Netflix took twenty-five years to reach $13 billion in annual revenue. Salesforce took twenty. Google took nine. Facebook took ten. OpenAI did it in roughly two years after launching ChatGPT. And by the time you read this, that number has already climbed past $25 billion on an annualized basis. No company in the history of technology has ever grown revenue this fast, and the trajectory suggests we are still watching the early innings of something unprecedented.
I remember the moment ChatGPT launched in late November 2022. The product felt like a curiosity. An interesting research demo. The kind of thing you showed your friends at a dinner party and then forgot about by Monday morning. I certainly did not think I was watching the birth of a company that would generate over a billion dollars a month within two years. Nobody did. And that fundamental underestimation — by competitors, by analysts, by almost everyone — is a big part of how OpenAI pulled this off.
A Billion Dollars a Month
Let me sit with that number for a moment, because it is easy to read "$13 billion ARR" and not fully register what it means in practical terms. OpenAI makes approximately $1.08 billion every month. That is roughly $36 million per day. About $1.5 million per hour. Every hour, around the clock, OpenAI is generating more revenue than most startups raise in their entire seed round.
The revenue comes from three streams, and understanding the mix tells you a lot about why the company grows the way it does.
ChatGPT subscriptions make up approximately 60 percent of revenue. Over 200 million people use ChatGPT every week. ChatGPT Plus runs $20 per month. Teams pricing sits at $25 to $30 per user per month. This is the consumer and small business engine, and it is enormous. When people argue about whether consumers will pay for AI, OpenAI has already settled that debate with over 200 million weekly data points.
API revenue accounts for roughly 30 percent. Thousands of companies have built their products on top of OpenAI's models — GPT-4o, GPT-4 Turbo, DALL-E, Whisper, the Assistants API. Pricing has dropped dramatically. GPT-4o input pricing went from $30 per million tokens down to $2.50. That is a 92 percent price reduction. But volume growth has more than compensated. When you make something twelve times cheaper, a lot more people use it, and the total revenue goes up, not down. This is the same dynamic that built AWS, and it is working for OpenAI.
Enterprise contracts make up the remaining 10 percent. ChatGPT Enterprise at $60 per user per month for large organizations, plus custom model training and deployment for Fortune 500 companies. This is the smallest segment today, but it is growing the fastest in percentage terms. Enterprise deals are stickier, higher margin, and more predictable than consumer subscriptions. As this segment grows, it changes the quality of OpenAI's revenue in ways that investors care deeply about.
The Growth Curve That Broke Every Model
Here is what the growth trajectory actually looked like, and I want you to pay attention to the pattern because it tells you something important about how AI adoption works.
In December 2022, when ChatGPT launched, revenue was effectively zero. Research lab, no commercial product. By December 2023, one year later, OpenAI had hit $1.6 billion in annual recurring revenue. That first year was the proof of concept — the moment when a research breakthrough became a consumer product that people were willing to pay for.
By June 2024, six months later, revenue had doubled to $3.4 billion. By December 2024, it doubled again to $6 billion. By July 2025, it doubled again to $13 billion. Revenue roughly doubled every six months for two straight years. That kind of sustained geometric growth has essentially no precedent in the history of business. The closest comparison is perhaps the early iPhone era at Apple, but even Apple's smartphone revenue did not double every six months for two years running.
Now the 2026 projections point to somewhere north of $25 billion annualized. Some estimates suggest the company could reach $30 billion if current trends hold. The growth rate is decelerating in percentage terms — you cannot double from $13 billion as easily as you double from $1.6 billion — but in absolute terms, OpenAI is adding more revenue per quarter than most public technology companies generate in an entire year.
Why the Growth Does Not Slow Down
The natural question is: when does this stop? Every high-growth company eventually hits a ceiling. The answer, as far as I can tell, is that OpenAI has not yet reached the ceiling for any of its three revenue streams.
On the consumer side, 200 million weekly active users sounds like a lot, but the potential market is every knowledge worker on earth — roughly one billion people. Penetration is still below 20 percent of the addressable market. And each new model generation brings capabilities that expand who finds the product useful. When GPT-3.5 launched, it was useful for writing and brainstorming. GPT-4 added reasoning, code generation, and image understanding. Each capability expansion brings in a new wave of users who previously did not see the value.
On the API side, the price reductions are actually accelerating growth rather than hurting it. This is counterintuitive, but it is the same playbook Amazon used with AWS. When you reduce API prices by 90 percent, you do not lose 90 percent of your revenue. You unlock use cases that were previously uneconomical, and the volume explosion more than compensates. Companies that could not afford to build on OpenAI's API at $30 per million tokens are building massive products at $2.50 per million tokens. The cheaper you make the API, the bigger the ecosystem becomes, and the more total revenue you generate.
On the enterprise side, OpenAI is just getting started. Most Fortune 500 companies are still in pilot phases with AI. The enterprise sales cycle is 6 to 18 months, which means deals that started in 2025 are closing in 2026. The pipeline behind the current numbers is reportedly massive.
The Valuation Question Everyone Is Asking
OpenAI was reportedly valued somewhere between $300 billion and $500 billion in early 2026. At $13 billion in ARR trending toward $25 billion, that is a 25x to 38x revenue multiple depending on which valuation figure and which revenue number you use. Is that expensive? By traditional metrics, absolutely. The median public SaaS company trades at 6x to 8x revenue. Even the most premium software companies rarely sustain 20x.
But traditional metrics may not apply here, and here is why. OpenAI is not a SaaS company. It is an infrastructure company that is becoming the default AI layer for a significant portion of the global economy. The better comparison might be AWS in its early years, or Google in its first decade. Both of those companies were "expensive" at every point on the growth curve, and both of those companies made their early investors spectacularly wealthy because the market they were capturing turned out to be far larger than anyone initially projected.
SoftBank put in $40 billion in March 2025 — the largest single venture investment in history. Total funding exceeds $60 billion. These are not naive investors. They are making a specific bet: whoever builds the best AI infrastructure captures a market measured in trillions, not billions. At a trillion-dollar market cap, the current valuation looks cheap. At $300 billion, it looks expensive. The bet is on the endgame, not the current state.
I will say this: I am personally skeptical of any valuation that requires a company to be worth a trillion dollars to justify the entry price. A lot has to go right for that to happen. But I also would not have believed you if you told me in November 2022 that the chatbot I was playing with would generate $13 billion in annual revenue within two years. My ability to predict what happens next in AI has been consistently wrong on the conservative side.
What This Means If You Are Building Something
If you are a startup founder, OpenAI's revenue trajectory is both encouraging and terrifying. Encouraging because it validates, beyond any argument, that the AI market is real. Consumers pay for AI. Businesses pay for AI. This is not speculative. $13 billion in recurring revenue settles the question. The total addressable market for AI products is large enough to support thousands of successful companies, not just one.
Terrifying because you are competing with a company that has near-infinite capital, the strongest brand in AI, and a product that 200 million people use every week. Any feature you build, any market you enter, there is always the question of whether OpenAI will simply add that capability to ChatGPT and make it available to their existing user base for free.
The founders who navigate this well are the ones building in areas where general-purpose AI models are not good enough. Vertical AI companies in healthcare, legal, finance, and construction are thriving because domain-specific expertise requires more than a large language model. It requires proprietary data, specialized training, regulatory knowledge, and industry relationships that OpenAI cannot replicate by adding a feature.
If you are a developer or a solo builder, the picture is more straightforwardly positive. OpenAI's API business means there is a thriving ecosystem of companies building on their infrastructure. The "wrapper" criticism that people love to throw around misses a fundamental point: those wrapper businesses collectively generate billions in revenue by solving specific problems that raw API access does not solve. The tools are getting cheaper and more capable every quarter. The opportunity to build profitable products and services on top of the OpenAI ecosystem is expanding, not contracting.
If you are an investor, AI is no longer a thesis — it is a proven multi-billion-dollar market. OpenAI's $13 billion, combined with Anthropic's $5 billion, Databricks' $4 billion-plus, and the revenue from hundreds of smaller AI companies, demonstrates that enterprise and consumer AI generates real money at real scale. The investment question has shifted from "will AI generate revenue?" to "which AI companies will capture the most value, and at what price is entry still attractive?"
The Speed of Change Is the Story
What strikes me most about OpenAI's numbers is not the absolute size — though a billion dollars a month is staggering — but the speed. The velocity of revenue growth tells you something about the velocity of AI adoption in the broader economy. When one company can add $3 billion to $5 billion in annual revenue every six months by selling AI access, it means millions of people and thousands of companies are integrating AI into their work at a pace that exceeds any technology adoption curve we have seen before.
The internet took a decade to become essential infrastructure. Mobile took about seven years. Cloud computing took roughly five. AI is compressing that timeline into two to three years. The speed at which OpenAI generates revenue is a proxy for the speed at which the world is adopting artificial intelligence as a fundamental tool, and that speed is faster than anyone projected.
The question for everyone else is not "should I use AI?" That question has been answered by 200 million weekly ChatGPT users and $13 billion in annual revenue. The question is how fast you can integrate it into whatever you are building, selling, or doing — because the companies and individuals who figured that out first are the ones generating the numbers I have just walked you through. And the gap between early adopters and late adopters is widening every single quarter.
---
Keep Reading
- Midjourney: $500M Revenue, Zero Funding, 40 Employees — the polar opposite model: the most profitable AI company raised zero dollars.
- Klarna AI Replaces 700 Agents — what happens when a company actually deploys AI at scale inside its operations.
- How to Make Money with Claude AI — practical income opportunities built on top of the AI ecosystem OpenAI helped create.
- AI Kills Middle-Class Jobs — or Does It? — the workforce implications of AI adoption at this speed and scale.
