OpenAI is pricing ads like prime-time TV. But is it selling belief over proof?

OpenAI is betting it can charge premium prices for advertising because it believes ChatGPT is not just another digital screen, but a fundamentally different kind of commercial environment.

At the core of its pricing argument is the idea that ads placed inside ChatGPT resemble high-end television more than social media or search.

The company is charging roughly $60 per 1,000 impressions, a level comparable to live US sports broadcasts and other premium video inventory, because it is selling presence in a high-attention environment rather than provable outcomes

So, even without the detailed metrics that advertisers have come to expect from platforms like Meta and Google, the company thinks scarcity, context and momentum give it unusual leverage in the market.

Here’s how that logic works, and why it may only hold for so long.

Scarcity beats scale, at least at the start

The first pillar of OpenAI’s pricing logic is artificial scarcity.

Most ad platforms drown buyers in inventory. Social feeds, pre-rolls, stories and banners are infinite. ChatGPT, by contrast, has deliberately constrained ad real estate. Ads appear only below certain responses, not threaded through conversations, and not in paid tiers.

That design choice mirrors what Apple did with Search Ads inside the App Store. Apple charged far higher rates than mobile ad networks long before it offered deep performance data. The pitch was this: there are very few places to appear, users are already in a decision-making mindset, and Apple controls the environment tightly. Advertisers paid up because access itself was scarce.

OpenAI is attempting the same move. By limiting placements, it creates competition for visibility and can anchor pricing to premium benchmarks rather than average digital CPMs.

Contextual intent as a proxy for targeting

The second pillar is contextual clarity.

Unlike social platforms, where ads chase inferred interests based on past behaviour, ChatGPT operates on explicit, real-time intent expressed in natural language. When a user asks “what’s the best budgeting app for freelancers” or “alternatives to meal kits,” the context is unambiguous.

This mirrors how advertising worked for decades in high-end print. A hedge fund ad in The Economist or a watch ad in The Financial Times carried no clickthrough data, but advertisers paid because the surrounding content did the targeting for them.

OpenAI believes conversational context can replace user-level targeting, especially as privacy rules make behavioural tracking harder. That belief allows it to argue that it is selling clarity of mindset, not just eyeballs.

Borrowing from terminal economics, not social feeds

A more revealing comparison is not Meta Platforms or Google, but Bloomberg.

Bloomberg terminals command enormous subscription fees because users treat them as authoritative, high-trust environments where decisions are made. When Bloomberg experimented with sponsored content and data partnerships, it priced them far above typical digital ads because the audience was small, elite and decisional.

OpenAI is implicitly positioning ChatGPT as a decision-support system, not a content feed. If advertisers believe users are actively reasoning rather than scrolling, the value per impression increases, even without proof of downstream action.

That framing also explains why OpenAI is comfortable starting with impression-level reporting. In terminal-style economics, presence near the decision is the product.

Brand safety and tone control as premium features

Another reason OpenAI thinks it can charge more is risk reduction.

Advertisers have grown wary of platforms where their ads can appear next to polarising content, misinformation or low-quality engagement. OpenAI tightly controls tone, language and subject matter. Ads appear in a neutral, utilitarian interface with no comment threads, no virality and no outrage dynamics.

Netflix used a similar argument when it launched its ad tier. It priced inventory aggressively, despite limited targeting at launch, because it could guarantee a controlled, brand-safe environment with close attention. Buyers accepted weaker metrics in exchange for lower reputational risk.

OpenAI is making the same trade-off explicit: fewer numbers, but fewer surprises.

Early pricing as a signal, not a reflection of value

There is also a strategic reason for starting high: price anchoring.

By pegging early CPMs to premium video rather than digital auctions, OpenAI sets expectations with agencies and brands about where it wants to sit in the media hierarchy. Dropping prices later is easier than trying to raise them after being labelled “experimental tech inventory.”

This is a classic enterprise software tactic applied to advertising. The initial price is as much about signalling ambition as monetisation.

Why this logic breaks without outcomes

All of this explains why OpenAI thinks it can charge so much. It does not explain why advertisers will tolerate it indefinitely.

Every comparable example eventually added proof:

• Apple layered in conversion data.
• Netflix expanded targeting and attribution.
• Bloomberg tied sponsorships to lead generation and events.

If OpenAI fails to show how ads connect to actions, budgets will stall at test levels. That is why its checkout experiments matter so much. Commission-based commerce would give OpenAI something Google and Meta built years ago: a closed loop.

Until then, OpenAI is effectively asking advertisers to buy belief.

That works in year one. It does not scale to a $10 billion ad business without evidence.