ChatGPT ads are sponsored placements that appear below AI responses on ChatGPT's free and Go tiers. Launched in beta in February 2026, they are matched to conversation topics, clearly labelled, and entirely separate from ChatGPT's answers. Advertisers buy them via OpenAI's self-serve Ads Manager using CPC or CPM bidding.

In late 2024, OpenAI's public position was clear: ChatGPT ads were not in its plans. By February 2026, the beta was live. By May 2026, any marketer with a credit card could buy ChatGPT ads with no minimum spend. That 18-month reversal is not a footnote. ChatGPT now has 900 million weekly active users (Business of Apps, April 2026). OpenAI projects $2.5 billion in ad revenue in 2026, growing to $100 billion by 2030 (Reuters, April 2026; based on investor projections). The world's most-used AI product is now an ad-supported media company, and the window to test early, before CPCs normalise and inventory tightens, is open right now.

This guide covers what ChatGPT ads are, how they work, what they cost, and what B2B performance marketers need to do before spending anything. It also addresses the misconceptions circulating in every PPC and agency forum. Most of what's been written about this channel was written for the press, not for practitioners.

What are ChatGPT ads?

ChatGPT ads are sponsored placements that appear directly below the AI's response on ChatGPT's free and Go tiers. They don't appear for Plus, Pro, Team, or Enterprise subscribers. They are contextually matched to the topic of the conversation, clearly labelled as "Sponsored," and structurally separate from ChatGPT's answer.

That last point matters more than any other. The most persistent misconception among marketers is that buying ads changes what ChatGPT says about your brand. It doesn't. ChatGPT's answers and the sponsored placements below them are independent systems, built that way deliberately. OpenAI calls this "answer independence," and it's a core principle of the product architecture.

This distinction also separates ChatGPT ads from Answer Engine Optimization (AEO), the organic practice of structuring content so AI engines cite your brand in their responses. Both tracks exist inside the ChatGPT ecosystem. They work through entirely different mechanisms. You likely need both.

How did ChatGPT go from rejecting ads to self-serve in 18 months?

Understanding how quickly this moved helps you calibrate how seriously to take it.

Late 2024: firm public rejection.

OpenAI's public position through most of 2024 was that advertising was not in its plans. The reasoning centred on user trust and preserving the integrity of AI-generated answers.

Early-mid 2025: signals emerge.

Internal code references to ad infrastructure surfaced in developer forums. OpenAI quietly began hiring from ad-tech teams at Google and Meta. The signals were easy to dismiss individually. In aggregate, the direction was clear.

January 2026: official announcement.

OpenAI formally announced the advertising programme. Within weeks, the first beta partners were live, including agency holding groups WPP, Omnicom, and Dentsu.

February 2026: beta goes live.

The beta launched at roughly 1% of ChatGPT users, with a minimum spend of $200,000–$250,000 per campaign. Large brands and enterprise advertisers only. Within six weeks of launch, the programme crossed $100 million in annualised revenue with over 600 advertisers participating (Reuters, April 2026).

Mid-March 2026: rapid expansion.

Sensor Tower data found ads had reached 5% of ChatGPT mobile users by mid-March, a 600% increase from the start of the month (CNBC, March 2026).

May 2026: self-serve opens.

OpenAI launched a self-serve Ads Manager at ads.openai.com, removed minimum spend requirements, and began international expansion. UK access is confirmed. This is the moment the channel became accessible to teams outside enterprise budgets.

How do ChatGPT ads work?

Format and placement.

Ads appear directly below the AI's response, before any suggested follow-up prompts. The format is closer to a display unit than a text search ad. Creative specs have evolved through beta, so treat current guidelines as the starting point. OpenAI is iterating fast.

Ads only appear on free and Go tier accounts. Plus, Pro, Team, and Enterprise users don't see them. This limits total addressable impressions but still covers a substantial share of ChatGPT's global user base.

Buying options.

Inventory is available through OpenAI's Ads Manager at ads.openai.com. Current options:

• CPM: approximately $60 (roughly three times the social media average)

• CPC: $3–$5 recommended range

The CPM is high relative to social channels but reasonable for a captive, active-intent environment. A direct CPC comparison with Google Search isn't meaningful yet, because the targeting and measurement systems operate very differently.

How targeting works.

This is where ChatGPT ads differ most from what performance marketers are used to. There's no keyword bidding. There's no audience segmentation built on cross-platform behavioural data. There's no user-level tracking.

Ads are matched contextually to the conversation topic. If a user is asking ChatGPT about project management software, a relevant advertiser can appear below the response. But OpenAI reports on an aggregated, privacy-preserving basis only. You don't get individual user data or the granular conversion paths you'd expect from Google or Meta.

This has direct implications for measurement, covered in detail below.

How do ChatGPT ads compare to other channels?

vs. Google Search Ads.

Google Search targets declared keyword intent. ChatGPT targets conversation context. In Google, you know the exact search term that triggered your ad. In ChatGPT, you know the broad topic category. Google has more than 20 years of measurement infrastructure behind it. ChatGPT is months into public advertising.

Google's 2025 search ad revenue was $294.69 billion (Reuters, 2026). Truist estimates ChatGPT will generate under $1 billion in ad revenue in 2026, growing to over $30 billion by 2030 (CNBC, March 2026). These are not comparable businesses yet. The question isn't whether to replace Google with ChatGPT ads. It's whether to test early, at manageable spend, while CPCs are low and inventory is not yet competitive.

For context on where ChatGPT sits relative to Google as an information discovery platform, the data shows a meaningful but not yet dominant shift in how people find answers.

vs. Meta and social.

Meta ads are behavioural. ChatGPT ads are contextual. A Meta user sees an ad while scrolling passively. A ChatGPT user sees an ad while actively asking a question about a relevant topic. For B2B use cases, where active research conversations matter more than passive feed exposure, the intent signal is arguably stronger.

The CPM comparison reflects this. OpenAI is pricing for intent, not volume.

The AI advertising competitive landscape.

ChatGPT is not the only AI product building an ad business, though it's furthest along.

• Google AI Overviews already runs sponsored placements within AI-generated search results.

• Perplexity briefly ran ads, then removed them. Its current advertising status is not settled.

• Anthropic (Claude) has stated it will remain ad-free, at least for now.

• Grok (xAI) has no formal advertising product at the time of writing.

Truist analysts called 2026 an "inflection year" for LLM-powered advertising and described it as likely to become "one of the most important pillars of the digital ad industry" (CNBC, March 2026). That's a projection, not a certainty. But Google, OpenAI, and Meta are all building toward AI-native advertising at scale, and the infrastructure is moving fast.

What do ChatGPT ads mean for B2B performance marketers?

The intent signal case for B2B.

The timing argument for B2B is stronger than it looks. When a buyer types "what's the best project management software for a 50-person engineering team" into ChatGPT, they're in an active research conversation. That's mid-funnel intent, expressed in natural language. Traditional paid search captures declared intent at the keyword level. ChatGPT captures it at the conversation level, with more context around the question.

Whether that translates to better conversion rates for B2B has no published benchmarks behind it yet. The channel is too new. The logic holds, though: a buyer using an AI to research a purchase category is further into a buying process than someone passively scrolling a feed.

Budget allocation: search, social, or neither?

The question appearing most often in PPC forums right now: should ChatGPT ad budget come from search, social, or a separate line? There isn't a data-backed answer yet.

The sensible position for most teams is to treat early tests as exploratory, funded from a small innovation allocation or underperforming channel budget. Don't move significant spend from Google or Meta based on current evidence. Do run a structured test with clear hypotheses and measurement protocols in place before you start.

The organic and paid interplay: you need both tracks.

Buying ads in ChatGPT doesn't mean you appear in ChatGPT's organic answers. The two tracks are entirely separate, and this is the second most common misunderstanding about this channel.

AEO is the practice of optimising content so AI engines cite your brand in responses. It's organic. ChatGPT ads are paid. Neither replaces the other. A brand already cited organically in AI responses is likely to see better ad performance, because the brand exists in the AI's context before the ad appears. If you haven't started the organic side yet, the practical guide to ranking in AI search is the right starting point.

For teams managing both paid and organic AI visibility, the complexity is real. You're running two parallel programmes, each with different mechanics, inside the same environment.

How do you measure ChatGPT ad performance?

This is the biggest practical challenge with ChatGPT ads today.

OpenAI uses aggregated reporting only. There's no user-level data, no cross-platform conversion pixel, and no click-path granularity. You won't get the attribution detail you're used to from Google Ads or Meta.

Four approaches that work under these constraints:

1. Branded search uplift as a proxy. If ChatGPT ads are running, monitor Google Search for any increase in branded query volume during and after campaign periods. Not perfect attribution, but a signal of awareness impact.

2. Incremental lift testing. Run the same campaign in test and holdout markets simultaneously. Compare conversion rates between exposed and unexposed audiences using a statistically sound holdout structure.

3. UTM parameters on landing pages. ChatGPT passes UTM parameters when users click through to your site. Configure GA4 to capture campaign source data before launch. You won't have user-level profiles, but you will have channel-level data.

4. Accept the measurement immaturity upfront. If your organisation requires last-click attribution to justify spend, ChatGPT ads won't satisfy that requirement yet. Teams already comfortable with multi-touch attribution and incremental measurement are better positioned to evaluate this channel. Teams that can't move without perfect attribution should wait for the measurement infrastructure to mature.

How do you get started with ChatGPT ads?

Access.

Go to ads.openai.com. As of May 2026, self-serve is live with no minimum spend. UK access is confirmed. International expansion continues throughout the year.

Five things to prepare before you spend anything.

1. Run an AEO audit first. Understand how your brand currently appears in organic ChatGPT responses before you buy ads. A brand already present in organic AI context is better positioned to convert paid clicks. Start with the AEO guide.

2. Review your landing page quality. ChatGPT users clicking through an ad are in an active research mindset. Pages that match that intent (educational, direct, low friction) will outperform campaign pages optimised for social audiences.

3. Define your conversation contexts. Write out the specific topic categories your ICP is likely discussing when they'd benefit from finding your brand. "Best CRM for SaaS startups" is a context. "AI marketing tools comparison" is a context. Build targeting briefs from these, not keyword lists.

4. Choose a bidding model. Start with CPC ($3–$5 range) rather than CPM unless your goal is pure awareness. CPC gives you direct feedback on whether placements are generating clicks worth paying for, before committing to impression-based spend.

5. Set up conversion tracking before you launch. UTM parameters on every destination URL, GA4 events configured, and a pre-campaign baseline documented. You need these to evaluate performance under aggregated reporting. Without them, you'll have spend and no signal.

What to watch for.

The channel is moving fast. Three things worth tracking:

• CPC benchmark maturation. Early-mover CPCs are low. As advertiser volume grows, expect increases. The window for low-cost tests is finite.

• International availability. OpenAI is expanding throughout 2026. Confirm your primary markets have self-serve access before building campaigns.

• Format updates. Creative specs and placement positions have already changed through beta. Build adaptability into your creative production workflow from day one.

Where do ChatGPT ads fit in your broader marketing strategy?

ChatGPT ads represent the third major paradigm shift in digital advertising, after search and social. The first took a decade to mature. The second took five years. This one is moving faster, because the infrastructure, the buyers, and the intent signals already exist.

The brands in the best position to take advantage of it aren't the ones with the biggest budgets. They're the ones that built organic AI search visibility early, understand how their audiences research across AI platforms, and have measurement frameworks flexible enough to work with imperfect data.

Managing complexity across Google, Meta, LinkedIn, organic channels, and now an AI-native paid channel isn't trivial. The advantage goes to teams whose systems compound: every new channel, every new signal, every new performance outcome feeding back into a clearer picture of what's working. For the full AI search strategy that ChatGPT ads sit inside, GEO vs SEO: what's the difference and why it matters is the broader frame.

Frequently asked questions.

How much do ChatGPT ads cost? As of May 2026, there's no minimum spend on the self-serve platform at ads.openai.com. CPM runs at approximately $60 (OpenAI Help Centre, May 2026). CPC typically falls in the $3–$5 range. The earlier beta required minimum campaigns of $200,000–$250,000, but those requirements have been removed.

Do ChatGPT ads affect what ChatGPT says about my brand? No. Paid placements and ChatGPT's responses are structurally separate systems. An ad appearing below a response has no influence on the content of that response. OpenAI has built this in deliberately and describes it as "answer independence."

Should ChatGPT ad budget come from my search budget or social budget? Neither, for now. Treat early tests as exploratory, funded from a small innovation allocation rather than cannibalising proven channels. The published data to justify a significant budget shift from Google or Meta doesn't exist yet. The case for acting now is about building early learnings cheaply, before CPCs increase and inventory becomes competitive.

What's the difference between ChatGPT ads and appearing organically in ChatGPT? They're entirely different tracks. Organic presence in ChatGPT responses comes from AEO and GEO work: structuring and distributing content so AI engines cite your brand naturally. Paid ads are bought placements below the response. Neither substitutes for the other. Building organic presence first likely improves paid ad performance, because the brand is already in the AI's context when the ad appears.

Managing organic AI visibility and paid AI placements without building separate programmes for each is exactly the kind of problem an intelligent marketing system solves. See how DOJO handles the whole picture.