ChatGPT Ads Are Coming: What OpenAI's Ad Tests Mean for Paid Media Strategy

OpenAI went from calling ads "a last resort" to clearing $100 million in annualized ad revenue in six weeks. Sam Altman said in 2024 that he thought of ads as a "last resort" business model for OpenAI.

By March 2026, the company's ChatGPT ad pilot in the U.S. had surpassed $100 million in annualized revenue within six weeks of launch, signaling a new revenue path beyond subscriptions. That reversal happened faster than most PPC teams update their quarterly roadmaps. If you run paid media for a brand, agency, or growth team, this isn't a curiosity piece. Self-serve advertising tools are launching in April 2026, removing the $200,000 minimum commitment that restricted the platform to enterprise brands - meaning this is the inflection point between watching from the sidelines and actively competing for attention in conversational AI. The window between "early signal" and "crowded channel" is compressing. This post breaks down exactly what ChatGPT ads are, how they perform against existing channels, what the competitive landscape looks like, and what your team should do in the next 90 days.

What ChatGPT Ads Actually Are (and What They Aren't)

The first thing practitioners need to understand: ChatGPT ads are not injected into the model's answers. OpenAI has stated explicitly that ads do not influence the answers ChatGPT gives, that answers are optimized based on what's most helpful to the user, and that ads are always separate and clearly labeled. That distinction matters more here than on any platform before it, because the entire value proposition of ChatGPT rests on users trusting the response.

How the Ad Format Works

Ads appear below the end of a ChatGPT response and are clearly labeled as sponsored and visually separated from the organic answer. During the current test, the format is constrained to a single ad unit per response. Users may see a single ad unit below a ChatGPT response when there's a relevant match, and that unit may feature one or more items from an advertiser.

Think of it as a sponsored product card rather than a search text ad. From OpenAI's own examples, the early format looks like a sponsored product or service card that includes key details like the item name and price with a clear sponsored label. There are no banner placements, no interstitials, no pre-roll - at least not yet.

Ad targeting during the test works by matching ads submitted by advertisers with the topic of the user's conversation, past chats, and past interactions with ads. That's a fundamentally different mechanism from keyword bidding. The system reads context and intent from the conversation thread itself, then selects the most relevant advertiser.

Privacy Architecture

Ads are shown using information that stays within ChatGPT and is not shared with advertisers, including what the user is discussing in the current thread and basic context like general location. Advertisers never receive chats, chat history, memories, names, emails, precise location, or sensitive information.

Advertisers receive only aggregated performance metrics and are charged on a cost-per-view model based on total impressions. That's it. No conversation transcripts. No user-level behavioral data. If you're used to Google's granular keyword reports or Meta's audience breakdowns, the reporting here is sparse by comparison - early testers receive weekly CSV performance reports including only impressions and click data.

The Numbers Behind the Pilot: $100M in Six Weeks

Let's ground the conversation in what we actually know.

Revenue Trajectory

Roughly 85% of OpenAI's free and Go users in the U.S. are eligible to see ads, but less than 20% are shown them on a daily basis. That means the $100 million annualized figure was built on partial inventory with deliberate throttling. OpenAI says fewer than 20% of eligible users are seeing ads daily, even though around 85% of Free and Go users are eligible - the current run rate was built with partial inventory, limited distribution, and a controlled advertiser base.

The headroom is enormous. OpenAI executives have told investors the company expects to generate more than $17 billion from ChatGPT consumers in 2026, with advertising representing a meaningful slice of revenue from its free user base.

Who's Advertising and at What Price

Early participants include Target, Ford, Mrs. Meyer's, and Adobe, with holding companies WPP Media, Omnicom, and Dentsu as launch partners, and Criteo as the first formal ad tech partner.

OpenAI is now working with more than 600 advertisers.

The cost is premium. ChatGPT charges $60 per 1,000 ad impressions (CPM), about three times what Meta currently charges.

The ad commitments required to participate were unusually high for this type of experiment - some brands dedicated between $200,000 and $250,000, which is double a typical experimental ad commitment.

That $60 CPM positions ChatGPT closer to premium streaming and NFL broadcast inventory than to standard digital display. The justification? Users who are having detailed conversations with AI are typically in problem-solving mode - they're actively researching, comparing options, or making decisions. That's gold for advertisers.

Relevance appears strong so far. Fewer than 7% of ads are rated by users as "low relevance," according to the company. But click-through rates tell a more sobering story. Independent research from Ahrefs and ppc.land estimates ChatGPT's CTR at approximately 1.3%, compared to Google Search's 29.2%. That's a 96% gap. For performance-driven advertisers, this is the number that will define whether the channel justifies its CPM.

How ChatGPT Ads Differ from Google Search and Perplexity

This is not a Google competitor with a chat skin. The mechanics of targeting, user behavior, and competitive dynamics are diverging across every major AI platform.

Contextual Matching vs. Keyword Bidding

The targeting mechanism across all ChatGPT ad formats is context matching, not keyword bidding. OpenAI's system analyzes the topic and intent signals of the current conversation thread to match relevant advertisements. You're not bidding on "best CRM for small teams." The system interprets the full flow of a multi-turn conversation and determines relevance dynamically.

Instead of bidding on specific keywords, advertisers will likely be bidding on specific prompts and target personas. That's a meaningful shift for PPC practitioners who have spent years optimizing match types and negative keyword lists. The skill that transfers is understanding commercial intent. The skill that doesn't transfer is keyword-level granularity. Google, meanwhile, is weaving ads into its own AI experiences. Ads now appear at the bottom of approximately 25.5% of AI Overview SERPs, up from roughly 3% in January 2025, a 394% year-over-year increase.

Google expanded Ads in AI Overviews from mobile-only US availability to desktop and multiple global markets throughout 2025, integrating ads directly into AI-generated summaries.

Three Platforms, Three Philosophies

The AI ad landscape has fractured into three distinct camps.

Perplexity abandoned advertising entirely in February 2026 - after testing sponsored follow-up questions through 2024-2025, the company pulled all ads, citing user trust concerns. It's now targeting $500M in annualized subscription revenue, positioning itself as the ad-free alternative.

Anthropic has committed to keeping Claude ad-free, and ran Super Bowl ads that contrasted its stance with OpenAI's - helping the platform gain an 11% boost in daily active users post-game, per BNP Paribas data.

Google is extending its existing ad infrastructure into AI responses. Google SVP Nick Fox said the company is "not ruling out" ads in the Gemini app, noting that learnings from ad experiments in AI Mode would "likely carry over" to Gemini. Google is currently testing ads in AI Mode but the standalone Gemini app remains ad-free.

For paid media strategists, this divergence matters. You can't build a single "AI search ads" strategy. Each platform has different inventory rules, different user trust profiles, and different commercial intent signals. The answer depends on your audience, your objectives, and your tolerance for the trust dynamics reshaping how users interact with AI-generated content.

Who Sees These Ads - and Why the Audience Matters for PPC

The ad-supported audience is not the entire ChatGPT user base. OpenAI began testing ads in the U.S. for users on its Free and Go subscription tiers. The Go plan is a low-cost subscription at $8 per month. Subscribers to Plus, Pro, Business, Enterprise, and Education tiers will not see ads.

That narrows the pool - but the pool is still massive. According to OpenAI's February 2026 report, ChatGPT's weekly active users have reached 900 million, more than double the 400 million reported in February 2025.

eMarketer estimates 94.7 million US users are eligible to see ads.

The demographic profile skews exactly where most consumer brands want to be. Gen Z is dominant among demographics on the platform. The study shows 58% of adults under 30 use ChatGPT consumer plans, and nearly half of all messages come from users under 26. Adoption drops to just 10% for users over age 65.

The 25-34 age group represents the most active user segment at 29.67%, followed by 18-24 year-olds at 23.32%. Users under 25 represent 42% of the total base.

What makes this audience different from Instagram or TikTok users is behavioral, not just demographic. People often use ChatGPT when they're actively exploring options, comparing ideas, or working toward a decision - moments where ads can help people discover relevant products or services. Session durations support this. As of March 2026, ChatGPT users spent around 15 minutes and 26 seconds on the platform per session. That's not scrolling. That's problem-solving.

What Self-Serve Access in April Changes for Advertisers

The current pilot operates like a walled garden. It requires a $200,000 minimum commitment and operates exclusively through agency partners. That structure kept participation limited to enterprise brands and major holding companies. April changes the math.

Barriers Coming Down

For advertisers who balked at the $200,000 minimum, self-serve access represents the first realistic entry point. The question remains whether ChatGPT's conversational ad format can deliver performance metrics that justify premium pricing as the audience and tooling scale.

The move mirrors the path every major ad platform has taken: start with managed, high-spend partnerships, then open access to capture the long tail. Google, Meta, and TikTok all followed this trajectory. Nearly 80% of small and medium-sized businesses have signaled interest in ChatGPT ads, but the $200K floor has kept them out.

What the Platform Likely Includes

OpenAI hasn't published full feature specs for the self-serve tool. OpenAI has not publicly detailed every self-serve feature, so nobody should assume feature parity with established ad systems on day one. Even so, a self-serve rollout usually implies campaign setup, billing, creative submission, audience or context controls, budget management, approval workflows, and some level of performance reporting.

Don't expect Google Ads-level sophistication at launch. There are no confirmed bidding strategies, no conversion tracking pixels, no audience suppression tools. How ChatGPT handles conversion tracking across devices and sessions remains unclear - whether there will be pixel-based tracking is still an open question.

For performance marketers, that limitation is real. Without robust attribution, justifying spend against channels with decades of measurement infrastructure requires a different framework - one built on early learning, brand presence in a new medium, and directional signal rather than last-click ROAS.

Practical Steps to Prepare Your Paid Media Strategy

The self-serve launch is weeks away. Here's what to do now, whether you plan to advertise immediately or watch for another quarter. 1. Register at openai.com/advertisers. OpenAI has a signup page for businesses interested in advertising in ChatGPT. Get on the list before April inventory competition increases. 2. Audit how ChatGPT already describes your brand. Type your brand name and services into ChatGPT. Then watch how it describes your business. If the responses are incomplete or inaccurate, you now know where to focus your improvements. This matters both for organic visibility in AI answers and for the contextual relevance system that decides which ads to serve. 3. Rethink creative for a conversational context. To succeed in the ChatGPT Ads era, growth teams must move beyond traditional display assets and adopt a conversational mindset. Audit conversational keywords - use ChatGPT to see how it describes your product category and incorporate those natural language terms into your ad copy.

Your headline needs to work as a helpful recommendation, not an interruption. OpenAI demands ad experiences that are "tasteful." According to Dentsu's EVP of Paid Search, "ad delivery is quickly building momentum, with volume increasing week-over-week as the environment scales."

4. Carve out a test budget. Even $5,000–$10,000 earmarked for Q2 testing will let you understand the platform mechanics, creative requirements, and directional performance signals before CPMs potentially rise with broader access. 5. Prepare your measurement stack. Even before you advertise on ChatGPT, the platform is already referring users to websites. Setting up a clear attribution model for ChatGPT-sourced traffic - distinguishing it from organic and direct in your analytics setup - will give you real data on current ChatGPT-driven behavior.

6. Document your product information thoroughly. If interactive ads let users ask follow-up questions, you'll need structured information ready: product specs, pricing details, policy information, competitive differentiators. ChatGPT ads reward information density. Thin landing pages won't convert users who just finished a deep-context research conversation.

The Trust Problem No One Has Solved Yet

The $100 million number is exciting. The 7% low-relevance rating is encouraging. But the structural tension at the heart of ChatGPT advertising hasn't been resolved.

An Ipsos survey found that nearly two-thirds (63%) of US adults say ads in AI search results make them trust the results less. That's a large gap between what OpenAI's internal metrics show and what the broader public feels about AI-embedded advertising. Perplexity's retreat is the sharpest counter-signal. A Perplexity manager told the Financial Times that "the challenge with advertising is that a user would simply start to doubt everything," which is why the company no longer views advertising as a "fruitful thing." They tried it, measured it, and walked away.

IAB research from January 2026 shows Gen Z skepticism toward AI ads is rising, and clearer standards plus disclosure are needed to rebuild trust. That matters especially because Gen Z is the dominant ChatGPT demographic. OpenAI is betting that clear labeling and structural separation solve the problem. Ads do not influence ChatGPT's answers - ads run on separate systems from the chat model, and advertisers have no ability to shape, rank, or alter responses. Whether users believe that claim long-term is the variable that will determine whether this channel thrives or erodes.

Former OpenAI researcher Zoë Hitzig warned in a New York Times commentary that economic pressures from an upcoming IPO, planned for the fourth quarter of 2026, could create strong incentives to override the company's privacy commitments. Advertisers should monitor this carefully. If user trust breaks, performance will follow - and the brands who were early in will bear the reputational cost alongside OpenAI. --- Paid media strategy has always been about going where attention concentrates before costs catch up. ChatGPT has 900 million weekly active users, and over 2 billion daily queries are processed on the platform. That's not speculative future traffic. It's a live audience, actively researching purchase decisions, in a format where sponsored content gets a single premium placement per response. The honest assessment: this channel is early, measurement is weak, and the trust model is unproven at scale. But the honest assessment also includes the fact that it hit $100 million in annualized revenue with less than a fifth of eligible users exposed to ads, and that self-serve access in April will remove the barrier that kept most teams out. The brands that build institutional knowledge about conversational ad formats now - understanding contextual targeting, adapting creative for AI environments, and tracking ChatGPT referral traffic - won't be scrambling when this platform matures into a core media channel. That's not hype. It's the same pattern every major ad platform has followed, just moving faster than any before it.