The $115 Billion Blind Spot: What the Anthropic-OpenAI ARR Headline Does Not Say

The number is precise. The source is not. Crypto Briefing reports that Anthropic and OpenAI have a combined annual recurring revenue of $115 billion. That figure, if accurate, places two private companies in the same revenue tier as Microsoft's commercial cloud division. It is a staggering claim. It is also, at present, an unverified one. Ledger balances do not lie; they only wait. And this ledger is waiting for a second signature. The report, published in 2026, uses the word "accelerates" to describe growth. It provides no split between the two companies. It offers no margin data. It cites no customer retention figures. It is a single data point from a crypto-focused outlet, repackaged as a market milestone. The absence of corroboration from Bloomberg, Reuters, or The Information is not an oversight. It is a signal. Context matters here. The AI industry has spent the last three years transitioning from a narrative-driven market to a revenue-driven one. The 2022 Terra-Luna collapse taught us that narrative without collateral is just a story. The same principle applies to AI revenue claims. A $115 billion ARR figure, if real, confirms that enterprise AI spending has moved from experimental budgets to operational line items. If inflated, it represents a systemic mispricing of the entire sector. The variance between those two outcomes is the entire story. My own audit experience informs this skepticism. In 2020, I traced a DeFi yield aggregator's hidden backdoor by following anomalous liquidity patterns on-chain. The project had raised millions on the strength of its marketing. The code told a different story. The same discipline applies here. We do not have the code. We have a press release. The burden of proof rests on the claimant. Let us parse the available data. A combined ARR of $115 billion implies monthly revenue of approximately $9.6 billion. Based on historical ratios, OpenAI likely contributes roughly $80 billion, with Anthropic at $35 billion. These are extrapolations, not facts. The report does not provide the breakdown. This is the first red flag. Any serious financial disclosure would include per-company figures. The omission suggests either a lack of access to granular data or a deliberate aggregation to create a more impressive headline. The second issue is growth quality. The report states growth is accelerating. It does not specify the driver. Is it API call volume? Enterprise subscriptions? Consumer products? Each has different margin profiles and retention characteristics. API revenue is volume-based and subject to price competition. Enterprise subscriptions are stickier but have longer sales cycles. Consumer revenue is volatile and sentiment-driven. Without this breakdown, the $115 billion figure is a single number with no analytical weight. Valuation math compounds the problem. At a 10-20x ARR multiple, the combined valuation range is $1.15 trillion to $2.3 trillion. This is the kind of range that tells you nothing. The lower bound is a mature SaaS company. The upper bound assumes hyper-growth with expanding margins. The truth is somewhere in between, but the report provides no data to narrow the range. Volatility is not risk; opacity is. This is opacity. The competitive landscape adds another layer of uncertainty. The report frames OpenAI and Anthropic as a duopoly. That framing ignores Google DeepMind's Gemini series, which has been commercially aggressive since 2024. It also ignores the Chinese AI ecosystem, which operates in a parallel market with its own infrastructure. The duopoly narrative may be accurate, but it is not supported by the data in this report. It is an assertion dressed as analysis. There is also the question of strategic investor influence. Microsoft is OpenAI's largest backer. Amazon is Anthropic's largest backer. Both are also their primary cloud providers. The revenue these companies generate from their strategic investors is not disclosed. If a significant portion of the $115 billion comes from Microsoft and Amazon purchasing AI services for their own cloud customers, the ARR figure is less impressive than it appears. This is not fraud. It is a structural conflict of interest that inflates the headline number. Now, the contrarian angle. The bulls have a point. The scale of the number, even if imprecise, indicates that AI has crossed a threshold. Enterprise customers are not buying AI because of FOMO. They are buying it because it reduces costs and increases output. The 2025 regulatory clarity provided by MiCA and other frameworks has legitimized the technology for institutional adoption. The revenue is real, even if the exact figure is disputed. The question is not whether AI is a viable business. It is whether the current valuation premium is justified by the underlying economics. The infrastructure implications are significant. A $115 billion ARR implies annual compute costs of $23-34 billion, assuming a 20-30% cost ratio. This level of spending means both companies are locked into long-term GPU supply agreements. It also means they are investing heavily in custom silicon to reduce their dependence on NVIDIA. The energy consumption at this scale is measured in terawatt-hours. This is not a software business. It is a utility business with software margins. The distinction matters for valuation. What the report does not address is the profitability path. High ARR does not equal high profit. Both companies are likely still in a strategic loss phase, spending heavily on compute, talent, and sales. The gross margin trajectory is the key metric to watch. If inference costs are declining faster than demand growth, margins will expand. If not, the business model faces structural pressure. The report provides no data on this. It is the most important missing piece. The regulatory dimension is also absent. The EU AI Act imposes compliance requirements on high-risk AI applications. Financial services, healthcare, and legal sectors are all subject to these rules. Compliance costs are not trivial. They affect sales cycles and margin structures. A $115 billion ARR company is a regulatory target. The report does not mention this. It is a significant omission. Hype evaporates; receipts remain. The receipt for this claim is a single article from a crypto-focused outlet. It is not enough. The market needs official confirmation from OpenAI and Anthropic. It needs audited financial statements or at minimum a detailed breakdown of revenue by segment. Until then, the $115 billion figure is a hypothesis, not a fact. The takeaway is not that the number is false. It is that the number is unverified. In a bull market, unverified numbers become the basis for investment decisions. That is how capital gets destroyed. The discipline of verification is not optional. It is the only thing separating an investor from a gambler. The data will eventually confirm or deny this claim. The question is whether the market will wait for the answer or price in the narrative first. Based on my experience, the market rarely waits. That is the risk.