The Maze: ChatGPT ads are no longer a thought experiment. In Neil Patel's post, Adthena data showed 580 ads across commercial queries, with activity split between a few repeat advertisers and a lot of one-off testing. That is the interesting part. The market is not empty. It is also not mature. It looks like the awkward early stage of a new auction: enough signal to learn, not enough crowding to make every lesson expensive.
The top 25 advertisers already account for 55% of observed ads. The source panel splits the 580-ad sample into 32% from the top 10 advertisers and 23% from ranks 11-25, while only 9% sits in the remaining
Otherbucket. That is concentration without saturation. A limited group is appearing often enough to learn faster than the field, but the long tail has not yet filled the channel with repetitive competition. This is why early inventory matters. The first advantage is not just lower media cost. It is cheaper pattern recognition.The largest single bucket is still experimentation. One-time advertisers make up 36% of the sample, which means many brands are touching ChatGPT ads without yet showing repeat depth. That is a classic new-channel shape: curiosity is broad, commitment is narrow. The danger for late movers is that the experimental stage does not last. Adthena's related launch material says its ChatGPT Ads Intelligence monitors 300,000+ daily prompts across ad presence, competitor activity, and prompt triggers. Once that kind of measurement becomes normal, testing becomes more disciplined and auctions become less forgiving.
The named leaders are not fringe AI brands. Lowe's leads the visible ranking with 25 ads, followed by Nordstrom at 22, Best Buy at 20, Liberty Mutual at 12, and Insurify at 10. That mix spans retail, consumer electronics, department stores, insurance, and comparison shopping. In other words, ChatGPT ads are already attracting categories where a user's question can sit close to purchase intent: what to buy, where to buy it, which offer to trust, and which provider deserves a quote request. This is less about novelty and more about intercepting commercial research before it becomes a traditional search click.
The placement format changes the learning curve. Adthena's separate analysis says ChatGPT ads appear inline inside conversational answers, with short copy, brand-led headlines, concrete numbers, and action-specific calls to action. That makes the creative job different from Google search. Brands are not only bidding against other advertisers. They are trying to sound useful inside an answer that the user already expects to be concise and helpful. Repurposed search copy may technically fit. It may also feel like a banner ad walked into a library.
The caveat is that visibility is not performance. The source does not disclose the prompt set, geography, sampling period, or whether an
admeans a unique creative, an observed placement, or another counting unit. It also does not show clicks, conversion, cost, or return on ad spend. So the right conclusion is not that Lowe's or Nordstrom are winning ChatGPT ads. The safer read is that they are among the most visible early testers in this observed sample, and that visibility can compound if it teaches them which prompts and offers work before everyone else piles in.
Why it matters: New ad channels usually create two windows. The first is cheap attention. The second is cheap learning. The first closes when buyers arrive. The second closes when the playbook becomes common knowledge. ChatGPT ads may still be early, but the source data suggests the learning window is already unevenly distributed. For ecommerce and lead-gen brands, the question is not whether the channel is fully proven. It is whether competitors are using this messy phase to learn the prompts, offers, and creative patterns that will define the cleaner phase.
