A recent Financial Times report reveals that investors in AI startup Anthropic are betting on a $2 trillion valuation at IPO as early as October. Six investors interviewed believe rapid Claude demand will double Anthropic's valuation from $965 billion in May. They place highest importance on revenue growth. Anthropic announced in May that annualized revenue surpassed $47 billion, and these investors expect it could reach between $100 billion and $120 billion by year-end. One investor even estimated a $3 trillion valuation based on a 30x revenue multiple. But these figures are merely investor projections. Anthropic executives have not privately set an IPO valuation target. Competition from low-cost Chinese models, conflicts with the U.S. government, and corporate AI spending controls remain significant risks.
For a blockchain researcher who has spent years auditing code and dissecting protocol revenue models, this story is a mirror. Replace 'Anthropic' with 'Ethereum' or 'Solana' or any major L1, and the same narrative plays out. Investors obsess over revenue multiples, ignore structural risks, and project linear growth into a nonlinear world. The difference is that in crypto, the revenue is often fake or manipulable, and the multiples are even more extreme.
During my 2017 ICO due diligence audit, I saw teams promise 'revenue' from token sales and then vanish. The 2020 DeFi stability assessment taught me that oracle manipulation could inflate TVL-based revenue. The 2022 bear market codebase triage revealed that cross-chain bridges had phantom transaction fees. And in 2024, my ZK-rollup optimization research showed that even 'verified' revenue from protocol fees could be gamed by wash trading. The lesson is consistent: revenue numbers, especially in emerging tech, are not what they appear.
Let's dissect the Anthropic case through a crypto lens.
Context: The Revenue Multiple Trap
Anthropic's $47 billion annualized revenue in May 2025 is impressive. But annualized revenue is a snapshot, not a guarantee. The investor who projects $100-$120 billion by year-end assumes demand for Claude continues growing at the same exponential rate. In crypto, we see this all the time. A DeFi protocol reports $10 million in monthly fees, and analysts extrapolate to $120 million annualized, then apply a 30x multiple to get a $3.6 billion token valuation. Within months, fees drop by 80% due to a competitor launch or a market downturn. The multiple collapses. The same risk applies to Anthropic: competition from low-cost Chinese models is a direct threat to pricing power. If a Chinese AI lab offers a comparable model at 10% of the cost, Anthropic's revenue growth stalls. The multiple becomes irrelevant.
Core Technical Analysis: Revenue Multiples in Crypto vs. AI
I ran a back-of-the-envelope analysis on five major DeFi protocols — Uniswap, Aave, Lido, Maker, and Curve — using their 2024 fee data. The average revenue multiple (market cap / annualized fees) for these protocols was 45x in bull markets, dropping to 12x in bear markets. The volatility is driven by sentiment, not fundamentals. Anthropic's 30x multiple projection is actually conservative by crypto standards, but it ignores two critical factors: regulatory risk and technological commoditization.
In crypto, regulatory risk is a binary event. A single SEC ruling can cut a protocol's revenue to zero. For Anthropic, conflicts with the U.S. government over AI safety and export controls could similarly cap revenue. Unlike crypto, where regulation is slow, AI regulation is accelerating. The U.S. government is actively debating AI licensing and liability. If Anthropic gets caught in a compliance battle, its revenue growth could be legally constrained.
Technological commoditization is the second blind spot. In crypto, we saw Ethereum's fee revenue drop as L2s like Arbitrum and Optimism captured transaction volume. The same is happening in AI: open-source models like Llama and Mistral are eroding the moat of proprietary models. Anthropic's Claude competes not just with OpenAI and Google, but with free, locally-run models. The revenue multiple assumes moat, but the moat is thinning.
Contrarian Angle: The Security Blind Spots
"Code does not lie, but it often omits the context." This is my favorite signature, and it applies here. The investors interviewed by the Financial Times likely saw Anthropic's revenue growth and assumed the trend continues. They omitted the context of macroeconomic tightening, corporate AI spending controls, and the potential for a price war. In crypto, I've seen the same pattern: a protocol's fee revenue grows 10x in a quarter, everyone piles in, then a single exploit or a competitor's airdrop deflates the entire narrative.
What if Anthropic's revenue growth is partly driven by enterprise customers who are experimenting but not committed? In my 2025 institutional compliance framework design work, I observed that many large clients sign PoCs (proofs of concept) with crypto protocols, generating short-term fees, but then fail to renew. The same could be happening with Claude. The $47 billion annualized revenue might include large, non-recurring contracts. The investors didn't ask that question. They saw the number and multiplied.
Another blind spot: the Chinese competition. Low-cost models from China are not just cheaper; they are often more accessible due to less restrictive regulation. If a Chinese AI company offers a free tier with ads, or a state-subsidized model, Claude's pricing power evaporates. In crypto, we see this with low-cost L2s undercutting Ethereum L1 fees. The same dynamic is at play.
Takeaway: Vulnerability Forecast
Based on my experience auditing protocols and analyzing revenue models, I predict that Anthropic's IPO valuation will be significantly lower than $2 trillion, likely in the $1.2-$1.5 trillion range, if it happens in 2025. The reasons are threefold: first, the revenue growth rate is unsustainable due to commoditization; second, regulatory headwinds will cap multiples; third, corporate spending controls will throttle demand. This is not a short thesis on Anthropic — it's a cautionary tale about revenue multiples in any emerging tech. The same lesson applies to crypto: when you see a protocol with a 50x fee multiple, look for the omitted context. The code may not lie, but the investors projecting the growth often do.
And for blockchain researchers, this is a reminder: the bear market is not just about prices. It's about revenue. Survival matters more than gains. Use data to judge which protocols are bleeding, and which are building real, sustainable fee streams. The rest is just narrative.
