Ads infinitum Analysis September 3, 2026 Emerging Europe Insight OpenAI is putting ads in ChatGPT and Anthropic will not touch them. It is hardly a reinvention, yet ads never stopped anyone using Google. Artificial intelligence is a powerhouse of reinvention, or so we’ve been told.
AI is reinventing drug discovery , the battery , the operating system of work , home appraisals , shops , and even the Olympic Games . It remains to be seen how much of this reinvention is anything of the sort, and whether or not AI simply makes things more efficient, or perhaps merely reinvents the wheel. An announcement from OpenAI this week does rather point to the latter.
ChatGPT Ads, OpenAI claims, has reached one billion US dollars in annualised revenue run rate, is used by tens of thousands of advertisers and from this week becomes available to advertisers in India, Europe, the Middle East, and North Africa. Ads within ChatGPT appear for OpenAI’s Go subscribers and its free tier, which make up the vast majority of its users. Ads are clearly labelled and do not influence ChatGPT’s answers, OpenAI says , which adds that advertisers do not have access to users’ private conversations.
But does this represent a bold new frontier in advertising, a genuine reinvention, or simply a quick way to boost revenue? To the untrained (perhaps even cynical) eye, it looks a lot like Google’s business model of two decades ago: become the market leader and sell advertising around the product. There’s little reinvention going on, although that does not mean the move is without financial merit.
As asking a chatbot for answers (as opposed to Google, or any other search engine—others are still available) becomes more common , there’s money to be made and OpenAI has clearly decided that it wants a share. Interestingly, Anthropic (the developer of ChatGPT’s biggest rival, Claude) has decided to do exactly the opposite, and eschew advertising revenue. In February, Anthropic made it clear that it did not view Claude as a platform for ads.
“There are many good places for advertising. A conversation with Claude is not one of them,” the firm said. It did, however, leave the door open just a little, adding that, “AI will increasingly interact with commerce, and we look forward to supporting this in ways that help our users.
We’re particularly interested in the potential of agentic commerce, where Claude acts on a user’s behalf to handle a purchase or booking end to end. And we’ll continue to build features that enable our users to find, compare, or buy products, connect with businesses, and more—when they choose to do so.” For now, OpenAI embracing advertising and Anthropic rejecting it is one of the clearest demarcations between the two leading chatbots, and it is a demarcation that was first drawn, of all places, at the Super Bowl. In early February Anthropic (not without irony) ran four commercials (under the title ‘A Time and a Place’) in which an AI assistant stops mid-sentence to recommend something, each ad opening on a single accusing word (betrayal, deception, treachery, violation).
The tagline was, ‘Ads are coming to AI. But not to Claude.’ Sam Altman, OpenAI’s boss, initially said he found the ads funny, but then called Anthropic “ authoritarian ”. Nevertheless, Altman himself called the combination of AI and ads “uniquely unsettling” at a talk at Harvard in October 2024, adding that advertising was “a last resort”, the notion being that higher paying customers would subsidise other uses.
The notion now is to sell ads on the free and cheapest tiers. ‘It’s free so here are some ads’ As lucrative as it may turn out to be, it’s a moot point indeed if this resembles anything close to genuine reinvention. An ad under a ChatGPT answer is not really any different to the kind of bog standard display ad that has been common since the first one appeared back in 1994 .
The genuinely reinventive idea, some argue, is agentic commerce, which McKinsey believes could be worth as much as one trillion US dollars in the United States by 2030 (representing something like one-third of all online purchases). The premise is that an AI assistant stops recommending things and starts buying them, moving advertising from attention to action. OpenAI itself pioneered this when it rolled out Instant Checkout last September, allowing people to buy from Etsy sellers inside a chat, but dropped the feature in March .
But it does now seem to be the route that Anthropic might follow. Google’s ongoing, unprecedented success rather suggests that the old ‘it’s free so here are some ads’ model is here to stay, for the time being at least. Back in 1998, well before Google became the giant market search leader it is today, the firm’s founders Sergey Brin and Larry Page wrote in a Stanford paper that advertising-funded search would be “inherently biased towards the advertisers and away from the needs of the consumers”.
They then built the most profitable advertising business in history anyway, and nobody particularly seemed to mind. Users of just about anything have a history of putting up with ads as long as a product is inherently useful (and, perhaps equally importantly, when the alternative costs money), and for most of ChatGPT’s users it does. Bloomberg Intelligence found (albeit back in 2023) that 93 per cent of people will not pay more than 10 US dollars per month for a chatbot and that only one per cent will pay the 20 US dollars Anthropic currently charges for Claude’s cheapest tier.
OpenAI’s own forecasts , reported by The Information, see its own 20 US dollars (or more) tier shrinking from 44 million subscribers to nine million this year, while the ad-supported Go tier (which currently costs eight US dollars per month) goes the other way, from three million users to 112 million. It’s almost as if OpenAI is happy to push users towards ads. OpenAI has told investors it expects the advertising business to be worth 100 billion US dollars by the end of the dec
Topics: #ads #reinvention #infinitum
