Originally published on The Searchless Journal
Two months ago, ChatGPT had no ads at all.
Today, advertisers can pay for outcomes — actual purchases, sign-ups, downloads — inside the most-used AI interface on the planet. OpenAI has switched on cost-per-action bidding inside its Ads Manager, completing a full performance marketing stack that took Meta nearly a decade to build. OpenAI did it in sixty days.
The speed is the story. But the implication is bigger: ChatGPT is no longer just an AI assistant that happens to show ads. It is becoming a performance advertising channel with the reach, intent data, and conversion infrastructure to compete directly with Meta and Google for the $600 billion global digital ad market.
This is not a test. OpenAI has hired three senior Meta ads veterans to build this business, is projecting $102 billion in annual ad revenue by 2030, and is expected to go public before the end of the year. The CPA launch is the clearest signal yet that OpenAI intends to make advertising a core revenue pillar — not a side experiment.
What Just Happened
On May 28, Digiday confirmed that OpenAI has activated cost-per-action (CPA) bidding inside its ChatGPT Ads Manager. Advertisers using the platform can now pay only when a user completes a specific action — a purchase, a sign-up, an app download — rather than paying for impressions (CPM) or clicks (CPC), which were the previous options.
The feature is currently available to select advertisers in a pilot phase. OpenAI has told advertisers that broader access will open by June 5, 2026. The CPA model requires the conversion tracking pixel that OpenAI shipped in recent weeks, which lets advertisers measure whether a ChatGPT ad view actually led to an action on their website or app.
This is the third pricing model OpenAI has added in rapid succession. The sequence tells you everything about the strategy:
- CPM (cost per thousand impressions) — launched first, establishing inventory and baseline pricing
- CPC (cost per click) — added to attract direct-response advertisers who wanted engagement accountability
- CPA (cost per action) — the final piece, letting advertisers pay only for verified conversions
Each step moved ChatGPT ads further up the performance marketing funnel. CPM is for awareness. CPC is for engagement. CPA is for outcomes. OpenAI built the entire ladder in eight weeks.
Asad Awan, OpenAI's head of monetization, teased the move three weeks ago during an advertiser briefing: "It's in the plan. Should be done soon." Soon turned out to mean weeks, not quarters.
The Meta Brain Drain
OpenAI didn't build this by accident. The company has been systematically poaching talent from Meta's ads organization — the team that built one of the most sophisticated advertising machines in history.
Dave Dugan, who previously led Meta's ads product, is now running OpenAI's advertising efforts. Archana Joshi, who spent seven years at Meta in ads go-to-market roles, has joined as a founding member of OpenAI's Ads Go-To-Market Revenue and Strategy Operations team. Sam Mulinder, formerly Snap's head of marketing science for North America, is building out OpenAI's marketing science function — the measurement and attribution infrastructure that performance advertisers demand before committing real budgets.
This is not a skunkworks team. This is an ads organization assembled from people who have collectively managed hundreds of billions of dollars in ad spend at the companies that invented modern digital advertising. They know how to build self-serve platforms, how to train advertisers on new inventory, and how to scale from pilot to programmatic.
The hiring strategy sends an unambiguous message to Madison Avenue: ChatGPT ads are not a novelty. They are being built by the same people who built the tools you already use every day.
Why CPA Changes Everything
The shift from CPC to CPA sounds incremental. It is not.
When advertisers pay per click, they bear the risk. They pay for traffic whether or not it converts. The publisher — in this case, OpenAI — gets paid regardless of outcome. This creates a natural ceiling on advertiser willingness to spend: they'll bid aggressively for clicks only if they have high confidence in their conversion rates.
CPA flips the risk. When advertisers pay per action, OpenAI bears the conversion risk. If an ad generates clicks but no purchases, OpenAI doesn't get paid. This forces OpenAI to optimize its ad delivery not just for engagement but for actual outcomes — showing ads to users who are most likely to buy.
This is the model that made Google's search ads the most efficient advertising medium in history. Google could charge premium prices because advertisers knew they were paying for clicks from people with high purchase intent. CPA takes this one step further: advertisers pay only when the purchase actually happens.
For OpenAI to offer CPA at this stage — barely two months into its ads business — is aggressive bordering on audacious. It means OpenAI is confident enough in its user intent data and conversion tracking to guarantee outcomes. Or it means OpenAI is willing to subsidize early advertiser success to build the marketplace fast.
Probably both.
The Financial Stakes
OpenAI's financial trajectory adds urgency to the advertising push. The company burned through $2.5 billion in 2025, and that figure is expected to balloon to $8.5 billion in 2026 as compute costs for training and inference continue to scale. ChatGPT's subscription revenue, while growing, cannot close a gap that wide on its own.
Advertising is the bridge. OpenAI is projected to generate $102 billion in annual ad revenue by 2030, according to multiple analyst estimates. Even if that number proves optimistic by half, the strategic direction is clear: ChatGPT must become an ad-supported platform at scale, and it must happen faster than anyone expected.
The IPO timeline compounds the pressure. OpenAI is widely expected to go public before the end of 2026, and investors will want to see a credible path to profitability. Advertising revenue — with its high margins and compounding network effects — is the most convincing narrative OpenAI can offer. The CPA launch, the Meta hires, and the full-funnel buildout are all designed to demonstrate that the ad business is real, operational, and scaling.
What This Means for Brands
The implications depend on what kind of brand you are. But no one gets to ignore this.
If you're a performance advertiser
You now have a new CPA channel with potentially unmatched intent data. ChatGPT users are not scrolling passively through a feed. They are asking specific questions — "what's the best CRM for a 50-person startup," "recommend a protein powder that's third-party tested," "help me find a winter jacket for Tokyo weather." These are high-intent queries that look more like Google search than Instagram browsing.
The brands that get into ChatGPT's ad program early will have access to this intent data at a discount. Early CPMs have already become cheaper as inventory expands, and CPA pricing during the pilot phase will be favorable while OpenAI is trying to prove conversion value to advertisers.
The window is narrow. Once the programmatic infrastructure matures and more advertisers compete for the same inventory, pricing will normalize upward. The next three to six months are the land-grab phase.
If you're a brand that relies on organic discovery
CPA ads create a new competitive surface inside ChatGPT that didn't exist two months ago. When a user asks ChatGPT for a product recommendation, the answer is now influenced by both organic citation signals and paid advertising placements. A competitor willing to pay for CPA placements can appear in front of your potential customer — even if your product is objectively better or more relevant.
This means that optimizing for AI visibility is no longer just about organic citation. It's about defending your brand presence in an environment where paid and organic signals are converging. Brands that invested early in GEO — generative engine optimization have a head start on the organic layer. But they'll now need to think about the paid layer too.
If you're an agency
Add ChatGPT ads to your channel matrix. The CPA model means you can run performance campaigns with measurable ROI from day one — no need to justify experimental CPM spend to skeptical clients. The ex-Meta team building OpenAI's ads infrastructure means the tooling will feel familiar to anyone who has run Facebook or Instagram campaigns.
The ChatGPT Ads vs Google Ads comparison we published last week gives a framework for thinking about channel allocation. The short version: ChatGPT ads are strongest where users are in research and consideration mode, not transactional search mode. That makes them complementary to Google for most advertisers, not a replacement.
The Full Funnel, Built in 60 Days
The pace of OpenAI's ads buildout is unprecedented. Here's the timeline:
- Late March 2026: OpenAI quietly launches its Ads Manager to select advertisers
- April 2026: CPM bidding available, minimum spend requirements in place
- Early May 2026: CPC bidding added; minimum spend removed; market expansion to UK, Brazil, Japan
- Mid-May 2026: Conversion tracking pixel shipped; product catalog automation enabled
- May 28, 2026: CPA bidding activated for pilot advertisers
- June 5, 2026 (projected): Broader CPA access
For context, Meta launched its first ad product in 2004. It didn't offer CPA bidding until 2012 — eight years later. Google launched AdWords in 2000 and didn't introduce full CPA capabilities until 2007. OpenAI compressed a decade of ad infrastructure development into two months.
The compression is possible because the playbook already exists. OpenAI didn't need to invent self-serve ad platforms, conversion pixels, or CPA bidding models. The ex-Meta hires brought institutional knowledge of how these systems work. What OpenAI did need — and what it had — was the user base. ChatGPT has over 400 million weekly active users. That's an ad inventory that materialized overnight, from the advertiser's perspective.
The Risk OpenAI Is Taking
Offering CPA this early is a bet on conversion quality. If ChatGPT ad clicks don't convert at rates that justify CPA pricing, OpenAI eats the cost. The company is essentially guaranteeing advertiser outcomes during a phase when it's still learning which users are most likely to convert.
This is a classic marketplace strategy: subsidize the supply side (advertisers) to build demand (conversion data), then use the data to optimize delivery and raise prices. It's what Uber did with driver subsidies, what Amazon did with seller fees, and what Meta did with advertiser support programs in its early years.
The risk is that ChatGPT's user base, while massive, may not behave like the audiences on Meta or Google. People use ChatGPT to write, code, research, and brainstorm. They don't come to ChatGPT primarily to shop. The conversion intent may be weaker than what Google sees on search or what Meta sees on Instagram.
But here's the counter-argument: ChatGPT's conversational interface generates richer intent signals than any ad platform in history. When a user types "I need a CRM for my 50-person startup and I want something with good email automation," that query contains more purchase intent data than a hundred Google searches for "best CRM." If OpenAI can match that intent with relevant ads, CPA performance could exceed expectations.
Early data from pilot advertisers will be telling. If conversion rates are strong, OpenAI's ad business could scale faster than even the $102 billion projection suggests. If they're weak, the company will need to adjust pricing models and targeting capabilities — a process that could take quarters, not weeks.
The Competitive Landscape
OpenAI's CPA launch doesn't happen in isolation. The entire AI search and discovery market is consolidating at breakneck speed.
Microsoft is building a Copilot super app that combines GitHub Copilot, Copilot Chat, Copilot Cowork, and a new "Autopilot" agentic workflow tool into a single interface. Google is expanding Gemini's reach through Drive integration and AI Overviews. Anthropic just raised $65 billion at a $965 billion valuation, signaling investor confidence that AI platforms will become trillion-dollar ecosystems.
All of these platforms will eventually need to monetize through advertising. OpenAI is simply the first to build a full performance marketing stack. That first-mover advantage matters: advertisers have limited budgets and limited attention spans. The platform that proves CPA performance first will capture the initial wave of performance marketing spend.
The brands that understand this dynamic — that AI platforms are becoming the new ad networks — will be the ones that build early expertise and secure favorable pricing. The ones that wait for the market to mature will pay premium rates for inventory that early movers secured at a discount.
What to Do Right Now
If you're a brand or agency reading this, here's the play:
1. Apply for ChatGPT ad access immediately. The pilot is still small. Getting in now means cheaper inventory, more attentive support from OpenAI's ad team, and time to learn the platform before it scales.
2. Set up conversion tracking. The CPA model depends on OpenAI's pixel. Install it now, even if you're not running ads yet. The data will help you understand how much ChatGPT traffic is already arriving at your site organically.
3. Audit your AI visibility. Before you invest in ChatGPT ads, understand how your brand already appears in AI-generated answers. If your organic AI presence is weak, paid placements will be more expensive because ChatGPT's ad delivery algorithm favors brands with strong organic signals. Run an AI visibility audit to establish your baseline.
4. Allocate test budget. You don't need to shift your entire Meta or Google budget. But setting aside 5-10% of your digital ad spend for ChatGPT testing in Q3 2026 will give you data that your competitors won't have. Performance marketing is an arbitrage game, and the arbitrage opportunity is now.
5. Think about organic + paid together. The brands that win in ChatGPT won't be the ones that run the best ads or have the best organic visibility. They'll be the ones that combine both — appearing in organic citations for high-intent queries while running targeted CPA campaigns for conversion-ready users. This is the same playbook that worked on Google for a decade, but applied to an AI-native interface.
The digital ad market has been a duopoly for over a decade. Meta and Google controlled the majority of performance advertising spend, and every new platform — TikTok, Snapchat, Pinterest — tried to carve out a niche without fundamentally disrupting the two giants.
OpenAI is different. It has the user base (400M+ weekly actives), the intent data (conversational queries), the leadership (ex-Meta ads veterans), and the financial motivation ($8.5B burn rate, IPO pressure) to build a genuine third pillar in performance advertising. The CPA launch is the clearest proof point yet.
Whether OpenAI captures 5% or 25% of the digital ad market over the next five years is an open question. What's not open is whether brands should start paying attention. That window is already closing.
Want to know how visible your brand is inside ChatGPT?
Run a free AI Visibility Audit to see whether ChatGPT, Perplexity, Gemini, and AI Overviews are citing your brand — or recommending your competitors instead.
👉 Start your free audit at audit.searchless.ai
Sources
- Digiday. "OpenAI turns on cost-per-action ads inside ChatGPT." May 28, 2026.
- AdExchanger. "Friday 29052026 — Retail Wags The Dog; Data Brokers Vs. The Military." May 29, 2026.
- Digiday. "OpenAI opens up ChatGPT Ads Manager to the U.S. while promising third-party measurement, CPA bidding." 2026.
- The Information. "OpenAI's first-half results: $4.3 billion sales, $2.5 billion cash burn." 2026.
- Enders Analysis. Claire Holubowskyj on CPA as logical next step for ChatGPT ads. 2026.
- Fortune. "Exclusive: Microsoft is building a super app." May 29, 2026.
- OpenAI Blog. "Codex for (almost) everything." May 29, 2026.
- Digiday. "OpenAI has quietly launched its ads manager." 2026.
FAQ
What is CPA advertising?
Cost-per-action (CPA) advertising means advertisers pay only when a user completes a specific action — like a purchase, sign-up, or download — rather than paying for impressions or clicks. It's the most performance-oriented pricing model in digital advertising because the platform (OpenAI) shares the conversion risk with the advertiser.
How is ChatGPT's CPA model different from Google Ads?
Google Ads offers CPA primarily through search ads where users are already in a transactional mindset. ChatGPT's CPA model operates inside a conversational interface where users are in research and consideration mode. The intent signals are richer (full conversational context) but the purchase readiness may be lower. Over time, as ChatGPT adds shopping features and agentic commerce capabilities, this gap could close.
When will ChatGPT CPA ads be available to all advertisers?
OpenAI has told pilot advertisers that broader CPA access will open around June 5, 2026. Early CPMs have become cheaper as inventory expands, and CPA pricing during the pilot phase is expected to be favorable.
Should brands shift budget from Meta/Google to ChatGPT?
Not wholesale. The prudent approach is a 5-10% test allocation in Q3 2026 to gather performance data while keeping core spend on proven channels. ChatGPT ads are strongest in research and consideration phases, making them complementary to Google's transactional search ads and Meta's awareness and retargeting strengths.
How does CPA advertising affect organic AI visibility?
Paid placements and organic citations operate on separate signals, but they interact. A brand with strong organic AI visibility (cited frequently by ChatGPT, Perplexity, and Gemini) will likely see better ad performance because the platform recognizes the brand as authoritative. Conversely, heavy ad spend without organic visibility may produce weaker results because the AI has less context to recommend the brand.
Ready to see where your brand stands in AI search? Explore Searchless pricing for ongoing AI visibility monitoring and GEO strategy.
Top comments (0)