
OpenAI’s current business rests on just two engines. Despite $1.4 trillion in infrastructure commitments and a $300B+ valuation, the company generates meaningful revenue from only two sources.
Engine 1: Subscriptions (~60-65% of revenue)
Estimated $8-10B annually. ChatGPT Plus at $20/month remains the core offering, with over 20 million paying subscribers across Plus, Pro ($200/month), Team, and Enterprise tiers.
The numbers tell a mixed story:
- 800 million weekly active users sounds impressive until you realize only 5% convert to paid
- 760 million users generate zero direct revenue
- Retention sits at 74% after three quarters — respectable but not exceptional
Engine 2: API & Enterprise (~35-40% of revenue)
Roughly $5B annually. Higher margins and stickier customers:
- 92% of Fortune 500 companies use OpenAI in some capacity
- Seven customers generate over $100M annually each
- $1B+ monthly ARR suggests strong enterprise demand
The Market Share Problem
Enterprise market share has declined from 50% to 25% as Anthropic gains ground:
- Anthropic now commands 32% of enterprise LLM utilization
- In coding: Claude holds 42% market share vs. OpenAI’s 21%
- Revenue efficiency: Anthropic generates $211 per monthly user vs. OpenAI’s $25 per weekly user — an 8x difference
The two-engine model works. But the core business is losing ground while demanding 15x growth.
This is part of a comprehensive analysis. Read the full analysis on The Business Engineer.









