OpenAI Wants to Sell You Ads — and It Just Poached Meta’s Playbook to Do It

OpenAI has hired a senior advertising executive from Meta, signaling that the maker of ChatGPT is preparing to build an ad-supported business that could fundamentally reshape how artificial intelligence companies generate revenue. The move, first reported by The Information, marks a concrete step toward a strategy that CEO Sam Altman has publicly resisted for years — and one that could pit OpenAI directly against Google, Meta, and the rest of the digital advertising establishment.

The hire is Shivakumar Venkataraman, a longtime Google executive who more recently worked at Meta on advertising products. His arrival at OpenAI isn’t just another talent acquisition in the AI arms race. It’s a declaration of intent.

For a company that has burned through billions of dollars in compute costs while charging consumers $20 a month for ChatGPT Plus, the economics of generosity have limits. OpenAI reportedly lost around $5 billion in 2024, according to estimates compiled by The New York Times. Revenue has grown fast — the company said it expects to hit $3.4 billion in annualized revenue — but so have the infrastructure bills. Training frontier models costs hundreds of millions of dollars per run, and inference costs scale with every new user. Advertising, which has underwritten the entire modern internet, offers a path to revenue that doesn’t depend on converting free users into paying subscribers.

Altman himself has been cautious about ads. As recently as 2023, he told audiences that he preferred a subscription model, suggesting that advertising incentives could compromise the quality and trustworthiness of AI outputs. But business realities have a way of eroding philosophical positions. And OpenAI, which is in the process of converting from a capped-profit structure to a full for-profit corporation, now has investors expecting returns on more than $13 billion in committed capital.

The Venkataraman hire fits a pattern. OpenAI has been quietly assembling the infrastructure for an advertising operation over the past several months. In late 2024, the company began testing a search product — ChatGPT Search — that directly competes with Google’s core business. Search advertising is a $300 billion global market, and OpenAI’s conversational interface offers a new format for delivering sponsored results. Rather than ten blue links with ads scattered above and among them, ChatGPT could integrate product recommendations and brand mentions directly into its natural-language answers.

That prospect excites advertisers. It terrifies publishers.

The advertising industry has been watching AI’s commercial evolution with a mixture of anticipation and anxiety. Major holding companies like WPP, Publicis, and Omnicom have all announced AI partnerships and internal tools, but the question of where AI-native ad dollars will flow remains open. If OpenAI builds an ad platform inside ChatGPT, it could siphon spend from Google and Meta — the same companies OpenAI has been raiding for talent.

Google, for its part, is already adapting. Its AI Overviews feature, which generates summary answers at the top of search results, now includes ad placements. Alphabet CEO Sundar Pichai told investors on a recent earnings call that AI Overviews are showing “strong monetization signals” and that ads within AI-generated answers perform comparably to traditional search ads. Google has decades of advertising infrastructure, relationships with millions of businesses, and a self-serve platform that small advertisers depend on. OpenAI has none of that. Yet.

But what OpenAI does have is attention. ChatGPT reached 100 million weekly active users faster than almost any consumer product in history. The company’s mobile app has consistently ranked among the top downloads in both the App Store and Google Play. And the nature of ChatGPT interactions — longer, more conversational, often revealing explicit purchase intent — could make its user data extraordinarily valuable for ad targeting. When someone asks ChatGPT to compare running shoes or plan a vacation, the commercial signal is unmistakable.

Privacy advocates are already raising flags. Unlike traditional search queries, which are relatively brief and discrete, ChatGPT conversations can span dozens of exchanges and touch on sensitive personal topics — health concerns, financial problems, relationship issues. Building an advertising model on top of that data would require careful handling, both legally and reputationally. OpenAI’s privacy policy already allows it to use conversation data for model improvement, but using it for ad targeting would represent a significant expansion that regulators in the EU, under the Digital Services Act and GDPR, would scrutinize closely.

The competitive dynamics here are layered. Meta, which lost Venkataraman to OpenAI, has been investing heavily in its own AI capabilities. Mark Zuckerberg has positioned Meta’s Llama models as the open-source counterweight to OpenAI’s proprietary approach, and Meta’s advertising machine — the most sophisticated in the world after Google’s — already uses AI extensively for targeting, creative generation, and bid optimization. Losing a senior ads executive to a potential competitor is a minor personnel matter for a company of Meta’s size. But it’s symbolically significant. The talent flow from Big Tech to AI startups has been well documented, but when the hires are specifically in ad sales and ad tech, it signals that the startup is getting serious about the business model that actually funds the internet.

Microsoft, OpenAI’s largest investor and partner, occupies an awkward position in all of this. Bing, which powers ChatGPT’s search functionality through a partnership deal, already serves ads. Microsoft takes a cut of revenue generated through Bing-powered search within ChatGPT. If OpenAI builds its own first-party ad platform, it could reduce its dependence on Microsoft’s infrastructure — and Microsoft’s revenue share. The two companies have already been renegotiating aspects of their partnership as OpenAI’s valuation and ambitions have grown. An independent OpenAI ad business would add another tension point to an already complex relationship, as Bloomberg has previously reported.

There’s also the question of format. Traditional display ads — banners, pop-ups, pre-roll video — don’t translate naturally to a chat interface. The most likely initial format would be sponsored recommendations: a user asks for the best noise-canceling headphones, and ChatGPT’s answer includes a paid placement from Sony or Bose, perhaps labeled as “sponsored” in the same way Google marks its paid search results. This is essentially native advertising, a format that has proven effective on platforms like Instagram and TikTok but has also drawn criticism for blurring the line between editorial content and paid promotion.

In a conversational AI context, that blurring could be even more pronounced. Users tend to trust ChatGPT’s responses as authoritative — studies have shown that people often treat AI-generated answers with the same or greater confidence as expert human opinions. If a sponsored result is woven into a ChatGPT response, users may not distinguish it from an organic recommendation. The Federal Trade Commission has existing guidelines on disclosure of paid endorsements, but those guidelines were written for human influencers and traditional media. Applying them to AI-generated content is uncharted regulatory territory.

OpenAI’s competitors in the AI space are watching closely. Anthropic, maker of the Claude chatbot, has explicitly positioned itself as a safety-focused company and has not indicated any interest in advertising. Perplexity AI, which operates an AI-powered search engine, has already begun experimenting with ads — it introduced sponsored questions and branded answers in late 2024, drawing both advertiser interest and user backlash. Google’s Gemini and Apple’s Siri integration with AI features represent other vectors of competition, though neither has announced standalone ad products tied to their AI assistants.

The financial logic for OpenAI is straightforward. Subscription revenue has a ceiling determined by the number of users willing to pay $20 or $200 a month. Advertising revenue scales with usage and engagement, metrics where ChatGPT already excels. If OpenAI can convert even a fraction of its free-tier users into advertising impressions, the revenue potential is enormous. JPMorgan analysts estimated in a recent note that an AI-powered search ad market could be worth $50 billion by 2028. OpenAI wouldn’t capture all of that. But it doesn’t need to.

The hire of Venkataraman suggests the timeline is accelerating. Building an ad sales operation requires not just technology but relationships — with agencies, with brands, with measurement partners. It requires a sales force, a self-serve platform for smaller advertisers, and sophisticated auction mechanics to price ad inventory. These aren’t things you build overnight. But they are things you start building when you’ve made a strategic decision.

And OpenAI appears to have made that decision.

What remains unclear is how Altman will frame this pivot to users and to the public. He has cultivated an image of OpenAI as a mission-driven organization focused on developing artificial general intelligence for the benefit of humanity. Advertising, with its associations with surveillance capitalism and attention manipulation, doesn’t fit neatly into that narrative. Altman will likely argue that ad revenue enables OpenAI to keep its most powerful tools free and accessible — the same argument Google has made for two decades. Whether users and regulators accept that framing will depend on execution and transparency.

The broader implications extend beyond OpenAI. If the most prominent AI company in the world embraces advertising, it sets a precedent for the entire industry. Startups building AI assistants, AI search tools, and AI productivity apps will face pressure from investors to adopt similar models. The ad-supported internet — which many technologists have criticized as a structural failure — would extend its logic into the AI era. The alternative models — subscriptions, API licensing, enterprise contracts — would remain viable but potentially secondary.

So here’s the picture: OpenAI, a company valued at roughly $150 billion, is building an advertising business from scratch. It has the users, the engagement, and now the executive talent. What it doesn’t yet have is the product, the advertiser relationships, or the regulatory clarity. Those will come. The question isn’t whether OpenAI will sell ads. It’s how big that business becomes — and what it costs the company’s reputation to build it.