Two companies that didn't exist in any meaningful commercial form five years ago are now collectively printing money faster than the golden arches and the green mermaid put together. According to Axios, OpenAI and Anthropic are on pace for a combined annual revenue rate of roughly $120 billion. Let that number sit with you for a second.
The Numbers That Should Make Your Brain Hurt
Axios is reporting the $120 billion figure based on data from Funda, an AI investment research platform, flagged by Tae Kim, author of the "Key Context" Substack. Of that combined haul, Anthropic accounts for roughly 60 percent, which means the company behind Claude is quietly outpacing its more famous rival on raw revenue.
To understand what $120 billion in annual run rate actually means, consider that McDonald's total revenues in 2024 were around $25 billion. Starbucks pulled in just over $36 billion. You could stack both of those iconic, globally ubiquitous brands on top of each other and still have money left over for a third helping. These are companies with decades of brand recognition, billions of physical transactions, and armies of minimum-wage workers. OpenAI and Anthropic have servers and a handful of researchers.
This is not a drill. This is not hype dressed up as a revenue projection. These are real dollars flowing in from real customers paying real subscription and API fees at a pace that would have sounded delusional at a venture capital pitch meeting in 2022.
Why Anyone Would Spend Trillions on AI Infrastructure
Axios frames this as one of the first genuinely compelling answers to a question that has haunted the AI investment bubble for two years: why are companies and governments committing trillions of dollars in infrastructure spending on this technology? The forecasts have always sounded like science fiction written by people who also buy their own press releases.
But here is the thing. When the revenue curve looks like this, when two companies can go from essentially zero to $120 billion in annual run rate in roughly three years, the trillion-dollar infrastructure bets start looking less like irrational exuberance and more like rational calculation. The customers are clearly there. The money is clearly there. The question stops being "will AI make money" and becomes "who is going to own the pipes."
That is an important shift. It moves the story from speculation to competition, and competition at this scale gets ugly fast. Microsoft, Google, Amazon, and Meta are all in this fight. They have been subsidizing the buildout of the exact infrastructure that OpenAI and Anthropic are now monetizing at unprecedented speed.
Anthropic's Quiet Dominance Is the Real Surprise
OpenAI gets the headlines. Sam Altman does the interviews. ChatGPT is the name your parents actually know. But if the Funda data Axios cited is accurate, Anthropic is the one quietly winning the revenue race, accounting for the larger share of that combined $120 billion figure.
Anthropics Claude has become the go-to API for enterprise customers who want reliability and safety guardrails baked into their products. While OpenAI has been lurching through executive drama, product pivots, and a very public identity crisis about whether it is a nonprofit, a capped-profit, or just a regular aggressive tech company, Anthropic has been cashing checks from corporate America with relative calm.
Damon Beres, writing for The Atlantic earlier this year, noted that Claude has become something of an enterprise darling precisely because it is not ChatGPT. Companies nervous about brand risk and liability have been gravitating toward Anthropic the same way they once chose IBM over whatever was cheaper. You don't get fired for buying the safer-seeming option.
What the Big Tech Giants Are Watching Right Now
Microsoft owns a significant chunk of OpenAI. Google has poured billions into Anthropic. Amazon has too. These are not passive investments from companies that wandered into a good deal. These are defensive positions from companies watching a potential competitor eat the world and wanting to be inside the tent when it happens.
The $120 billion combined run rate figure will reverberate through quarterly earnings calls for the rest of the year. Every tech CEO with AI infrastructure spending on their balance sheet is going to wave this number around like a permission slip. The returns are coming. The model works. Please stop asking us to justify the capital expenditure.
Whether this pace holds is a different question entirely. Revenue acceleration this dramatic almost always runs into something: regulatory walls, market saturation, a competitor who cracks a cheaper model, or a high-profile failure that spooks the enterprise customer base. But right now, in the summer of 2026, the AI money printer is running hot.
The Dingo Take
Here is the uncomfortable truth hiding inside a very impressive revenue number. The fact that two AI companies are now outpacing McDonald's and Starbucks in annual run rate does not tell us whether any of this is good for the world, for workers, or for the people whose creative and intellectual labor these models were trained on. It tells us that a very small number of people built something that a very large number of businesses are now completely dependent on, and those businesses are paying through the nose for the privilege. That is a description of a monopoly in formation, not a success story.
The trillion-dollar infrastructure bets now look smarter, sure. But who is building that infrastructure? Largely the same people who created the environmental catastrophe of data center water consumption, who are running power grids into the ground in Texas and Virginia, who are promising AI agents will replace the entry-level jobs that used to let young people learn how to work. The money is real. The externalities are also real. The gap between those two facts is where the next decade of policy fights will happen, assuming anyone in Washington can still read a bill without an AI summarizing it for them.
Anthropics quiet dominance is the actual news here, and it deserves more attention than it will get. OpenAI has the cultural footprint. Anthropic apparently has the revenue. That gap between fame and profit is the kind of thing that reshapes industries when it finally becomes visible to everyone, not just the people reading AI investment research newsletters before breakfast.
