OpenAI Reports 900 Million Weekly Users as Enterprise Revenue Nears Consumer Parity

OpenAI disclosed that ChatGPT has reached 900 million weekly active users and more than 50 million subscribers, cementing its position as the dominant consumer AI platform. The company reported that enterprise revenue now represents over 40% of total income and is projected to match consumer revenue by the end of 2026, marking a fundamental shift in the commercial AI landscape.

The company's API infrastructure now processes more than 15 billion tokens per minute, driven by what OpenAI describes as record engagement across agentic workflows. Codex, the company's code generation tool, has grown to serve over 2 million weekly users, representing a fivefold increase in three months with month-over-month usage growth exceeding 70%. ChatGPT's search functionality has nearly tripled in usage over the past year, while an advertising pilot reached more than $100 million in annual recurring revenue within six weeks of launch.

OpenAI framed compute capacity as a "strategic advantage" that powers every layer of its business, from frontier research to deployment and revenue generation. The company stated that both revenue and available compute have scaled rapidly since ChatGPT's launch, though it did not disclose specific figures for the latter. GPT-5.4 is cited as the driver behind the surge in agentic workflow adoption, though the company provided no technical details on the model's capabilities or release timeline.

(The disclosure comes as OpenAI faces intensifying competition from Anthropic, which recently signed a memorandum of understanding with the Australian government to share economic index data tracking AI adoption and workforce impact. The deal positions Anthropic as a partner in policy formation, a role OpenAI has historically sought in other jurisdictions.)

The enterprise momentum reflects a broader industry pattern in which AI companies are pivoting from consumer-focused growth to B2B revenue streams. Anthropic's government partnership in Australia centers on monitoring AI's economic effects, including job displacement and productivity shifts, suggesting regulatory interest in tracking deployment at scale. Meanwhile, Chinese semiconductor firms are consolidating capacity to support AI infrastructure outside Western supply chains. SMIC recently acquired the remaining 49% stake in its mature-node subsidiary for $5.7 billion, while Hua Hong absorbed Shanghai Huali Microelectronics for $1.2 billion, securing 38,000 additional 12-inch wafers per month at 65nm to 40nm nodes.

OpenAI's user metrics dwarf competitors, with ChatGPT reporting six times the monthly web visits and mobile sessions of the next-largest AI application, and four times the total time spent compared to all other AI apps combined. The company characterized these figures not as growth milestones but as evidence that frontier AI is embedding itself in daily routines globally. The advertising pilot's rapid revenue ramp suggests OpenAI is exploring monetization beyond subscriptions and API fees, though the company did not detail the pilot's scope or advertiser base.

The enterprise trajectory aligns with broader industry signals that agentic AI—systems capable of autonomous planning and execution—is moving from experimental deployments to production workflows. Security researchers have noted that agentic platforms are transitioning from passive recommendation tools to autonomous actors with direct system access, raising governance questions. OpenAI's emphasis on compute as a competitive moat underscores the capital intensity required to maintain leadership in a market where model performance increasingly correlates with infrastructure scale rather than algorithmic innovation alone.