OpenAI Starts Selling Ads Inside ChatGPT
The company is opening a formal ad platform aimed at the moment when people are comparing products and getting ready to buy, with Newegg among the first brands on board.
OpenAI has launched a dedicated advertising platform for ChatGPT, giving brands a way to pay for placement inside the chatbot's conversations rather than relying on organic mentions alone. The new offering is built around a self-serve tool called Ads Manager, live now at ads.openai.com, and OpenAI is sweetening the pitch for newcomers with a $500 ad credit for any advertiser who spends their first $500.
The pitch to marketers is specific. People don't just ask ChatGPT for facts, OpenAI argues, they use it to weigh options, compare products, and talk themselves into, or out of, a purchase. That's a different kind of attention than a search results page or a social feed scroll, and OpenAI is framing it as a moment of "real intent" that advertisers haven't had a clean way to reach before. Because people share more context in a conversation with ChatGPT than they'd type into a search bar, OpenAI says the ads it serves can be more relevant and personalized, rather than just more frequent.
How the ads actually work
Signing up drops advertisers straight into Ads Manager, where they set up an account, build campaigns, and lean on OpenAI's Help Center for step-by-step guidance. That's the fast, self-serve path OpenAI is pushing. For bigger spenders, the company also offers direct sales support for larger advertisers with more complex needs.
Newegg is one of the names OpenAI is using to sell the idea to other brands. Alec Shao, a performance marketing manager at the retailer, said in a statement OpenAI shared: "We want Newegg to be part of the conversation when someone is researching what to buy. Showing up in ChatGPT helps us build consideration at that moment." It's a tidy summary of the whole strategy: get the brand into the research phase, not just the moment right before checkout.
The trust problem OpenAI is trying to get ahead of
Running ads inside an AI assistant is a different proposition than running them beside a list of search results, and OpenAI seems aware of the tightrope it's walking. The company describes its approach as "measured," tying it explicitly to preserving the trust people already place in ChatGPT. It lays out three commitments: ads will be clearly labeled as ads, they'll remain distinct from ChatGPT's own generated responses, and users will have control over how their data is used for ad targeting.
That last point carries more weight here than it would on an ordinary website. ChatGPT already knows a great deal about what a user is trying to figure out in any given conversation, which is exactly why advertisers find the format compelling. It's also why the real test of this system won't be whether ads are labeled correctly, but whether that rich conversational context genuinely stays walled off from shaping the chatbot's actual answers, rather than just sitting next to them. OpenAI's framing insists the two stay separate. Whether that holds once ads are running at scale is something the company's own materials don't fully settle.
Why this matters beyond ChatGPT
This also marks a new chapter in how OpenAI plans to make money. Advertising hasn't been part of the ChatGPT experience before now, and the company is visibly building the infrastructure that any ad business needs to scale: a self-serve manager, a sales team for larger accounts, and a credit program designed to get brands trying it with low risk.
It's also a signal about where OpenAI thinks the real commercial value in conversational AI lives. Not at the top of the funnel, where someone asks a broad question, but deeper in, at the comparison-and-decision stage that search advertising has quietly profited from for decades. If ChatGPT can plant a brand in that moment without users feeling like the assistant's judgment has been bought, OpenAI will have pulled off something search ads never quite managed: placement that reads as help instead of interruption. If it can't, this becomes just another ad business wearing a conversational interface.
Either way, the test starts now. The platform is live, the credit is on the table, and the next question is simply how users react the first time they realize the thing answering their question is also, quietly, selling an ad.
