Hook
Last week, Crypto Briefing ran a headline that made me spit out my espresso: “AI infrastructure boom drives Anthropic valuation toward $1.2T by year-end.” As someone who navigated the 2018 bear market by running three side-projects on Layer 2 scaling, I’ve learned to smell narrative-driven valuations from a mile away. $1.2 trillion for a model company that has yet to prove a sustainable unit economy? That’s not an investment thesis — it’s a meme with a decimal point.
Context
Anthropic is the startup behind Claude, the AI model known for its “Constitutional AI” safety alignment. It has raised over $7 billion from investors including Google and Spark Capital, and its cloud partner is AWS. The broader AI infrastructure boom — think NVIDIA’s H100 shortage, massive data center builds, and hyperscaler capex — has lifted all boats. But in my decade of watching protocol economies, I’ve seen this script before: the shovel sellers get rich, the gold miners struggle with margin compression and narrative whiplash.
From 2020 to 2022, we witnessed a similar “infrastructure boom” in crypto: Layer 1 blockchains raised billions on the promise of “world computer” usage. Yet many of those tokens now trade at fractions of their peak. The common thread? Infrastructure is necessary but not sufficient for value capture. The code is cold, but the community is warm — and Anthropic lacks the decentralized community that could make its governance layer resilient.
Core: Why $1.2T Collapses Under Scrutiny
Let’s apply a protocol analyst’s lens to this valuation claim. First, we must ask: who captures the value in AI infrastructure? The answer is vertically integrated hyperscalers (Microsoft, Google, Amazon) and chip manufacturers (NVIDIA). Anthropic is a consumer of this infrastructure — it buys compute from AWS to train and serve Claude. Its gross margins are squeezed by rising GPU costs, just as DeFi protocols in 2021 saw their TVL metrics inflated by token incentives that masked underlying unprofitability.
Second, the valuation narrative conflates macroeconomic tailwinds with company-specific moats. In blockchain terms, this is like arguing that because total value locked in DeFi is growing, every new lending protocol should be worth $10 billion. We know that’s false: only protocols with deep liquidity, governance resilience, and user stickiness survive. Anthropic’s moat is its safety research and “Constitutional AI” — but safety is a feature, not a business model. OpenAI proved that users will accept less safety for more capability (GPT-4o’s uncensored mode vs. Claude’s refusal to answer).
Based on my audit experience with crypto governance loopholes, I see a parallel in Anthropic’s centralized control structure. The company has absolute power over Claude’s behavior, aligned with its own ethical interpretation. This is the opposite of the decentralized ethos we champion in blockchain. We are not just users; we are the protocol. But with Anthropic, users are renters in a walled garden. If the garden’s rules change (e.g., new safety constraints or pricing changes), users have no exit except to switch to a competitor. That’s a weak value capture mechanism.
Third, the $1.2T number implies a multiple that defies basic finance. Even if Anthropic captured 50% of the AI software market (currently ~$200B enterprise spend), that would imply a revenue of $100B. At a 12x revenue multiple (generous for a growth company), that’s $1.2T. But Anthropic’s revenue is estimated at under $1B in 2024. To reach $100B in revenue by year-end would require growth of 10,000% in six months — mathematically impossible without buying all of AWS.
The infrastructure boom benefits Anthropic only if it can lower its cost structure relative to peers. But NVIDIA’s dominance means everyone pays similar prices. The real value is being captured by chipmakers and cloud providers. In crypto, we call this “hydraulic stability” — value flows to the point of least resistance. Currently, that point is the hardware and energy layer, not the model layer.
Contrarian: Could Anthropic Become the Ethereum of AI?
Now let me play devil’s advocate, because contrarian thinking is the heart of honest analysis. Perhaps the $1.2T narrative is not about current revenue but about future platform value. If Anthropic’s API becomes the standard for safe, verifiable AI agents — and if they leverage blockchain-based compute markets (like Akash or Render) to decentralize their infrastructure — they could create a flywheel similar to Ethereum’s smart contract platform.
Picture this: Anthropic open-sources its safety layer as a smart contract. Developers build autonomous agents that must comply with “Constitutional AI” enforced on-chain. Oracle networks verify model behavior. Value flows to the protocol through fees or tokens. This vision is exactly what I’m working on now with verifiable AI datasets on-chain. If Anthropic pivots from a company to a protocol, its valuation could indeed reflect network effects rather than simple revenue.

But that’s a big “if.” Today, Anthropic is a centralized corporation, not a DAO. Its “constitution” is written by a handful of executives. It has no token, no on-chain governance, and no user base that controls the protocol. For the valuation to reach $1.2T, investors must believe Anthropic will transform into something it currently is not — a decentralized, community-owned protocol. In my experience, narrative shifts of this magnitude in crypto happen only after a clear catalyst, like Uniswap’s airdrop or Ethereum’s Merge. I see no such catalyst for Anthropic.
Takeaway
From hype cycles to hydraulic stability: the lesson of both blockchain and AI booms is that value accretes to infrastructure that is scarce and community-owned. Anthropic may be a great AI company, but $1.2T is a fantasy built on a conflation of macro trends with micro merit. As builders, we should resist the temptation to equate “infrastructure boom” with “every player wins.” The code is cold, but the community is warm — and Anthropic’s community is not yet warm enough to justify that price tag. We are not just users; we are the protocol. And this protocol needs a more honest audit.