Eric Seufert: OpenAI running ads is inevitable — and it will be the most important new ads platform in a decade

Summary

Eric Seufert, independent analyst and author of Freemium Economics, argues that OpenAI running advertising is not a question of if but when, drawing a direct parallel to Netflix's trajectory before launching its ad-supported tier in November 2022. The tell, per Seufert, was Spencer Neumann's March 2022 statement at the Morgan Stanley TMT conference that Netflix had no "religion against advertising" — a line that preceded the ad product launch by eight months. Sam Altman's recent appearance on Ben Thompson's podcast produced an almost identical rhetorical move: Altman praised Instagram's product discovery experience while distancing himself from the word "ads," a distinction Seufert dismisses as semantic cover for an already-developing strategy.

The Unit Economics Case

Financial Times data cited in the conversation puts OpenAI at 800 million users, with only 5% paying, yielding roughly 40 million paying users. Of the reported $13 billion ARR, approximately 70% is consumer-derived, with the remainder coming from enterprise API access. That produces a global ARPU of $11.38, slightly below Meta's $13.65 and well above Snap's $2.87 and Pinterest's $1.64. The remaining 95% of users generate inference costs with no offsetting revenue, making the free tier economically unsustainable without an additional monetization layer. Advertising in the free tier is the logical solution.

Format, Attribution, and the Trust Firewall

Seufert pushes back on the conventional concern that advertising inside a chat interface is technically intractable. Attribution, he argues, is a solved problem — no different in principle from tracking conversions on a social media ad. The harder challenge is format: any ad unit that appears to influence the content of a response risks destroying user trust permanently. Once users believe OpenAI's answers are commercially tilted, that credibility cannot be recovered. The constraint does not apply to surface areas like a future app directory or a news feed product, where sponsored placements feel native and carry no implied editorial compromise.

Critically, Seufert argues ads in ChatGPT do not need to be contextually matched to query content. Meta's model — behaviorally targeted against purchase history rather than the adjacent content — applies directly. A user researching Roman history can be served a paper towel ad if prior conversion data suggests purchase intent. The chat content is largely irrelevant to ad targeting.

Instant Checkout as Bootstrap, Not Destination

Seufert is skeptical that OpenAI's instant checkout feature is itself the monetization endpoint. The current model, a flat percentage fee on transaction value, is economically suboptimal because it does not combine relevance with bid price. An auction-based system incorporating advertiser willingness to pay would generate higher yield. His read is that instant checkout is primarily a mechanism to accumulate conversion data on users, which then powers future ad targeting — not a standalone commerce product.

Talent and Infrastructure Are Already in Place

A common objection — that OpenAI lacks the team to build a competitive ad platform — Seufert dismisses. The company has been recruiting heavily from Meta, including engineers who built ads-ranking systems. The institutional knowledge for building performance advertising infrastructure is already inside the organization.

Broader Market Implications

Seufert frames an OpenAI ad platform as potentially the most significant new advertising surface in a decade, with compounding economic effects similar to what Facebook produced for direct-to-consumer e-commerce. Advertisers who enter early on an underpriced platform and reinvest returns into subsequent spend create self-reinforcing growth loops. The same compounding dynamic applies to Meta's own back-end improvements — referencing projects Andromeda and Lattice — where incremental click-through rate gains of 2-3% at Meta's scale translate into substantial absolute revenue and grow with each reinvestment cycle. For app developers and small businesses, a well-functioning OpenAI ad platform represents a potential step-change acquisition channel, analogous to early Facebook ads for the DTC category.