I spent the last 48 hours auditing a ghost. Not a smart contract—those are at least honest about their bytecode—but a financial claim so detached from reality that it almost feels like a coordinated stress test of our collective skepticism. The headline reads: “Anthropic’s revenue run rate exceeds $65B ahead of IPO.”
Let that sink in. $65 billion. That is more than the annual revenue of Salesforce, Adobe, and Cisco combined when they were at a similar stage of maturity. For a company that—according to every credible source I could triangulate—was doing roughly $4–5 billion in annualized revenue in mid-2025, this number is not just optimistic; it is a category error. It is the kind of figure that would require Anthropic to have captured the entire enterprise AI market, plus half of the cloud computing market, plus a printing press for money.
But here is the thing: the article is not a joke. It was published on a crypto-adjacent news site, and it is being shared in Telegram groups and Discord servers where people are making investment decisions based on it. In my years as a smart contract auditor, I learned that the most dangerous bugs are not the ones that break the code—they are the ones that break the assumptions everyone takes for granted. This is a bug in the information layer of the market. And it is spreading.
Context: The Real Anthropic
Anthropic is not a small player. Founded in 2021 by former OpenAI researchers, it has raised over $10 billion from Amazon, Google, and a constellation of venture firms. Its Claude models—especially Claude 3.5 Sonnet and Claude 4—consistently rank in the top tier of LLMs, competitive with GPT-4o and Gemini 2.5. The company’s differentiation is not just technical; it is philosophical. The “Constitutional AI” alignment method, the Responsible Scaling Policy, the deliberate pace of deployment—these are the hallmarks of a team that takes AI safety seriously, even if it costs them market share.
In 2024, Anthropic’s ARR was estimated at around $1 billion. By mid-2025, that number had grown to perhaps $4–5 billion, driven by enterprise contracts (SAP, Zoom, and others) and a growing API developer base. The valuation correspondingly surged from ~$18 billion in early 2024 to ~$60 billion in a March 2025 funding round, with reports of a new round potentially pushing it past $100 billion. That is impressive. That is real. But it is a far cry from $65 billion in revenue.
To put it in perspective: OpenAI, the market leader, was reported to have an ARR of around $13 billion in mid-2025. If Anthropic were truly doing $65 billion, it would be generating five times the revenue of its closest competitor—a competitor that has a massive consumer brand (ChatGPT), a decade of head start, and a valuation of $300 billion. Such a scenario would imply that enterprise AI adoption has already surpassed the entire software industry’s spending on cloud infrastructure. It has not. Not by an order of magnitude.
Core: The Anatomy of a Narrative Inflation
I have seen this pattern before. In 2021, during the DeFi summer, I audited a protocol that claimed to have $2 billion in total value locked. The actual on-chain data showed $200 million. The difference was not a bug—it was a feature. The team had planted a fake LP token that inflated the TVL calculation, knowing that the number would be picked up by aggregators and media. The narrative became the reality for long enough that they could exit their positions.
This is not a direct analogy, but the mechanism is similar. The $65 billion figure—whether it is a typo (6500 million vs 65 billion) or a deliberate fabrication—serves the same purpose: to create a “reality distortion field” around Anthropic’s growth trajectory. In a market where FOMO drives capital allocation, a headline that says “Anthropic’s revenue exceeds $65B” is more valuable than a nuanced analysis that says “Anthropic’s revenue is growing rapidly but still below $5B.” The former gets shared. The latter gets ignored.
But the damage goes beyond clickbait. When institutional investors, family offices, and even retail traders see this number, it recalibrates their expectations. They start to believe that AI is already a mature, trillion-dollar industry, and that any company not growing at 10x per year is a laggard. This creates a perverse incentive for every AI startup to inflate its metrics, leading to a systemic bubble that will eventually burst—taking real innovation down with it.
During my years running the Open Ledger education platform in Nairobi, I learned that the first casualty of hype is trust. We saw it when a local crypto project promised 30% monthly returns and collected $2 million from farmers who could not afford to lose it. The promoters were not necessarily malicious—they believed their own hype. But the result was the same: real people lost real money because they trusted a narrative that was too good to be true.
Contrarian: The Case for Skepticism as a Service
Here is the counter-intuitive angle: the $65 billion claim, even if false, reveals something true about the market’s appetite for AI narratives. It tells us that the demand for “AI is the next everything” story is so strong that publishers are willing to print numbers that defy basic arithmetic. And that demand, in turn, signals that the real opportunities in AI are not in the hype cycle—they are in the infrastructure and services that survive the hype.
If you are an investor, the smartest move is not to chase the next Anthropic. It is to build the tools that allow everyone to verify claims like this. A decentralized oracle for AI company revenue? A cryptographic proof of revenue that can be audited by third parties? These are the kinds of infrastructure that will outlast any single company’s valuation.
From my experience auditing ERC-20 standards, I know that the most valuable code is not the flashy smart contract—it is the boring Oracle that ensures the data feeding into the contract is accurate. The same principle applies to AI investing. The most valuable asset is not the AI model itself; it is the ability to distinguish real growth from narrative inflation.
Takeaway: The Ledger Never Lies
Anthropic is a real company with real technology and real revenue. But until we have a transparent, verifiable way to audit those numbers, we are all trading on stories. The blockchain taught me that trust is not a substitute for verification. The same lesson applies to AI.
So here is my challenge to the industry: build a public, cryptographically signed revenue registry for AI companies. Let the code speak. Let the on-chain data replace the press release. Because when the next $65 billion headline drops, I want to be able to trace the moral code behind every token—and every dollar.
Tracing the moral code behind every token. Building libraries where others build empires. Walking away from the hype to find the soul. Ethics is not a feature; it is the foundation. Community over capital, always. Listening to the silence between the blocks. Preserving the human story in digital ledgers.
