OpenAI Projects ChatGPT Subscriptions To Reach 122 Million

In a January 16 blog post, OpenAI introduced a low-cost subscription tier, ChatGPT Go, priced at $8 USD/month, and noted the tier had launched in 171 countries since August (OpenAI blog). The post states, "we're also planning to start testing ads in the U.S. for the free and Go tiers," and sets out advertising principles including answer independence, conversation privacy, and user control over personalization (OpenAI blog). Separately, The Information reports that OpenAI projected the $8 consumer offering would drive consumer subscribers to 122 million this year (The Information).

Industry-pattern observations: Low-price consumer tiers combined with ad-supported free offerings are a common growth lever for consumer platforms. For practitioners, that pattern typically increases sustained traffic volumes and amplifies the need for scalable content-moderation pipelines, inference-cost monitoring, and robust telemetry to separate product metrics from ad metrics.

If the 122 million subscriber figure reported by The Information is representative of internal projections, it would materially change the scale of user interactions for a single consumer AI product. Large consumer scale tends to surface edge-case safety issues, raise latency and cost pressures for inference workloads, and push teams toward more aggressive caching, batching, and prompt-safety guardrails. OpenAI's public principles emphasize preserving answer independence and not selling conversation data to advertisers, which addresses common privacy and monetization concerns raised when ads meet conversational AI (OpenAI blog).

Observers should track the scope and format of initial ad tests in the U.S., metrics OpenAI publishes about ad personalization opt-outs, and any published third-party audits or red-team reports on ad influence. For engineering teams, watch for changes in rate limits, SLA communications, and any new telemetry or tooling OpenAI provides to distinguish ad-serving behavior from model response behavior.