The $115B ARR Mirage: A Forensic Audit of the OpenAI-Anthropic Revenue Claim

A headline flashed across crypto media this week: Anthropic and OpenAI’s combined annual recurring revenue has topped $115 billion, closing in on Microsoft. The number is a statistical impossibility. I spent the last 48 hours dissecting the claim, and the ledger doesn’t lie—but this headline does. The source is Crypto Briefing, a publication that often trades in hyperbole. No primary data, no methodology, no breakdown. Just a single figure and a comparison designed to provoke. This is not analysis; it is narrative engineering. Let me walk you through why the number fails every basic sanity check. First, the context. ARR, or annual recurring revenue, is a standardized metric for subscription-based software companies. It excludes one-off deals, professional services, and variable usage charges. For AI labs like OpenAI and Anthropic, ARR is typically derived from API calls, enterprise licenses, and ChatGPT subscriptions. Public reporting from credible outlets—The Information, Bloomberg, and the companies’ own investor communications—places OpenAI’s 2024 ARR at roughly $3.7 billion. Anthropic’s is estimated at $1 billion. Combined, that is $4.7 billion. The claim of $115 billion is 24 times higher. That is not a rounding error; it is a fabrication. Where could such a number originate? One possibility is a unit mistake. The original might have said $11.5 billion, but even that is 2.4 times the credible estimate. More likely, the author conflated a long-term revenue forecast with current ARR, or used “total contract value” from multi-year deals, which inflates recurring revenue figures. In my 2017 forensic audit of ICO contracts, I saw similar distortions: teams reporting token sale allocations as revenue, or multiplying a one-time grant by five years. This is the same logical fallacy, applied to AI. Let me test the claim against Microsoft’s own numbers. Microsoft’s commercial cloud revenue—which includes Azure, Office 365, and Dynamics 365—was about $160 billion in fiscal 2024. The headline suggests OpenAI plus Anthropic are at 72% of that scale. But these two companies have a combined workforce of less than 5,000 employees. Microsoft has over 200,000. For OpenAI and Anthropic to generate $115 billion in recurring revenue, they would need to capture roughly 10% of the global enterprise software market. That is not “closing in on Microsoft”; that is conquering the entire sector in two years. The probability is zero. What does this tell us about the industry? The headline is not a data point; it is a symptom. We are in a bull market for AI hype, and crypto media is eagerly feeding the fire. The purpose is not to inform but to align AI’s explosive narrative with crypto’s speculative appetite. I have seen this playbook before. In 2021, I analyzed 150 NFT collections and found that 80% of their volume was wash trading. The same “data as marketing” virus infects AI reporting. The ledger doesn’t lie, but the people writing the headlines do. Now, the contrarian angle. Even if the specific number is garbage, the underlying trend is real. AI revenue is growing rapidly—Azure AI alone grew over 100% year-over-year. Enterprise budgets are shifting from experimentation to deployment. But that growth is still an order of magnitude smaller than the traditional SaaS giants. Salesforce, Adobe, and ServiceNow each generate over $20 billion in annual revenue. OpenAI and Anthropic are still catching up. The headline’s implication—that AI-native companies are about to dethrone Microsoft—is premature and misleading. More troubling is the deliberate conflation of OpenAI and Anthropic as a single entity. These are fierce competitors. OpenAI has a privileged relationship with Microsoft; Anthropic is backed by Amazon and Google. Their combined ARR is a meaningless aggregation, designed to create a false equivalence. If you break them apart, OpenAI’s $3.7 billion is still less than 2% of Microsoft’s cloud revenue. The gap is not closing; it is expanding, because Microsoft’s AI revenue is growing faster than either company’s standalone revenue. What should investors and analysts do? Ignore the headline, but watch the underlying signals. Look at API call volumes, enterprise customer counts, and net revenue retention. These are the true leading indicators. Also, be wary of any AI data that comes from a crypto outlet without a primary source. Cross-reference with audited financials, which will be available once OpenAI or Anthropic files for an IPO. Until then, treat every ARR figure above $20 billion as a red flag. My takeaway is simple: the ledger doesn’t lie, but the storytellers do. In a bull market, the noise gets louder. Your edge is verification. The next time you see a stunning revenue claim, ask three questions: Who is the source? What is the methodology? And does it align with on-chain or audited reality? If the answer to any of those is vague, move on. The real opportunity lies in the gap between narrative and fact—where the data still has a pulse.