Silence is the first vote in a true consensus. When I read Crypto Briefing’s headline — "Anthropic and OpenAI’s combined ARR tops $115B, closing in on Microsoft" — I felt the kind of silence that precedes a hard fork. Not the silence of agreement, but the silence of disbelief. I’ve spent years auditing smart contracts, and I know the smell of a reentrancy bug. This number reeks of one.
Let me be clear: the combination of OpenAI and Anthropic’s annualized revenue reaching $115 billion is not just improbable — it is mathematically absurd against any publicly available data. The Information and Bloomberg estimate OpenAI’s 2024 ARR at around $3.7 billion, Anthropic at roughly $1 billion. That’s a combined $4.7 billion, not $115 billion. The difference is 24x. To put it in blockchain terms, this is not a rounding error; it’s a 51% attack on the truth.
Context: The media’s moral hazard
The article in question, published by Crypto Briefing — a crypto-native outlet — offers no source, no methodology, no breakdown. It simply states the number and draws a comparison to Microsoft’s cloud revenue. In my experience designing governance for MakerDAO, I learned that any proposal without a transparent audit trail should be treated as a rug pull. The same applies to financial journalism. We are seeing a new form of information asymmetry: centralized media outlets, even those within crypto, can inject unverified data into the market, influencing everything from token prices to institutional allocations.
Why would they do this? One plausible motive: to bridge the AI hype narrative into the crypto audience. Crypto Briefing’s readers are often retail investors looking for the next big trend. A $115 billion ARR figure, however false, creates a sense of urgency and FOMO. It aligns with the bull market’s euphoria, where technical flaws are masked by soaring valuations. I’ve seen this pattern before — during the 2017 ICO mania, when projects claimed millions in revenue without a working product. The DAO hack taught me that code is not law, but neither is a press release.
Core: Deconstructing the mirage
Let me walk through the numbers using the same forensic approach I applied to the reentrancy vulnerabilities in The DAO. Microsoft’s commercial cloud ARR in 2024 was approximately $160 billion. OpenAI + Anthropic at $115 billion would mean two pre-IPO companies, with fewer than 10,000 employees combined, are generating 72% of Microsoft’s cloud revenue. That implies a revenue per employee of over $10 million per year — a figure that would make Goldman Sachs blush. Even the most optimistic analyst estimates for OpenAI’s 2025 ARR hover around $10 billion. The $115 billion claim is a ghost in the machine.
But the deeper issue is not the number itself; it’s the absence of verification. In the decentralized world, we have on-chain oracles, zk-proofs, and quadratic voting to ensure data integrity. For AI companies, there is no equivalent. Revenue is self-reported, unaudited, and often conflated with total contract value or future commitments. This is DeFi’s oracle problem at scale: a single source of truth can be manipulated, and when it is, the entire market suffers.
During my time as a Senior Researcher at a Tallinn-based cybersecurity firm, I led a post-mortem of The DAO hack. We identified 14 critical flaws, but the most profound lesson was that technical efficiency without ethical governance leads to societal harm. The same applies here. The $115 billion claim is not just a mistake; it’s a failure of ethical code auditing in the media. We need a new standard: every revenue claim from AI companies should be accompanied by a verifiable on-chain proof of revenue, or at least a cryptographic commitment that can be cross-referenced with payment rails.
Contrarian: What if the number were true?
Let’s play the contrarian game. Suppose, against all evidence, the combined ARR did hit $115 billion. What would that mean? It would mean that AI has achieved a level of enterprise adoption that renders traditional SaaS obsolete overnight. It would also mean that OpenAI and Anthropic’s profit margins are astronomical, since their operating costs (compute, labor) are a fraction of that revenue. Under this scenario, the two companies would be worth over $1.5 trillion, surpassing all but the largest tech giants.
But here’s the blind spot: even if the number were true, it would represent a catastrophic centralization of power. Two private companies controlling the dominant AI infrastructure, with no decentralized governance, no on-chain transparency, and no user ownership. This is the opposite of the values I advocate for. In my 2020 work with MakerDAO, I designed quadratic voting to prevent whale dominance. Here, whales are not just token holders; they are the AI companies themselves, making decisions without community consent. The irony is that the crypto media is amplifying this centralization narrative under the guise of a bullish headline.
Consensus requires patience, not speed. The speed of this headline — viral, unverified, and emotionally charged — is a symptom of a market that has forgotten the lessons of 2022. After the FTX collapse, I retreated to Hiiumaa island for six weeks and wrote a manifesto about the hollow promise of yield. Now I see a parallel: the hollow promise of AI revenue. We are building a new financial system on top of unverified claims, repeating the same mistakes.
Takeaway: The need for a new oracle
What can we do? I propose a decentralized data verification protocol for AI revenue claims. Think of it as a Chainlink for corporate financials, but with zk-proofs to preserve confidentiality. Imagine a world where every AI company publishes a zero-knowledge proof of its revenue, proving the sum without revealing individual customer details. This is feasible with today’s technology. I have already started designing such a framework for Tallinn’s AI startup hub, integrating ZK-proofs into agent wallets. The same principle applies to corporate ARR.
Winter teaches what spring forgets. We are in a bull market, but the euphoria will fade. When it does, the projects and media that lied will be left exposed. The only way to build trust that survives the next winter is to embed verification into the fabric of our systems. Silence is the first vote in a true consensus. Let’s vote with code, not with headlines.