Six weeks. One ad platform. $100 million in annualized revenue.

That's what OpenAI's first real shot at advertising produced — and now they're taking it global. On March 26, The Information reported that ChatGPT's ad pilot had crossed $100 million in annualized run rate since its February launch. As of this week, OpenAI is expanding beyond the US with pilots in Canada, Australia, and New Zealand, and has announced that self-serve advertising tools will launch in April.

The company that once said it was building AI for humanity just became a full-blown advertising business.

Over 600 advertisers are already running campaigns inside ChatGPT. Eighty-five percent of users are eligible to see ads, though fewer than 20% actually see them on any given day. That gap is deliberate — and it won't last. OpenAI has already hired Dave Dugan, a former top advertising executive at Meta, to lead ad sales as the platform prepares to scale.

When an ad runs in ChatGPT, it appears clearly labeled as sponsored and separated from the organic answer. Advertisers have been told they don't have access to users' chat history or personal details. That's the official line — for now. But self-serve tools launching in April means the number of advertisers could jump from hundreds to hundreds of thousands within months. That's when targeting pressure starts.

Analysts project ChatGPT will generate under $1 billion in ad revenue in 2026. By 2030, that estimate climbs to over $30 billion. To put that in perspective, Google's entire advertising business in 2012 — 14 years into its existence — was about $43 billion. OpenAI is trying to build a comparable ad empire in a fraction of the time.

The company used subscriptions first. GPT-5.4, released this month, is available to paid users. Now ads are unlocking a free tier funded by advertiser spend. It's the same playbook that built Google and Facebook into surveillance capitalism empires — give the product away, charge the advertisers.

My Opinion

I'll be blunt: this was inevitable, and anyone surprised by it wasn't paying attention.

Sam Altman spent years framing OpenAI as the antidote to Big Tech's worst habits. The company's early positioning was explicit — we don't sell your data, we don't optimize for engagement, we're a safety-focused research organization building AI for humanity. That framing helped OpenAI attract talent, disarm regulators, and generate press coverage that no PR budget could buy.

Here's what bugs me: the ad rollout is being framed as expanding access. Ads help fund free ChatGPT. That's technically true. It's also exactly what Facebook said in 2012, and exactly what Google said when it launched AdWords. The justification is identical because the business model is identical. You give people a free, useful product. You monetize their attention. You gradually improve targeting to increase revenue per user. You repeat until you've built a $200 billion advertising machine.

OpenAI hiring a former Meta advertising executive to run this operation is the tell. You don't hire someone who spent years optimizing Facebook's ad targeting to build a hands-off, low-key sponsored label program. You hire them because you want to win. And winning in digital advertising means knowing your users well enough to show them something they'll click. The IPO is reportedly coming later this year. The ad revenue will be front and center in the prospectus. Watch how quickly the language shifts from "expanding access" to "personalized advertising experiences" — that's exactly where this road leads, and they hired someone who's been down it before.

Author: Yahor Kamarou (Mark) / www.humai.blog / 28 Mar 2026