OpenAI announced it will begin testing ads in ChatGPT free and the $8/month Go tier in the U.S. within weeks to raise revenue, the company stated and later tweeted. “In the coming weeks, we plan to start testing ads in ChatGPT free and Go tiers.”
A year and a half earlier, Sam Altman called ads “uniquely unsettling” and a “last resort”, as seen in an earlier comment. The shift appears driven by mounting costs and losses.
Internal figures show a roughly $8 billion operating loss in 2025, with projected operating losses hitting about $74 billion by 2028, according to reporting that details those estimates (Ed. note: $74 billion is the projected shortfall). The company reports roughly 800 million weekly users, with about 5% paying subscribers, and has committed over $1.4 trillion to infrastructure, reported recent coverage.
Market share for ChatGPT fell from about 87% in January 2025 to roughly 65% this month while Google‘s Gemini rose from about 5% to over 18%, as data shows. Competition and distribution give Google advantages that affect adoption.
Hardware costs also matter; analysis finds Google’s TPUs cost roughly four to six times less per unit of compute than the Nvidia GPUs OpenAI uses, creating a so-called “Nvidia tax.” That comparison is outlined in a detailed analysis.
OpenAI published five ad principles promising privacy and limited influence over answers, and said Pro, Business, and Enterprise tiers will remain ad-free. The first ads will appear as basic product placements below responses, and long-term impact remains uncertain given past ad-industry tradeoffs noted in historical discussion.
