A year ago, running ads inside ChatGPT wasn't something most small businesses could even ask about. The pilot program was invite-only, and the minimum spend to get in the door was $50,000.
That's no longer true. Over the past several months, OpenAI has quietly opened advertising on ChatGPT to any business willing to sign up, dropped the budget floor that kept smaller advertisers out, and built the pieces (reporting partners, a Shopify app, a creative toolkit) that make it look less like an experiment and more like a real channel.
If you're wondering whether this is something your business should actually test, or just another AI headline to skim past, here's what's confirmed, what's still early, and what's worth watching before you spend a dollar on it.
What happened
OpenAI rolled this out in stages across 2026, not all at once, which is part of why it's easy to have missed.
In January, OpenAI confirmed ads were coming to ChatGPT for adult users in the US, kept clearly labeled and separate from the chatbot's answers. By May, it opened a public beta of its Ads Manager and dropped the $50,000 minimum spend that had limited the earlier pilot to large advertisers working directly with OpenAI's sales team, a change OpenAI's own monetization lead framed as opening the platform to businesses without big budgets or an agency behind them.
In September, OpenAI added two features aimed specifically at smaller, less technical advertisers: a ChatGPT Ads for Shopify app that connects directly to a merchant's existing product catalog (US merchants first, expanding to other markets later that month), and a HubSpot integration for businesses that manage customers there already. It also introduced Sponsored Agents, still in testing with a small group of US advertisers, where clicking an ad can open a separate, clearly labeled conversation with a business-run AI agent.
Most recently, on October 5, OpenAI announced a new visual ad format that appears during image generation in ChatGPT, rolling out this month to Free and Go-plan users, along with a broader measurement push: integrations with attribution platforms like AppsFlyer, Triple Whale, and Northbeam, early incrementality-testing partners, and brand-safety evaluation pilots with DoubleVerify and Integral Ad Science.
Put together, this went from a closed, enterprise-only pilot to something close to a self-serve ad platform, with the connective tissue (ecommerce integration, measurement, creative tools) a real small business would actually need to run it.
Why it matters
ChatGPT reaches about 1.2 billion people every week, by OpenAI's own figure. Even a sliver of that is a real audience, and arguably one with more purchase intent than a casual social scroll: someone asking ChatGPT to compare running shoes or plan a trip is closer to a buying decision than someone mid-scroll between videos. OpenAI has reportedly set internal targets around $2.5 billion in ad revenue this year and $100 billion by 2030, which signals this isn't a side experiment for them.
More practically: every new ad channel goes through a window where early advertisers get cheaper attention than the channel will ever offer again, before competition and ad prices catch up. That's not guaranteed to repeat here, but it's happened with every major platform that opened self-serve advertising, from early Facebook Ads to early TikTok Ads. Whether it's worth jumping on depends on what's actually proven so far, not on how new or interesting it sounds.
What's actually proven so far
The honest answer: not much yet, and OpenAI's own language reflects that.
The strongest results circulating come from OpenAI's measurement partners, reporting on a small number of named brands. WeightWatchers' cost per acquisition on ChatGPT Ads came in 15.3% lower than its blended paid-search benchmark, according to DV Rockerbox. WorkMagic reported that 67% of wellness brand Dose's incremental purchases came from genuinely new customers. Triple Whale found that 93% of visitors Portland Leather got from ChatGPT Ads were new to the brand.
Those are promising, but they're a handful of case studies supplied through OpenAI's own measurement-partner program, not independent, broad-based results, and OpenAI's own materials describe its incrementality testing as still in its early stages. Treat these as early signal, not proof this will work for your business or category.
What's still unverified or worth watching
A few things are worth knowing before treating this as a settled channel.
There's no confirmed minimum spend right now, but the platform is still a US-only beta, which limits it if your customers are elsewhere. OpenAI says advertising doesn't influence ChatGPT's own answers, but reporting on the May beta launch has pointed out that OpenAI hasn't published evidence for that claim, and that the platform's billing is moving toward click-based pricing, which would give OpenAI a financial incentive to show ads exactly where someone's making a purchase decision.
There's also real outside scrutiny. In January 2026, Senator Edward Markey sent OpenAI and several other AI companies a formal letter raising concerns about ads in chatbots: whether they could be mistaken for the chatbot's own neutral advice, whether sensitive conversations about health or family matters could ever be used for targeting, and whether paid placement could quietly shape what a chatbot recommends. OpenAI has pointed to its stated safeguards (labeled ads, no targeting from sensitive conversations, no ads for minors) in response, but the specific questions the letter raised haven't been independently verified one way or the other.
What to actually do
If you're curious, the cost of finding out is low. Sign-up is at ads.openai.com, no large budget is required to get in the door anymore, and if you already sell through Shopify or manage customers in HubSpot, most of the integration work is close to done for you.
Treat it like a genuine test, not a channel switch. Run it alongside whatever's already working, with a budget small enough that a disappointing result doesn't actually hurt you, and don't move budget away from a channel you already trust until you have your own data, not someone else's case study, saying it's worth it.
What not to overreact to
You're not behind. This is weeks old as a self-serve platform, limited to the US, and still missing pieces (independent measurement, a track record beyond a handful of showcased brands, clarity on the outstanding privacy questions) that take time to resolve. Nothing here means your existing channels stopped working, and nothing here means you need an "AI advertising strategy" by next quarter. It means there's a new option worth a small, low-risk look, not an emergency.
What to watch next
The real tests ahead: independent data instead of partner-supplied case studies, whether a minimum spend quietly reappears as the platform matures, how OpenAI responds to the specific questions in Senator Markey's letter, and whether the platform expands past the US. If you do test it, measure your own numbers against your other channels, not against OpenAI's own highlighted results.
The bigger lesson here
Every new advertising channel asks you to decide with incomplete information. Jump in too early and the proof doesn't exist yet. Wait too long and the cheap attention is gone. The honest move isn't picking a side before you have your own data. It's treating the first few months of any new channel as a small, deliberately low-stakes experiment, with a budget you can afford to lose and a clear question you're actually trying to answer, rather than jumping in because it's new or dismissing it because it's unproven.
Sources & Further Reading
- New ChatGPT ad format and measurement — OpenAI, Oct 5, 2026
- Reimagining advertising with AI — OpenAI, Sept 16, 2026
- ChatGPT Ads are now open to small businesses as OpenAI builds a full self-serve ad platform — The Decoder, May 6, 2026
- Letter to OpenAI on AI chatbot advertising — Sen. Edward J. Markey, Jan 22, 2026
- OpenAI projects $2.5 billion in ad revenue this year and $100 billion by 2030 — eMarketer, 2026