OpenAI to expand ads on ChatGPT to all free and low-cost users, The Information reports - Yahoo Finance
OpenAI will begin testing ads on ChatGPT — and the company is starting with logged-in adults in the United States on its free and low-cost Go tiers.
The company confirmed Friday that it will begin this US test in the coming weeks, with ads appearing for users on the Free and ChatGPT Go ($8/month) versions of the product. The expansion follows a limited pilot earlier this year and now formalizes advertising as a structural part of OpenAI's revenue model, alongside its subscription tiers. Plus, Pro, Business, and Enterprise subscribers will not see ads.
The format is deliberately contained: ads will appear at the bottom of answers, clearly labeled, tied to the current conversation topic. Users can dismiss any ad and flag why. Personalization can be turned off. OpenAI says it will not show ads near health, mental health, or political content, and will exclude users it detects or believes are under 18.
The announcement came paired with the global rollout of ChatGPT Go — now available in 171 countries — and an explicit statement of principles: ad responses will not influence ChatGPT's answers, user conversations will not be sold to advertisers, and OpenAI will always maintain an ad-free paid option.
The Infrastructure Behind the Pilot
The ad stack is not being built from scratch. On March 2, Criteo (NASDAQ: CRTO) announced it is the first ad technology partner integrating with the ChatGPT ad pilot in the US. Criteo connects more than 17,000 advertisers and activates over $4 billion in annual media spend — it's a commerce-focused platform, and the fit is intentional. OpenAI is not pitching ChatGPT as a brand awareness channel. It is pitching it as a discovery and purchase-intent channel.
The numbers Criteo is showing advertisers support that framing. Across a sample of 500 US retailers in February 2026, users referred from LLM platforms like ChatGPT converted at approximately 1.5 times the rate of other referral channels — though the sample is small and self-selected, drawn from Criteo clients already testing LLM traffic. That is a meaningful signal. Advertisers are being asked to commit between $50,000 and $100,000 in initial spend, and Criteo has been advising them that providing more variations of ad text and visuals increases frequency and improves performance.
The combination — high-intent conversational context plus a commerce-optimized ad network — is the bet OpenAI is making. The question is whether it holds up at scale across the full Free-tier population, which is far more diverse than early pilot participants.
Why Now
OpenAI's cost structure is no secret. Compute infrastructure for frontier model inference is expensive, and the company has been burning through capital as it scales. Revenue from ChatGPT Plus and Pro is strong but concentrated in a paying minority. The free tier is enormous — and until now, essentially monetization-free.
Ads represent a way to extract value from that user base without a subscription paywall. The Go tier is also relevant here: at $8/month with ads, it sits below the $20 Plus tier and gives users expanded access (messaging, image generation, file uploads, memory) without requiring full subscription spend. Ads effectively subsidize Go's economics.
OpenAI is also facing rising competitive pressure on cost. Meta has released powerful open-weight models at near-zero marginal cost to users. Google has embedded Gemini into products people already use. In that environment, keeping the free tier genuinely capable — and funding it — is a real strategic problem. Advertising is one answer.
The Contrast: Perplexity Moves the Other Direction
The timing is notable because Perplexity AI has explicitly moved the other direction. Business Insider reported February 18 that Perplexity executives told reporters at a roundtable the company is not exploring any ad deals at the moment, doubling down instead on subscriptions and enterprise sales. According to the same report, the company's ARR grew from $80 million at the end of 2024 to an estimated $200 million by February 2026, suggesting subscriptions alone can drive meaningful revenue at AI-native query volumes.
Perplexity CEO Aravind Srinivas has positioned the company toward high-powered professionals — finance, medicine, executives — and is ramping enterprise sales from a current team of five. The shift puts Perplexity more directly in competition with enterprise search tools like Glean.
That divergence matters. Perplexity and ChatGPT are not the same product — ChatGPT's free-tier user base is orders of magnitude larger and more general-purpose — but it does suggest there is not a single obvious answer to the AI monetization question. Perplexity is betting that users will pay to avoid ads in a search-like context. OpenAI is betting that its free tier is too large to leave on the table, and that conversational ad context is worth something to advertisers.
Both bets could be right for their respective scales.
What to Watch
The test is US-only for now, and explicitly framed as a testing phase. OpenAI said it will refine ad formats based on feedback. The real questions are: How does user retention respond when ads appear? Does the 1.5x conversion signal Criteo is observing hold as volume grows? And does advertiser demand at the $50-100K commitment level scale to support Free-tier economics?
The ad format itself — bottom-of-answer, contextually triggered — is deliberately non-intrusive compared to search advertising, where ads appear above organic results. If OpenAI can maintain that separation credibly over time, it may have found a format that works. If ads migrate upward in the answer, or if users perceive that ad relationships are influencing ChatGPT's responses, the trust cost could outweigh the revenue gain.
OpenAI's stated principle — "ads do not influence the answers ChatGPT gives you" — is the one that matters most. It is also the one that is hardest to verify from the outside.
Sources: OpenAI blog (openai.com/index/our-approach-to-advertising-and-expanding-access/); Criteo press release (March 2, 2026); Reuters (March 21, 2026); Business Insider, Charles Rollet (February 18, 2026).