ChatGPT Advertising CPMs Drop Sharply in Pilot

Advertising rates in the `ChatGPT` pilot are trending downward. Initial pilot CPMs were advertised at $60; nine weeks in, buyers report averages closer to $45 and many spots in the $25 to $35 range, with some reports as low as $15. The minimum entry threshold has fallen from initial reports of $200K-250K to roughly $50K in some cases, expanding bidder participation. "It feels like everything is coming down, preparing for a wider auction accessibility to fit [a] global rollout," said Ashley Fletcher, CMO of Adthena.

The current pilot uses a hybrid sales approach: direct deals and limited programmatic access via partners rather than a full public auction. Early supply and demand drivers include inventory growth as OpenAI increases ad-eligible impressions on free and Go tiers, and more buyers entering as minimum spends fall. Key technical and measurement signals practitioners should note:

• •Partner stack: inventory is already available through Criteo for commerce/retail demand; OpenAI has held talks with The Trade Desk to expand programmatic reach.
• •Performance signal: referral data cited by OpenAI shows LLM-driven referrals converting at roughly 1.5x the rate of other channels, a potential justification for higher CPMs.
• •Pricing mechanics: without a full live auction, reported CPM variance reflects negotiated deals, buyer composition, and early supply scale rather than open-market floor pricing.

This is a commercial inflection point for OpenAI. Ads convert monetization for the majority of users who are not subscribers; only about 5% of users pay, so ad revenue is a lever to scale consumer revenue toward projections in the billions. The pendulum of ad pricing is following historical precedents: Netflix launched ads at high CPMs and saw rates fall to the $20-30 range as inventory scaled. The comparison matters because ChatGPT lacks a clearly understood value exchange the way streaming content provided; users historically perceive ChatGPT responses as unmediated, which raises user-experience risk as ads become more visible.

Advertisers and holding groups involved or reported as early buyers include:

• •WPP, Omnicom, Dentsu
• •Brands such as Target, Ford, Mrs. Meyer's, Adobe, Best Buy, AT&T, and Expedia

These buyers indicate both brand and direct-response appetite, and the presence of DSPs would open programmatic budgets that many agencies currently keep in their normal ponds.

Attention to three things will determine how sustainable current prices are: inventory growth and the timetable for a live auction; measurement and attribution for LLM-driven conversions; and user reaction as ads scale across free experiences. If OpenAI brings The Trade Desk or similar DSP integrations live, CPMs are likely to compress further as more demand is layered onto growing supply. Conversely, if conversion premiums hold and OpenAI controls ad placement and measurement tightly, sustained higher CPMs could be defensible.

For ad-tech engineers, media buyers, and product leaders, the takeaway is tactical: expect rapidly evolving floor prices, prioritize integration with OpenAI's measurement endpoints, and model for both a high-CPM early phase and a likely multi-quarter compression as inventory matures. For monetization strategists, balancing revenue goals against user trust risk will be the critical product trade-off.