There is a distinct rhythm to the evolution of Silicon Valley unicorns, a trajectory as predictable as a Greek tragedy. Act One is the Promethean phase: a charismatic founder offers fire to the masses, promising that this time, the technology is too pure, too transformative, to be sullied by the grubby business of hawking goods. Act Two is the reality check, usually precipitated by the immutable laws of accounting. In Act Three, the ads arrive.
OpenAI, the maker of ChatGPT, has officially entered its third act.
It was not so long ago—though in AI years, it feels like an eon—that Sam Altman, OpenAI’s boss, declared that advertising would be a “last resort.” He framed the subscription model as a moral imperative, implying that relying on ad revenue would distort the sanctity of the relationship between the chatbot and the user.
In 2026, Mr. Altman’s reluctance seems to have dissolved as he will begin testing advertisements for ChatGPT “Free” and “Go” tier users, starting in North America.
While the volte-face is hardly surprising, the timing is instructive. OpenAI is generating revenue at a run rate of $10bn a year—a figure that would make most CEOs weep with joy. Yet, the costs of training and running Large Language Models (LLMs) are eye-watering. And so the decision to embrace ads, admitting that subscriptions alone cannot support the most capital-intensive startup in history.
And OpenAI’s ads pivot looks like a defensive manoeuvre in a war that is turning against the incumbent.
Google, the heavyweight champion of digital advertising, has been steadily gaining ground with its Gemini models. More bruisingly, Apple, which only a year ago made ChatGPT the default AI for its devices, has shown its notorious fickleness. The iPhone maker is now shunting traffic toward Gemini, leaving OpenAI to contemplate a future where it loses its prime real estate on the world’s most valuable screens.
If OpenAI cannot rely on Apple, it must double down on monetising the ones it has more aggressively. But the transition carries risks that go beyond mere annoyance.
The proposed mechanism for these ads is novel. Rather than slapping banner ads atop a conversation, OpenAI plans to insert “interfaces” within the chat, allowing brands to converse directly with users. The backend will likely mirror the programmatic bidding wars that power Google Search. But the context is different.
Search is transactional; chat is conversational, occasionally even confessional. A user asking ChatGPT for advice on a divorce or a medical diagnosis may not welcome an interjection from a divorce lawyer or a pharmaceutical giant.
There is a perverse incentive at play, too. To serve an ad at the end of a response, the model may be subtly optimised for brevity. A chatbot that rambles is a chatbot that delays the revenue event. The risk is that the “oracle” becomes less helpful, shortening its wisdom to make way for the pitch.
This changes the fundamental nature of the product. An ad-supported search engine is a directory; an ad-supported chatbot is a double agent. It serves the user, but it ultimately answers to the highest bidder. OpenAI is betting that its users, particularly those on the free tiers, are price-sensitive enough to tolerate the intrusion. But with Google and nimble open-source competitors breathing down its neck, user attrition is a genuine threat.
Mr Altman’s startup may have built the smartest machine in history, but to keep it alive, it must now teach it to be a salesman. We shall wait and watch to see if the customers are buying.
Published – January 23, 2026 07:53 am IST