The $115B ARR Mirage: Why the OpenAI-Anthropic Duopoly Numbers Demand an Audit

The data shows a single headline: OpenAI and Anthropic now command a combined $115 billion in annual recurring revenue. Crypto Briefing dropped this figure into the AI discourse without a second source, without a breakdown, and without context. For traders, this is not a signal to chase momentum. It is a red flag demanding verification. A $115 billion ARR figure, sourced from a crypto outlet rather than audited financial statements, is a data point with an unknown variance. I have seen this pattern before, and it rarely ends with the market getting the full picture. Context: The market treats this number as confirmation that AI has entered a hyper-growth phase. If you split the ARR based on historical ratios, OpenAI holds roughly $80 billion and Anthropic claims $35 billion. Those figures would place both companies among the top-tier global SaaS entities, with Microsoft's commercial cloud at a run rate above $100 billion and Salesforce trailing behind at $37.5 billion. The word "accelerates" in the report suggests 2026 growth is outpacing 2025, which aligns with the classic crossing of the chasm from early adopters to mainstream enterprise procurement. But the source matters. Crypto Briefing is not Bloomberg or The Information. It is a publication primarily focused on digital assets, and its sudden pivot to AI revenue reporting warrants suspicion. When a non-specialist outlet publishes a massive industry metric without naming its data origin, I treat the number as an unverified claim. Core: Based on my 2018 experience auditing fifteen ICO smart contracts, I learned that the gap between a whitepaper's promise and deployed bytecode is where value is lost. Apply that same skepticism here. The ARR figure is the headline; the audit trail is the missing breakdown. The critical question is revenue quality. Is this ARR composed of committed multi-year enterprise contracts, or does it include a significant portion of "internal revenue" from strategic investors? Microsoft and Amazon are not just investors; they are the respective cloud backbones for OpenAI and Anthropic. If a percentage of this ARR represents credits or internal consumption from these hyperscalers, the number is inflated by an unknown factor. In 2020, when I automated my DeFi rebalancing during the gas fee spike, I preserved 92% of my capital because I had pre-coded rules for efficiency. The market now needs that same standardized framework for AI revenue. Run the numbers: at a 10x to 20x ARR multiple, the combined valuation range lands between $1.15 trillion and $2.3 trillion. The market is pricing in a future where these two firms become the primary tollbooths for enterprise AI. But a valuation built on a single-source ARR figure without a margin or retention breakdown is a position with undefined tail risk. The 2021 NFT floor collapse taught me that when data is opaque, you set a stop-loss, not a larger position. Contrarian: The conventional narrative is that a $115 billion ARR proves AI is a necessity, not a luxury. I challenge that assumption. The data does not reveal whether this growth is driven by ROI or by competitive anxiety. Enterprises may be purchasing AI capabilities defensively, afraid of falling behind rivals, rather than because the tools generate immediate positive returns. This is not sustainable demand; it is a hedge. The risk is that when CFOs complete their post-implementation audits, they will find the ROI is negative and pull back on renewals. Additionally, the absence of Google DeepMind from the conversation is a glaring omission. If Google's Gemini revenue is not even in the same reporting category, it signals a widening gap. But if the source selectively excluded it, the narrative is skewed. The market is treating this as a two-horse race, but the finish line is profitability, not revenue. High ARR with high losses is just a deferred bankruptcy. The 2022 Terra Luna collapse demonstrated that a protocol with massive transaction volume can be insolvent in minutes. Revenue scale is not a proxy for solvency. Takeaway: The $115 billion ARR is a headline, not a thesis. The next move is to watch for primary source confirmation. If OpenAI or Anthropic file statements or leak financials that verify the split, the market can price the duopoly with more precision. If the data remains a ghost from a crypto outlet, it is noise. Ledger books, not feelings, settle the debt. Audit the code, then audit the intent. Liquidity dries up when confidence breaks. Until this number is verified, I am treating the AI trade with the same caution I used when auditing those ICO contracts: verify the bytecode, then commit the capital.