There are growing indications that ads are coming to ChatGPT, marking a crucial shift in how one of the world’s most widely used AIs might pay for its free tier. Newly discovered app code and a leaked in-the-wild placement appear to show that OpenAI is at least testing ads within the chatbot experience.
What Users Are Seeing and What the App Code Reveals
A beta build of the ChatGPT Android app (version 1.2025.329) got attention this weekend when an X user noticed strings referencing “search ad,” “ad features,” and “bazaar content.” The language is written in the style of scaffolding for native positions connected to queries and responses, rather than venerable banner slots. The results have yet to be confirmed across all installs, but the language is specific enough that it seems more than a placeholder experiment.
Not long after that, another user shared a screengrab of what seems to be an in-stream ad unit appearing below a ChatGPT answer: an image-led card featuring the CTA “Find a fitness class” and brand prompt “Connect Peloton.” If authentic, the positioning is reminiscent of conversational “sponsored cards” that integrate with assistant responses but are also visually set apart.
OpenAI’s CEO Sam Altman has previously said on the company podcast that ads are one form of monetization on the table, as they continue to refine how they might be implemented. So the philosophical barrier is surmounted; now it’s just work about product fit and policy.
Why Ads Make Sense for OpenAI Now, Given Costs and Scale
Inference isn’t cheap. Storing and training of large models at consumer scale requires huge amounts of compute, networking, and memory bandwidth. It is estimated by analysts that advanced model queries can be up to a few cents per session, not an insignificant cost when serving tens of millions of free users. OpenAI does have diversified revenue including ChatGPT Plus, Teams, and Enterprise offerings; reports in the press have said that it’s on a billion-dollar run rate. Even still, monetizing the free tier can significantly mitigate costs and even out margins as usage scales.
There’s industry precedent. Microsoft has been displaying ads on Copilot and its Bing chat experiences since 2023, such as contextual links and shopping units. Google unveiled and experimented with ads in AI Overviews in Search, which itself extends the company’s auction and safety layers to generative summaries. A few AI search start-ups have tested sponsored results mixed into conversational responses. For a consumer assistant created to act as a discovery tool, ads are a logical — though perhaps sensitive — extension.
What Ads May Look Like Inside ChatGPT on Mobile and Web
The code suggests “search” integrations, perhaps along the lines of formats paired to intent-rich prompts: product discovery, local services, travel planning, or software tools. Anticipate small, prominently labeled cards attached beneath responses, featuring imagery and brief copy along with a brand or CTA button. And, certainly, sponsored links within cited sources or tool recommendations are also a possibility — but here you’d need additional guardrails to keep from confusing ads with reference material.
In mobile, placements would likely resemble native app implementations: lightweight, quick to load, and limited to one or two slots per interaction. The reference to “features” is most likely a reference to the control plane for personalization and frequency capping, which are vital for context relevance and for maintaining conversation flow.
The Trade-offs: Transparency and Trust When Ads Appear
Embedding ads in AI answers is a tricky proposition. Clear labeling and distinction from organic model output may also be essential to user trust. The Federal Trade Commission has repeatedly said advertising in immersive or AI experiences should be clear and not presented as neutral information. In the European Union, the Digital Services Act would require ad transparency about who paid for an ad and why the user was being shown it.
There are product risks, too. Ads can nudge an assistant’s perceived neutrality, particularly when recommendations fall in domains like medical, legal, or financial. “Tight filtering, category exclusion, and opt-outs for sensitive queries will be table stakes.” If history is any guide — and OpenAI has reportedly followed traditional playbooks on the way to making this decision — free tier users could very well start seeing ads while paid subscribers keep it ad-free as a perk.
What to Watch Next as Ad Tests Expand Across Platforms
Key things to keep an eye on:
- Whether ad tests spread beyond Android and into iOS and the web
- How prominently ads are labeled
- Whether sponsored content remains confined to commercial intent queries
- Announcements for advertisers describing formats, measurement, and brand safety standards
If OpenAI makes use of buy-side infrastructure already in place — search-style auctions, for example, or partnerships with big ad companies — the rate of rollout could become quite rapid.
The business logic is simple: ads can underwrite free access and spur growth. The difficulty is execution — keeping the assistant useful and trusted while adding a new revenue source. Early testing suggests OpenAI is poised to strike that balance, and for many ChatGPT users the era of ad-free interaction may soon be over.