In an October 21, 2026 post, OpenAI technical staff Ritendra Datta and Karthik Gopalratnam described the advertising system built for ChatGPT. According to OpenAI, traditional ad systems respond to discrete events such as a search query or a page view, while ChatGPT conversations develop intent across multiple turns and sessions. To address this, the company says it built a custom, post-trained large language model that orchestrates retrieval, ranking, relevance evaluation, and auctions for ad selection and pricing. OpenAI outlines three design decisions: stateful ad sessions that remember information across turns and conversations; a tool-call architecture that interprets conversation context and intent; and running the ad-selection process in a separate service from ChatGPT's answer generation. The company states this separation is intended to keep ads from influencing answers and to maintain what it calls ChatGPT's neutral voice. OpenAI reports that ChatGPT Ads reached a $1 billion annualized revenue run rate less than 200 days after launch, and that more than 100,000 advertisers have run ads on ChatGPT since launch. These figures are reported by OpenAI about its own product and have not been independently verified. The post also describes filtering, eligibility checks, and human feedback used to steer ad placement.
- OpenAI reports ChatGPT Ads hit a $1B annualized revenue run rate in under 200 days (OpenAI's own figure).
- More than 100,000 advertisers have reportedly run ads on ChatGPT since launch.
- Ad selection runs in a separate service from answer generation, per OpenAI.
What it means for you
ChatGPT now carries ads, and OpenAI says it keeps the ad system separate from the answers themselves. For most teams, the practical takeaway is simpler than the engineering: if your staff use free ChatGPT for research, buying decisions, or recommendations, some of what they see may be paid placement rather than a neutral answer. That matters when someone is, say, picking software or a vendor based on what the tool suggests.
Try this
This week, ask whoever leans on ChatGPT for vendor or product research to treat its suggestions as a starting point, not a verdict — and to confirm any shortlist against sources outside the chat.
Who should care
Anyone whose team uses free ChatGPT to research purchases, tools, or vendors, and anyone who markets a product that competes for attention in those conversations.
Skip this if
Your team runs on paid or enterprise tiers without ads, or you don't use ChatGPT for buying or recommendation decisions — then this is just background.
Sources: OpenAI — read the original