OpenAI launches advertising platform, a wealthy business selling to the poor

Title: OpenAI Launches an Advertising Platform: A Wealthy Business Selling to the Poor

Author: Dongcha Beating

Source:

Reposted from: Mars Finance

Sam Altman once described advertising as ChatGPT’s “last resort.”

For a long time, that line functioned as restraint. OpenAI continued to package itself as a research company, an infrastructure company, and a company trying to bring AI capabilities to everyone. Advertising—one of the most familiar old-internet monetization methods—was treated as a backup plan.

But getting approval for the advertising plan came quickly.

On May 5, OpenAI launched the self-serve Ads Manager platform, enabling advertisers to place ads on ChatGPT directly or through agencies such as Dentsu, Omnicom, Publicis, and WPP. Less than three months after the first ad pilot was launched on February 9.

The platform is still in testing, but the direction is already clear: ChatGPT is no longer just a conversational product—it’s also becoming ad inventory. OpenAI’s goal is to achieve $2.5 billion in ad revenue in 2026, and push ad revenue to $100 billion by 2030.

ChatGPT, with a user base of 900 million, has found that the free route is getting harder and harder.

Annual losses of billions, relying on ads to recover

OpenAI is growing so fast that it’s hard for traditional internet companies to find a reference point.

But it burns money just as fast.

HSBC analysts estimated that by the end of 2025, OpenAI could still face a funding gap of $207 billion before 2030. Its cloud and AI infrastructure spending in the second half of 2025 through 2030 could reach $792 billion, and its long-term compute commitments by 2033 might approach $1.4 trillion.

These numbers explain why they are moving to build an advertising business.

Subscription revenue proves users are willing to pay, but it’s difficult to cover the reasoning costs of all free users. Enterprise APIs can contribute cash flow, but they face price wars and model convergence. Capital financing can keep things going, but it dilutes equity and pushes higher valuation pressure back inside the company.

Advertising is the fastest source of non-dilutive revenue. It doesn’t require free users to pay, doesn’t need to re-educate the market, and is easier to explain to investors.

According to Reuters, OpenAI’s ad pilot annualized revenue exceeded $100 million within six weeks. Ads are shown only to users on free and Go plans, do not affect ChatGPT’s generated responses, and will not share user data with marketers.

Setting aside user privacy for now, there’s a more fundamental question behind this strategy.

Ads are sold to free users, but what advertisers want are paying users

ChatGPT has 900 million weekly active users, about 50 million paid subscriptions, and a free-to-paid conversion rate of less than 6%. Because ads are shown only to free users, all of OpenAI’s ad inventory comes from the 94% who are unwilling to pay.

The problem is that advertisers willing to invest $50k or more usually aren’t selling products directly to individual consumers. Enterprise software, SaaS tools, and B2B services—decision-makers in these high-ticket categories—are precisely the most likely to be ChatGPT’s paid users. They spend $20 to $200 per month to buy stronger models and larger context windows, and ads will never appear on their screens.

Beyond audience mismatch, there’s a deeper issue: even if ads successfully reach free users, how much ad value can their usage scenarios actually support?

High intent doesn’t equal high conversion

OpenAI’s ad narrative is built on a core assumption: ChatGPT users enter the chat with genuine intent, so ad reach in high-intent scenarios is worth a higher price.

That assumption is only half right.

Over the past twenty years, what brands most wanted to capture was the search box—because the search box represents intent. Searching for a hotel suggests the user may be about to book. Searching for business tax software suggests procurement is underway. Searching for the best noise-canceling headphones suggests the user is already at the doorstep of a purchasing decision.

Google built an advertising empire on this. After ChatGPT arrived, users directly handed the decision-making process to AI. For advertisers, that’s more enticing—and more terrifying—than search ads. The enticing part is that ChatGPT sees an entire segment of demand: it doesn’t just know what users want to buy, it knows why they want to buy it. The terrifying part is that if AI gives the answer directly, users might not even look at the search results page.

But “help me buy a pair of running shoes” and “help me write an email” are two completely different intents. The first is a consumption scenario; the second is a productivity scenario. In everyday use of ChatGPT, the latter accounts for far more than the former. Users come here to write, translate, code, draft plans, and sort out their feelings—high frequency—but it doesn’t naturally map to product purchases.

This directly suppresses ad performance metrics. Advertisers are willing to pay premium prices for highly certain purchase intent. Google search ads are expensive because users often enter the search box with clear intent to buy, compare, reserve, or place an order. Meta ads are a bit cheaper, but Meta has social profiles and massive conversion data, allowing algorithms to repeatedly filter low-intent users into potential consumers.

ChatGPT sits in the middle. It’s more like an entry point for needs than social media, yet harder than search when it comes to judging commercial intent. It’s more private than search, but harder to attribute. It can help solve user problems, but it doesn’t necessarily generate ad clicks.

That’s also why OpenAI is moving from CPM (pay per impression) to CPC (pay per click)—not just a product upgrade, because advertisers aren’t willing to pay long-term based on imagining “the next-generation search entry point.” In the end, they have to ask: whose click is this? Where does the conversion happen? How much of the budget should be moved from Google, Meta, and TikTok to ChatGPT?

Category fit is also a problem. Low-risk categories like home, travel, education, and software tools can be tested first. But the highest-profit categories are often also the most heavily regulated—such as finance, healthcare, insurance, and recruiting. Once ChatGPT runs ads in these fields, the platform must bear not only ad performance risk, but also misinformation, discrimination, and compliance risk.

Google’s approach is a mirror. In Q1 2026, Google’s search ad revenue was $77.25 billion. Yet even so, Google remains extremely cautious about ad placements in AI Mode and AI Overviews, and the standalone Gemini app still hasn’t officially started running ads.

OpenAI expanding into advertising is an exploration of broader business models for the entire large-model race.

OpenAI needs users to feel that AI is intimate enough, while convincing advertisers that there is enough commercial intent here. Once this balance slips, ChatGPT will lose both sides at the same time: users will feel it’s not pure, and advertisers will feel it can’t convert.

But the changes brought by advertising go beyond this—it is also reshaping how brands compete.

The center of gravity of GEO is shifting

Over the past year, brands have worried about whether they will disappear from AI answers. The market packages this as GEO, but in essence it isn’t a new concept—it’s just an old search marketing anxiety wearing a new shell in the AI era.

When OpenAI rolls out Ads Manager, it taps directly into this anxiety—but it also changes the direction of that anxiety.

In a no-ads era, the core issue for GEO is: “how to enter AI’s context.” Brands compete to be referenced by the model through product documentation, media coverage, third-party evaluations, and community discussions—competing on information quality and how well the data is structured.

After the ad platform launches, precise traffic can be purchased directly, and brands no longer rely solely on organic mentions. But the competitive focus doesn’t revert back to “buy more exposure.” Instead, it shifts from “how to get into AI’s answers” to “how will AI evaluate my product?”

The reason is simple: once users see an ad, the most natural next step is to ask AI, “Is this product actually any good?” The AI’s answer becomes the real conversion gate. Advertisers can buy impressions, but they can’t buy AI’s good reviews. If AI gives negative feedback based on publicly available data, every dollar spent accelerates user drop-off instead of driving conversions.

This means brands must build positive reputation within AI’s evaluation system. Signals that AI can read—product quality itself, the density of user reviews, and the coverage of third-party evaluations—will matter more to conversion than ad placements themselves.

GEO is moving from “entering the context” to “winning evaluations,” and this is a trend worth watching after OpenAI’s launch of its new advertising business platform.

Not running ads is the most expensive ad in 2026

Having discussed OpenAI, we have to talk about its nemesis, Anthropic, which is taking a completely different “advertising model.”

On February 4, 2026, two days before the Super Bowl, Anthropic published a blog post stating that Claude will never run ads. There were no sponsored links and no third-party placements.

That sentence itself is an expensive advertisement.

Super Bowl ads aren’t cheap. By spending heavily to tell users that it doesn’t sell ads, Anthropic is essentially buying brand recognition with an ad-free message.

Going ad-free has never been only a moral position—it’s also a business positioning. It tells enterprise customers, professional users, and highly sensitive scenario users that Claude’s responses won’t be influenced by advertisers, that Claude’s product direction won’t be optimized around ad inventory, and that its revenue comes from what you pay.

The effect is immediate. Claude’s ranking on the US App Store climbed from 42nd at the beginning of the year and kept rising. On February 28, after OpenAI signed a Pentagon contract that sparked the QuitGPT movement, Claude first topped the US App Store free apps for the first time ever, surpassing ChatGPT. Free active users increased by 60%, daily registrations quadrupled, and paid users doubled within a week.

Anthropic’s revenue structure is completely different from OpenAI’s: over 80% comes from enterprise customers, and annual recurring revenue often surged from about $9 billion to $19 billion. Enterprise tools such as Claude Code and Cowork have already contributed at least $1 billion in revenue. Anthropic doesn’t need the advertising value from free users—it needs enterprise customers’ trust premium that their data won’t be used for ads.

In this context, not running ads is a precise business decision: by giving up ad revenue, it strengthens the trust barriers of enterprise customers, thereby supporting higher subscription pricing.

However, “not running ads” isn’t an eternal virtue.

Stanford’s AI Index shows that the cost to reach performance equivalent to GPT-3.5 dropped 280 times in two years, from $20 per million tokens in November 2022 to $0.07 in October 2024. If model capabilities continue to converge and an API price war breaks out, the enterprise subscription premium Anthropic enjoys today could be gradually eroded. When model costs fall enough for all competitors to offer roughly similar performance, why would enterprise customers keep paying more for Claude?

No conclusion yet, but time will answer it.

There’s no free lunch

OpenAI chooses ads; Anthropic turns not running ads into a premium. They look like two opposite paths, but they’re actually answering the same question: when AI product reasoning costs can’t be covered long-term by free models, who will foot the bill?

OpenAI’s Ads Manager isn’t just an ad product—it’s a signal that the AI industry is shifting from free expansion to cost recovery.

But the bleeding-stopping method OpenAI chose, ironically, exposes the most fragile part of this business. It needs to prop up ad pricing that is three times higher than Meta’s using user groups with the least commercial intent.

This isn’t a problem that can be solved by user scale alone. 900 million weekly active users is a nice number, but if those 900 million people come to ChatGPT to write emails instead of buy things, advertisers will eventually vote with their feet.

Ads can be a source of revenue for AI products, but they shouldn’t be treated as the only answer. Because when a product’s business model requires users to stay as long as possible and expose as many intents as possible, that product is no longer an assistant to users—it becomes an assistant to advertisers.