ChatGPT Just Started Selling Ads in India. Its Two Biggest Rivals Are Making a Point of Not Doing That.

OpenAI began showing ads inside ChatGPT to Free and Go-tier users in India this week — a sponsored product or service appearing at the bottom of a response, clearly labelled and visually separated from the answer itself, according to the company. India is ChatGPT's second-largest market by user count, with more than 100 million weekly users as of February 2026 spanning students, developers, educators and enterprises. WPP and Omnicom are OpenAI's first ad-agency partners in the country, more than 50 brands are expected to go live this week through a managed-ads arrangement, and a self-serve ads product — letting any advertiser buy placements directly — is scheduled to launch on 4 September. This is not a small test. It is a full commercial rollout, in the market OpenAI has picked to run it first at this scale.

Why India, and why now

The Go plan, OpenAI's cheapest paid tier, costs ₹399 a month in India — a price point set specifically for a market where a global subscription price would put the product out of reach for most of that 100-million-user base. A free or near-free product used by that many people is also, by the economics of consumer software, a product that needs another revenue line if it isn't going to lose money on every user forever. OpenAI has said plainly that the ads push is part of diversifying revenue ahead of a planned public listing — putting a second income stream, beyond subscriptions and enterprise API contracts, into the numbers before the company goes to public markets. India, as the largest population of ChatGPT's lowest-paying and non-paying users, is where that math is most under pressure, which is a reasonable account of why the ad rollout is landing here first and at this scale rather than in a market where subscription revenue is already carrying the load.

OpenAI's own framing of the safeguards is specific: advertisers cannot access user conversations or personal data, and ads are excluded from sensitive categories including health and politics. Those are real constraints, and they're more restrictive than the norm for most ad-supported consumer software. What they don't resolve is the more basic tension underneath contextual advertising in a conversational product: for an ad to be relevant enough to a specific reply to be worth showing at all, something about the content of that reply has to inform which ad appears, even if no advertiser ever sees the raw conversation. OpenAI's claim is about who gets access to the data, not about whether the conversation's content shapes what gets shown — and those are different assurances, worth keeping distinct rather than treating "advertisers can't read your chats" as the whole answer to "does this product now watch what I say to decide what to sell me."

The choice of launch partners is its own signal. WPP and Omnicom are two of the largest advertising holding groups operating in India, with existing relationships across most major Indian and multinational consumer brands already spending on television, digital display and search. Routing the initial 50-plus-brand rollout through them, rather than opening self-serve access from day one, is a way of borrowing that existing brand-safety and buying infrastructure instead of building trust with individual advertisers one at a time — the self-serve product arriving on 4 September, after the agency-led launch has had a week to run, reads as OpenAI wanting the bigger, harder-to-vet pool of direct advertisers to arrive only once the format has already been tested by partners it already has a relationship with.

Two competitors, the same market, the opposite answer

Google's DeepMind leadership has said explicitly that the company has no plans to put advertising into Gemini, its consumer AI chatbot — a notable position for a company whose core search business is built almost entirely on ad revenue, and one that reads as a deliberate line drawn between its two products rather than an absence of ad-tech capability. Anthropic went further and made it a formal commitment: on 4 February 2026, the company pledged to keep its Claude conversational product ad-free. Anthropic's public positioning leans hard on the contrast — enterprise and developer-focused, monetised through subscriptions and API usage rather than attention, a framing this site covered when Anthropic opened its Bengaluru office and named India its second-largest market by its own measure.

That split is being described in industry coverage as OpenAI sitting squeezed between the two — Anthropic pulling the enterprise and developer audience with a no-ads, subscription-first pitch, Google already dominant in consumer reach through Search and Android distribution without needing ChatGPT's playbook to monetise attention. OpenAI's ads move can be read as the company choosing the one lever its two biggest rivals have each, for different reasons, chosen not to pull.

What actually changes for an Indian user

For someone on ChatGPT's Free tier in India, the practical change is straightforward: a sponsored line will now sometimes appear under an answer, in a market where advertisers are told they cannot see the conversation that produced it. For someone paying ₹399 a month for Go, the same ads apply — the ad-supported tier isn't limited to non-paying users, which is worth noting given that paying a subscription is, for most other software categories, precisely the thing that removes ads. Whether that trade is reasonable is a judgment each user is entitled to make for themselves; what's verifiable here is only that OpenAI has now made the choice its two largest rivals have publicly said they won't, in the one market where the free-user math is largest enough for the difference to show up.