The $115B AI Fiction: Why Market Narratives Precede Reality

The most dangerous number in technology right now is one that almost certainly does not exist. A recent Crypto Briefing flash news item claims that Anthropic and OpenAI's combined Annual Recurring Revenue has topped $115 billion, closing in on Microsoft. The claim is extraordinary. It is also, based on every verifiable data point in the market, almost certainly false. The consensus is wrong because it ignores the cost of attention. The real story here is not AI revenue. It is the manufacturing of market narratives and the structural vulnerability of capital that believes them. This is not a hit piece on a single outlet. It is a case study in how data becomes unmoored from reality and why, for institutional allocators, the discipline of verification is the only true alpha. The core claim defies basic arithmetic. Public estimates from reputable financial media place OpenAI's ARR at approximately $3.7 billion in 2024. Anthropic's is around $1 billion. That is a combined figure of $4.7 billion. The reported number is 25 times higher. To reach $115 billion in ARR, these two private companies would need to have added over $100 billion in revenue in a single year, a growth trajectory that has no precedent in the history of commerce. The article provides no source, no methodology, no breakdown between the two companies. It presents a headline number and a comparison to Microsoft, then stops. History does not repeat, but it rhymes. This is the 2017 ICO whitepaper all over again: a compelling narrative, a total absence of diligence. What is the actual competitive reality? Microsoft's commercial cloud revenue is approximately $160 billion annually. That includes Azure, Office 365, and LinkedIn. The $115 billion figure would position these two AI startups at roughly 72% of Microsoft's entire cloud business. The mental image is absurd. OpenAI and Anthropic combined have a fraction of Microsoft's employee count, data center footprint, and enterprise sales force. The claim ignores the capital intensity of the AI business. Inference costs alone for serving hundreds of billions of dollars in revenue would require a scale of GPU deployment that does not exist in the public cloud today. My 2017 due diligence checklist would have rejected this whitepaper in the first paragraph. Why would a crypto-focused outlet publish this? The answer is structural. Crypto media needs to connect the AI story to its own thesis of digital value transfer. An AI narrative that is merely growing is less interesting than an AI narrative that is disrupting the existing tech order. The "closing in on Microsoft" framing is designed to trigger a specific emotional response: the old order is falling. This is a story designed to move capital, not to inform it. In 2020, during the DeFi yield crisis, I saw the same pattern. Yields that were unsustainable were presented as the new normal. Capital flowed in. The correction was brutal. Volatility is the fee for admission to the future. But a false narrative is a fee for admission to a future that never arrives. Let me be clear on what is true. The AI revenue trend is real. OpenAI and Anthropic are growing rapidly. The enterprise adoption of AI is accelerating. The competition between AI-native companies and traditional software giants is genuinely significant. Microsoft's Azure AI business is growing at triple-digit rates. But the actual gap remains wide. OpenAI's ARR is perhaps $4 billion, not $40 billion. The distance between $4 billion and $100 billion is not a difference in magnitude. It is a difference in kind. In 2024, I negotiated direct prime brokerage relationships for my fund's institutional clients. The data I could verify determined my allocation. The data I could not verify determined my caution. Code is law, but capital decides who writes it. Consider the strategic error in the reporting. The article deliberately merges OpenAI and Anthropic into a single alliance to create the "challenger" narrative. In reality, these are two companies in fierce competition. They compete for enterprise clients, for top-tier AI talent, for the attention of the same venture capital dollars. Their safety narratives are positioned as differentiators. Their API pricing strategies are a direct arbitrage against each other. The attempt to frame them as a united front against Microsoft is a logical fabrication. Microsoft itself is OpenAI's largest investor and cloud partner. The relationship between the three is far more complex than a simple competition. The narrative of "them vs. Microsoft" is a simplification that serves a story, not an understanding. For investors, this data point is dangerous. If someone believes the $115 billion figure, they will extrapolate valuations based on a 10x price-to-sales multiple, arriving at a combined valuation of over $1 trillion. That is a fantasy. The real market cap of these two companies is closer to $200 billion combined, with an average P/S ratio of about 40-50x. The article, if taken at face value, would make these companies look wildly undervalued. The reality is that they are priced for significant growth, and the current valuation already reflects a high degree of optimism. The opportunity here is not in believing the narrative. It is in finding the companies that benefit when the narrative is corrected. When the market realizes the actual numbers, the AI software valuations might face a correction. But the infrastructure layers, data centers, chip manufacturers, and energy providers, will still benefit from the secular growth in compute demand. I have seen this cycle before. When the 2022 Terra-Luna collapse happened, I was short and buying distressed assets. The panic was not the end. It was a signal for a reallocation of capital. What are the risks here? First, the market sentiment can be moved by false data. A headline like this can trigger a brief rally in AI-related tokens or stocks. But it is the kind of move that fades quickly. Second, regulators are beginning to look at misinformation in financial markets. If a crypto outlet publishes a fabricated revenue number for private companies, it could attract scrutiny that harms the entire sector. Third, the AI industry itself is threatened by this kind of hype. When the expectations are set impossibly high, the inevitable miss will lead to a brutal correction. That correction will be felt not just by the AI companies, but by the broader technology sector. For institutional investors, the signal is clear. Demand transparency. Demand data with sources. Demand methodology. The opaqueness of private company finances is a feature, not a bug. It allows narratives to flourish in the absence of evidence. In my 27 years of observing this industry, I have seen one common thread: the hype cycle always leads to the capital destruction cycle. The smart money is not on the headline. It is on the structures that will benefit from the underlying reality, not the fiction. I am thinking about the data centers, the energy providers, and the specialized hardware. The real trend is the AI adoption. The narrative is just noise. The signal to track is not the $115 billion figure. It is the actual usage data. Look at OpenAI's API volume growth. Look at Anthropic's enterprise customer count. Look at the Azure AI growth rate as a proxy. These are the data points that are real and verifiable. In the short term, the most likely outcome is that neither OpenAI nor Anthropic will issue a formal correction, because they do not usually comment on third-party estimates. But within the next 3-6 months, when the next round of funding disclosures comes out, the actual ARR will be public. That will be the moment of truth. The gap between the claimed $115 billion and the actual $10 billion will be the size of the market's delusion. When that gap is revealed, the correction will be severe. I want to be clear: AI is a generational opportunity. The companies building these models are creating massive value. The trend of AI replacing and augmenting human work is irreversible. I am not a skeptic about the AI revolution. I am a skeptic about the data that purports to measure it. The difference between the two is the difference between a successful investment and a catastrophic one. My entire career has been a series of filters. In 2017, I rejected 95% of ICOs due to flawed tokenomics. In 2020, I pulled capital from unsustainable DeFi yields. In 2022, I profited from the panic. The same discipline applies now. The story is compelling. The data is not. Code is law, but capital decides who writes it. The narrative of $115 billion is a story, not a fact. The only useful takeaway is this: the AI market is growing, but the market's expectations have grown faster. Volatility is the fee for admission to the future. The future is real, but the admission ticket is often fake. Verify before you buy. The price of attention is the price of the mistake.