Author(s): MohamedAbdelmenem

Originally published on Towards AI.

Forget the 1B users. The 40% enterprise revenue split shows where the real money is.

OpenAI is generating a $40 billion annualized revenue run rate, equating to roughly $3.3 billion monthly, and 40 percent of it now comes from enterprise B2B contracts. That leaked CFO metric renders the celebration over ChatGPT’s 1 billion consumer users irrelevant, turning the user base into an expensive liability ahead of a target 2027 IPO. Here is the exact margin math driving that $852 billion valuation, and how enterprise CIOs must lock in their infrastructure before public shareholders force the company to throttle its consumer tiers.

The article argues that the headline “1B users” is misleading, since the company’s reporting shift from weekly to monthly metrics may indicate stalled retention rather than true momentum. It claims OpenAI’s $40B run rate is increasingly powered by enterprise and API revenue—about 40%—which enables predictable recurring margins needed for an IPO and “S-1 sanitization” (i.e., cleaning up financials by cutting operational fat). It further explains the “compute liability” of serving free consumer traffic, describing how unpaid users effectively load-test infrastructure and drive inference costs that enterprise revenue must offset. Finally, it lays out a CIO-oriented playbook: audit and cap consumer-tier workflows, lock in multi-year enterprise contracts, and decouple non-critical workloads (especially RAG) to reduce dependence on low-cost API assumptions before public-market pressure forces consumer-tier throttling.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI