OpenAI turns on cost-per-click ads inside ChatGPT

OpenAI has begun testing cost-per-click ads inside ChatGPT, with bids reportedly set at $3-$5 per click, a meaningful shift from its prior CPM-only model. The move should improve advertiser measurement and revenue potential, especially as ChatGPT ad CPMs have already fallen from $60 at launch to as low as $25 in some cases. OpenAI is also hiring its first advertising marketing science leader to build measurement, attribution and privacy-safe reporting infrastructure, signaling rapid commercialization of its ad business.

Analysis

OpenAI moving from CPM to CPC is the first real evidence that AI-native ad inventory is being forced into the same economic architecture as search. That matters because the valuation case for the whole category shifts from “novel placement” to “measurable intent capture,” which is the only framework that can support durable budget reallocation. The immediate winner is Google: OpenAI’s pricing will be benchmarked against search economics, and Google’s auction machine has a multi-year moat in conversion-quality optimization that a new entrant cannot replicate quickly.

The second-order effect is more interesting on Meta. If OpenAI clicks start looking like mid-funnel intent rather than pure awareness, the platform could siphon budgets from social prospecting before it takes meaningful share from search. That would pressure Meta’s upper-funnel performance narrative more than its core retention/commerce stack, especially if advertisers begin treating ChatGPT as a “high-intent assistant” rather than a discovery feed. In contrast, OpenAI’s near-term monetization is still likely to be constrained by traffic quality variance and sparse conversion data, which means CPC adoption can increase revenue without necessarily proving superior ROI.

The market is likely underestimating how fast measurement becomes the gating function. Whoever wins the measurement layer owns the budget map; that’s why the marketing science hire is strategically more important than the ad product itself. But there is also a real privacy/cost risk: if conversion tracking remains too abstracted or clean-room dependent, advertisers may test budgets but fail to scale them, capping take-rate. The key timeline is months, not days — the next 2-3 quarters should determine whether this becomes a meaningful search-adjacent spend category or just another experimental placement.

Key Decisions for Investors

• Long GOOGL vs short META for 1-3 months: if AI-native ads become a measurable intent channel, Google’s proven auction/measurement stack should defend share better than Meta’s prospecting-heavy mix; target 1.5:1 risk/reward with a tight stop if ChatGPT ad adoption remains niche.
• Buy GOOGL calls 3-6 months out: the market may underprice the defensive value of Google Search if OpenAI clicks prove low-quality or hard to attribute; this is a low-cost hedge against AI ad channel optimism.
• Short a basket of ad-tech proxies with high dependency on incremental experimentation budgets over 6-12 weeks: the risk is that “new channel testing” shifts spend into direct platform buying, compressing intermediary take-rate; keep size modest because the move is thesis-driven rather than immediately measurable.
• If available, buy medium-dated OTM META puts into any rally: the asymmetric risk is that performance advertisers start reallocating prospecting dollars to AI assistants before attribution proves clean, pressuring near-term sentiment more than earnings.
• Stay flat UBER and NFLX for now: the article is signaling an ad-measurement arms race, not a direct P&L driver; any move here would be noise relative to the uncertainty in OpenAI’s conversion economics.