No application in history has reached one billion users faster than ChatGPT. The milestone was hit in June 2026, putting OpenAI in a category of its own — but growth doesn’t come free. OpenAI faces cost pressures already visible in free tier degradation and existential questions about its business model.
The milestone in context
ChatGPT reached 100 million users two months after launch (January 2023). It hit 400 million by early 2025. The jump from 400M to 1,000M in roughly 18 months is impressive, but the growth curve has flattened compared to the initial phase — inevitable after the explosive first year of adoption.
The billion users aren’t just free users. OpenAI has diversified its base: the free tier remains massive, ChatGPT Plus has a stable subscriber base, ChatGPT Team grows in the SMB segment, and ChatGPT Enterprise — though slow in adoption — is starting to move with the arrival of GPT-5.5 and AWS Bedrock integration.
The cost of a billion users
Maintaining a billion active users on a generative AI product is fundamentally different from maintaining a billion users on a social network or search engine.
Every query to ChatGPT costs inference money. It’s not like Facebook, where the marginal cost of an additional user approaches zero. It’s as if every post, every “like”, every scroll cost Meta cents of a dollar in GPU time. At the scale of a billion users, those cents become billions of dollars.
OpenAI has taken steps to manage this cost pressure:
- Free tier degradation: free users have increasingly restricted access to the most capable models. GPT-5.5 is only available to paid subscribers.
- Rate limits: even on paid tiers, usage limits reflect available inference capacity, not just product decisions.
- Efficiency pressure: OpenAI has optimized its inference models to reduce per-query costs, but total volume keeps growing faster than efficiency gains.
OpenAI’s structural dilemma
Behind the user milestone is a question OpenAI hasn’t fully answered: how do you make money giving away free access to one of the most expensive software products to operate?
The answer so far has been a combination of:
- Subscriptions: Plus ($20/month), Team ($25/user/month), Enterprise (custom pricing)
- API: the real business, where developers pay per token
- Strategic partnerships: the AWS deal (covered in another article) and Microsoft’s investment
- Future monetization: rumors of advertising, superapp, and other yet-unconfirmed mechanisms
But the numbers don’t easily add up. June 2026 reports indicate OpenAI still faces financial pressure — the cost of training and operating GPT-5.5, plus the infrastructure needed to serve a billion users, exceeds current revenue by a significant margin.
What it really means
That ChatGPT has a billion users is, above all, a sign that conversational AI has become a universal interface. People don’t use ChatGPT because it’s AI — they use it because it’s the easiest way to get answers, write texts, analyze data, or generate ideas. For hundreds of millions of people, it’s the new entry point to the internet.
But the milestone also reveals the model’s fragility: OpenAI has the most successful product of the generative AI era and yet the unit economics are precarious. If a company with a billion users and the world’s most advanced models can’t make the numbers work without friction, the question about AI’s economic sustainability stops being theoretical.
A billion users is a monumental achievement. Turning that achievement into a sustainable business is the real challenge — and it’s not yet solved.