verify the wording against, which is why this should not be described as a public announcement.
Is ChatGPT Ads Manager available in Thailand?
The source does not state it, and neither availability nor a market list appears in the report. Any claim either way about Thai access would be invention rather than reporting.
Where do the 2x and 30.4x figures come from?
From Google, not OpenAI. Those are Google's published Google Ads behaviours, cited by PPC Land as the contrast: Google quantifies its daily and monthly overspend multipliers, and OpenAI has not published anything comparable for ChatGPT Ads.
The short version
An average daily budget is a normal way to run a bidding system and nobody should be alarmed by the mechanic itself. What is unusual is a spending limit that is referred to but never given a number, disclosed in a beta mailing list rather than documentation. Until OpenAI publishes a daily multiplier, manage a ChatGPT Ads budget against the seven-day total. If you want a second pair of eyes on how AI-assistant channels sit alongside your existing paid search, that conversation is cheaper before the budgets get larger.
