OpenAI's Secret Second Business Just Went Public — and It Could Be Worth More Than ChatGPT

OpenAI built its business on the premise that AI should be a subscription, not an advertising platform. Sam Altman said publicly that advertising misaligns incentives , that an AI trying to keep users engaged will optimize for engagement, not truth. That principled stance lasted exactly until OpenAI needed a second revenue stream that could scale without adding marginal compute costs. On May 5, 2026, OpenAI opened its self-serve ChatGPT Ads Manager to small businesses across the United States, dropping the $50,000 minimum spend requirement that had applied during its agency-partner pilot phase and setting its sights on $2.5 billion in advertising revenue this year and, audaciously, $100 billion by 2030.

What Actually Happened

OpenAI launched a self-serve advertising platform that makes ChatGPT advertising accessible to businesses of any size for the first time. The platform, initially available in the United States, allows advertisers to register directly, set budgets and bids, upload creative assets, and manage campaigns inside OpenAI's portal , without needing a partner agency or a six-figure minimum commitment. OpenAI simultaneously introduced cost-per-click (CPC) bidding alongside existing cost-per-thousand-impressions (CPM) buying, giving advertisers the ability to align spending more directly with actions users take after seeing an ad. The self-serve Ads Manager launched in beta and is rolling out progressively to US advertisers.

Agency partnerships remain operational in parallel: Dentsu, Omnicom, Publicis, and WPP have direct buying relationships with OpenAI, while ad technology companies including Adobe, Criteo, Kargo, Pacvue, and StackAdapt are integrated into the Ads Manager platform , giving enterprise advertisers the same programmatic and performance marketing tools they use on Google and Meta. ChatGPT's current user base exceeds 800 million monthly active users, engaging with the platform through sessions that are, by their nature, expressions of high-intent information seeking. Unlike social media platforms where users are scrolling, ChatGPT users are actively researching, deciding, or completing specific tasks. That behavioral difference is the core of OpenAI's advertising proposition.

Why This Matters More Than People Think

OpenAI's subscription revenue , primarily from ChatGPT Plus at $20 per month, ChatGPT Pro at $200 per month, and enterprise contracts , is growing rapidly but faces structural constraints. Subscription revenue scales with customer acquisition efficiency and churn management. Advertising revenue scales with attention, and OpenAI has more high-quality, decision-oriented attention per user than almost any other platform on earth. The average ChatGPT session involves a user actively researching, comparing options, or making a decision. This is the same high-intent engagement for which Google charges premium advertising rates of $10 30 CPM on Search , three to five times the rates commanded by social platforms. If OpenAI can capture even 30 40% of that intent premium with its conversational ad format, the revenue mathematics at 800 million MAU justify the $2.5 billion annual target.

The $100 billion by 2030 projection is the number that requires the hardest scrutiny. Google's advertising business , the most successful in history, built over 25 years , generates approximately $250 billion annually. For OpenAI to reach $100 billion in ad revenue by 2030 would require it to become one of the top three advertising platforms on earth in four years, from a standing start. That number is either a genuine strategic vision supported by internal user engagement data that the public has not seen, or it is a deliberate signal to advertisers and investors that OpenAI is serious about advertising as a core business line. Even discounted by 80%, a $20 billion advertising business by 2030 would be large enough to fund OpenAI's compute requirements without necessitating another equity raise , which is perhaps the actual goal.

The Competitive Landscape

Google's response to ChatGPT advertising will define the competitive dynamic for the next 24 months more than any model benchmark comparison. Google Search's advertising moat was always the monopoly on high-intent queries , users searching "best electric car 2026" are buyers, not browsers, and Google has charged accordingly. ChatGPT is now capturing millions of those same queries and is about to monetize them at scale. Google's response has been to integrate AI answers more aggressively into Search via AI Overviews, but each AI answer that replaces a set of organic search results also displaces the advertisements that surrounded those results. Google is in the structurally uncomfortable position of cannibalizing its own advertising revenue to defend against a competitor that is cannibalizing it from outside. The company that built the internet's most profitable business is now facing a user behavior shift it cannot stop without destroying its own economics.

Meta's exposure is different but equally significant. Meta built its advertising empire on demographic targeting, behavioral data, and lookalike audiences. ChatGPT's advertising proposition is fundamentally orthogonal , intent-based rather than interest-based. A ChatGPT user asking "compare the iPhone 17 and Samsung Galaxy X for photography" is expressing purchase intent in natural language that is richer, more specific, and more action-oriented than any behavioral signal Meta can infer from scroll patterns and engagement history. For the first time in the digital advertising era, a competitor to Google and Meta is entering the market without relying on tracking pixels, cookies, demographic profiles, or behavioral surveillance. It relies entirely on what people say , and what people say to ChatGPT is, by definition, exactly what they want to know or buy right now.

Hidden Insight: Advertising Changes the Alignment Problem Forever

Sam Altman's original objection to advertising was not primarily commercial , it was epistemic. An AI system that shows users ads will inevitably face pressure to optimize responses in ways that keep conversations going rather than resolving them efficiently. The fastest path to resolution and the highest-revenue path are in fundamental tension. A user asking "should I buy a Tesla or a Rivian?" gets the most accurate answer when the model synthesizes available data without commercial considerations. That same session becomes a monetization opportunity once advertising enters the equation , and the optimal monetization strategy is not the same as the optimal truth-seeking strategy. OpenAI has not published any policy explaining how advertising considerations factor into answer selection, how sponsored content is distinguished from AI-generated analysis, or what safeguards prevent commercial interests from shaping GPT-5.5's responses in monetized sessions.

There is a deeper structural concern that goes beyond the immediate product question. OpenAI's $852 billion private valuation rests in significant part on a user trust premium , the perception that ChatGPT gives users the best available answer rather than the most profitable one. Google's advertising relationship has always operated within a framework of transparent separation: sponsored links are labeled, organic results are distinct, and users have a cognitive model for how the system works. ChatGPT's conversational format makes that separation genuinely difficult to maintain. When a response flows naturally from analysis to recommendation, the boundary between an AI judgment and a paid placement can dissolve in ways that are not legible to users. If users begin to perceive , rightly or wrongly , that advertising influences the content of AI responses rather than appearing alongside them as clearly labeled units, the trust premium collapses.

The historical parallel that is directly relevant here is Facebook's transformation between 2009 and 2014. Facebook launched advertising on a platform that users experienced as personal, authentic, and genuinely useful. Early advertising was sparse, clearly labeled, and low-density. The revenue imperative gradually increased ad density, shifted algorithmic rankings toward content that drove engagement rather than content that was accurate or meaningful, and commoditized the quality of the user experience in service of monetization. By 2016, the authenticity and utility that had made Facebook the dominant social platform of a generation had been systematically degraded. OpenAI is at an inflection point structurally identical to Facebook circa 2012: it has a product that users trust for its quality, and it is introducing a revenue model whose incentive structure is partially misaligned with that quality. The question is whether OpenAI manages that tension more skillfully than Facebook did , or whether it follows the same arc.

What to Watch Next

The 30-day indicator: OpenAI's next revenue disclosure or funding announcement that references advertising as a distinct business line. If advertising contribution appears separately from subscription and API revenue in any investor communication before the end of Q2, it confirms that OpenAI is already treating advertising as a material and trackable revenue stream rather than an experimental pilot. Watch simultaneously for user sentiment signals in the first three to five weeks after the self-serve launch: App Store review score trends, social media discussion of "ChatGPT ads," and any coverage of users encountering labeled sponsored content in ChatGPT sessions for the first time. The initial reaction will be informative about how aggressively OpenAI can scale ad density without triggering user backlash.

The 90-day indicator to watch closely: Google's Search advertising revenue in its Q2 2026 earnings, expected in late July 2026. If Google reports deceleration in high-intent commercial query advertising despite broader digital advertising market stability, it will be the first quantitative evidence that ChatGPT is beginning to divert commercial search budgets at meaningful scale. The more structurally significant signal: when major media agencies begin reporting ChatGPT Ads as a distinct line item in their clients' media mix models , not a sub-category of "other digital" or "AI-native channels" , it signals that ChatGPT advertising has achieved category legitimacy in the planning process. That moment will reshape OpenAI's valuation narrative more than any language model benchmark, because it will mean that the advertising business is real, scaled, and competitively positioned against platforms that have never faced a genuine threat to their share of commercial intent.

OpenAI built the world's most trusted AI by refusing to optimize for money , and is now betting it can optimize for money without breaking the thing that made it trustworthy.

---