How can ChatGPT reach $100BN in advertising revenue?

Citing fundraising documents, both The Information and Axios report that OpenAI expects to reach $100BN in advertising revenue by 2030. It’s a lofty, ambitious goal, although the company revealed two weeks ago that it had reached a $100MM revenue run rate just six weeks after launching its advertising pilot.

I published some “napkin math” on X following that revelation to estimate just how much revenue OpenAI might generate if it expanded its advertising pilot globally; based on (crude) comparisons to other large ad platforms across geo-level ARPU skew and free-tier usage, I calculated that OpenAI’s advertising revenue from ChatGPT might scale to just over $1BN, annualized, if the company “flipped the switch” and made it available globally, for all free users.

What makes this type of analysis difficult is that OpenAI is currently charging advertisers on a CPM basis (reportedly, $60 CPM) for ad impressions, which almost certainly couldn’t hold if supply expanded dramatically. This model will inevitably transition to objective-based conversion pricing at some point, which is how other large platforms like those from Meta and Google generate revenue. This transition represents OpenAI’s opportunity to improve advertising ARPU over time: CPM is a reach strategy, but conversion optimization is an ARPU strategy, and it is through conversion optimization that modern digital advertising platforms provide value to direct response advertisers. I’ve previously likened brand advertising revenue to a “participation trophy” in digital advertising.

The reporting around OpenAI’s $100BN target for advertising revenue in 2030 outlines other projections for the product on that timeline:

• 2.75BN WAU, compared to a reported 920MM today
• $60 ARPU (I take this to mean globally, blended), up from $3.5 today

Given these constraints, combined with the assumptions I present in my “napkin math,” it’s worth considering how OpenAI might reach $100BN in advertising revenue by 2030. It’s clearly a combination of WAU and ARPU growth, given its own internal forecasts. A global, blended ARPU of $60 in 2030 implies CAGR from 2026 of 151%:

But how does it get there? ChatGPT’s advertising revenue is a function of regional ARPU and WAU. As I note in my “napkin math” missive, ARPU in the US and Canada tends to be meaningfully higher than in other geographic regions; any optimization “solver” that attempts to project to $100BN in advertising revenue and $60 global, blended ARPU must take that into account. In order to do that, I created rough estimates of Meta’s annual ARPU by region from its most recent 10K and created a weighting framework for ChatGPT’s advertising ARPU on that basis:

At a high level, the problem reduces to solving for a WAU distribution and ARPU distribution that jointly satisfy OpenAI’s revenue and ARPU targets. Note that Meta’s global ARPU for its family of apps in 2025 was $57.03.

Given this geographic ARPU weight profile, the next step in solving for $100BN in advertising revenue with a $60 global, blended ARPU is projecting WAU ratios across these same regional definitions. Roughly 85% of ChatGPT’s Free and Go users in the US are currently eligible to see ads, and roughly 15% of ChatGPT’s current user base is in the US. Projecting these numbers forward to a global WAU of 2.75BN by 2030, and anchoring the geographic distribution in 2030 to 10% in UCAN, 20% in Europe, 42% in APAC, and 28% in ROW produces the WAU mix below:

With that WAU mix, the geographic ARPU values become:

On a blended basis, this implies that ARPU grows at a far higher CAGR than WAU:

Given the current scale of the consumer product, it’s natural that ARPU is the principal lever that OpenAI must adjust to achieve its $100BN advertising revenue goal. Since the company’s current implementation of advertising is fairly primitive, it also offers the most headroom for expansion. As I write in OpenAI should know better:

OpenAI should know better. And the team there likely does. If OpenAI’s advertising platform doesn’t evolve meaningfully in six to twelve months to include conversion-optimized targeting and vastly better measurement, I’ll be skeptical of its success. But until then, given the high density of ex-Meta employees working for the company, I’m willing to give OpenAI the benefit of the doubt. While commercial traction is by no means guaranteed and will be hard won, OpenAI’s advertising bona fides are substantial, and its success with advertising is easier to rationalize ex ante than other companies that have charted this course.

If my reverse-engineered input assessments are anywhere near what OpenAI used in its model to project $100BN in annual advertising revenue, then the team faces a formidable challenge, albeit one that is surmountable with the right product and execution discipline. This clearly has implications for the product: OpenAI needs to build not only a highly functional optimization engine for use in advertising targeting, but also an ad unit that resonates roughly as effectively as those found on Meta’s properties today. And it needs to do both of those things in parallel, quickly. That’s no trivial task: increasing ad load risks degrading retention, and therefore constraining user base scale. And these aggressive goals leave little room for error in either ARPU or WAU growth.