OpenAI's cheap ChatGPT tier could turn subscriber growth into worse economics
OpenAI's latest subscriber leak matters because it points to a problem every consumer AI company is heading toward: you can grow fast and still end up with worse customers. The Information reported that OpenAI expects more of ChatGPT's paying growth to come from the cheaper Go tier than from Plus. type0 has not independently verified those internal numbers. But the pressure is real because OpenAI has already built the public product stack that would make that shift possible.
That stack is not just a cheaper plan. OpenAI has launched ChatGPT Go at $8 a month in the United States, and in a separate policy post said it plans to test ads on Free and Go while keeping higher tiers ad-free. If the leak is directionally right, the story is not simply that ChatGPT may reach 122 million subscribers. OpenAI may be growing by steering more users into a lower-priced, ad-supported lane.
OpenAI's own documents show how far that lane already extends. The company said in its Go launch post that the plan is rolling out wherever ChatGPT is available and costs $8 per month in the United States. In its advertising post, OpenAI said Go has launched in 171 countries since August and is now its fastest-growing plan. In a separate company post, it said ChatGPT has more than 900 million weekly active users and more than 50 million consumer subscribers. The scale is already there. The question is what kind of subscriber base OpenAI is building on top of it.
Go also is not an empty downgrade tier. The OpenAI Help Center says it includes GPT-5.3 Instant, a faster lower-cost model, plus longer memory, file uploads, image generation, and access to projects, tasks, and custom GPTs. Users lose access to older models and Sora, OpenAI's video generator, but they still keep enough features to make trading down plausible for many casual subscribers. The same Help Center page says Plus and Pro users can switch to Go at the end of a billing cycle. OpenAI has made the lower tier cheap, functional, and easy to reach.
The ad story needs a careful line because OpenAI's own documents do not describe it the same way. The policy post says OpenAI plans to begin testing ads on Free and Go in the United States soon, while the Help Center ads page says the U.S. test begins on Feb. 9, 2026. Both sources agree on the main structural point: ads are for Free and Go, while Plus, Pro, Business, Enterprise, and Edu stay ad-free. That split does not prove Plus is collapsing. It does show OpenAI is drawing a sharper line between premium users and everybody else.
The optimistic case is easy to see. A cheaper plan can bring in users who would never pay $20 a month, especially in markets where Plus is expensive relative to local incomes. OpenAI's public framing leans hard on access, and 171-country rollout gives that argument substance.
The harder question is what that access does to the economics. Independent analyst Ed Zitron wrote in Wheres Your Ed At that The Information reported OpenAI expects Plus subscribers to fall from 44 million in 2025 to 9 million in 2026 while Go rises from 3 million to 112 million. Those numbers remain secondhand here, and they should be treated that way. But Zitron's broader point survives even if the exact forecast is wrong: a subscriber chart can keep rising while the underlying business gets weaker if too much growth comes from cheaper plans.
That is the pressure to watch next. OpenAI has already built the rails for a more mass-market version of ChatGPT: lower pricing, broader international reach, and ads reserved for the bottom of the funnel. The unresolved question is whether those rails expand the market without hollowing out Plus, or whether consumer AI is sliding toward the old web model faster than the industry wants to admit.