What happened
CryptoBriefing reported Friday that Anthropic and OpenAI have crossed a combined $115 billion in annualized recurring revenue, a threshold that puts the two private AI labs on a run rate comparable to a mid-cap public software company. The reporting frames the number as a live snapshot rather than an audited disclosure, drawing on internal figures the two companies have shared with investors and enterprise customers over recent months. Neither Anthropic nor OpenAI has published a formal earnings release. The $115B figure is a sum of two private ARR marks, not GAAP revenue.
The scale reads differently against public comparables. Microsoft's Intelligent Cloud segment, the closest apples-to-apples benchmark, ran at roughly $130 billion in annualized revenue in its most recent fiscal disclosure. That's the number the CryptoBriefing piece anchors on when it says the two labs are 'closing in' on the software giant. Salesforce, by contrast, runs at roughly $38 billion. The comparison flatters the AI labs on growth rate and flatters Microsoft on durability.
Why it matters
For crypto, this is a narrative event more than a flow event. There is no token, no ETF, no on-chain settlement layer attached to Anthropic or OpenAI revenue. What there is: a market that has spent the last year and a half pricing AI-adjacent tokens off exactly this kind of headline. Render, Bittensor, Akash, Fetch, and the broader decentralized-compute cohort have repeatedly traded on OpenAI product launches, funding rounds, and revenue leaks.
The $115B print matters because it validates the underlying compute-demand thesis those tokens are built on. If two labs alone are generating that much recurring revenue, the GPU spend backing it is enormous, and the argument for decentralized alternatives gets a fresh data point. It also matters for the reverse trade. Every dollar of AI ARR that lands with a centralized US lab is a dollar that isn't flowing to a permissionless protocol.