When I saw Anthropic’s Q2 numbers—$11.5 billion in revenue, 14x year-over-year growth, and a stunning 70% of that coming from a single product, Claude Code—I felt a familiar pang. It’s the same pang I had in 2017, watching ICOs promise decentralization while their founders held multi-sig keys. The AI industry is centralizing power faster than we can audit it, and crypto is watching from the sidelines. But here’s the uncomfortable truth: the market is rewarding this centralization. And that should scare every true believer in decentralization.
Let me explain. Claude Code is an agentic workflow tool—an AI that autonomously writes, deploys, and debugs code for enterprises. It’s closed-source, controlled by a single company, and deeply integrated into centralized cloud infrastructure. Yet it’s generating $8 billion a quarter. Compare that to the entire DeFi ecosystem, which struggles to retain $10 billion in total value locked across hundreds of protocols. The market is voting with its wallet, and it’s voting for centralized efficiency over decentralized sovereignty.
We didn’t design blockchain to be less efficient. We designed it to be trust-minimized. But the reality is that enterprises don’t care about trust minimization when the alternative is a 10x productivity boost. They care about ROI, audit trails, and SLAs. Anthropic delivers that. Crypto, for all its philosophical purity, struggles to deliver reliable, low-latency agentic workflows that enterprises can stake their reputations on.
Truth in blockchain isn’t about code being law; it’s about understanding that law is always enforced by humans. Right now, the humans are choosing Anthropic over Ethereum. And the data is brutal. Anthropic’s B2B market share hit 34.4%, overtaking OpenAI’s 32.3%. Its annualized revenue run rate is $11.5 billion. Compare that to the entire crypto analytics and infrastructure space, which might generate a few hundred million in combined revenue. We are not even in the same league.
But here’s the core insight that most crypto commentators miss: the technology behind Claude Code is not fundamentally different from what we could build on-chain. The agentic workflow—the ability to decompose a task, execute sub-steps, and verify outcomes—is exactly what DAOs need. Imagine a DAO that doesn’t just vote on proposals but autonomously writes the smart contracts, deploys them, and monitors them for bugs. That’s the vision. But we’re not there yet because our infrastructure is fragmented, our tooling is immature, and our incentives are misaligned.
Let me share a personal experience. In 2020, I lost $15,000 in a yield farming exploit because I trusted a protocol’s marketing more than its code. I spent the next three months reverse-engineering the exploit. That experience taught me that trust is built through transparency, not promises. Anthropic’s Claude Code is transparent in its pricing and results, but opaque in its inner workings. Crypto’s smart contracts are transparent in their code but opaque in their reliability. The market is choosing the former because it delivers measurable outcomes.
Now, the contrarian angle. Maybe the market is right. Maybe centralized agents are simply better for most use cases. After all, we don’t demand that our email provider be decentralized; we demand that it be reliable. The same logic applies to enterprise AI agents. But here’s where crypto’s core value proposition re-enters: sovereignty. When you rely on a centralized agent, you are trusting the company that controls it. That trust can be revoked, manipulated, or exploited. We’ve seen it with AI companies changing their APIs, terms of service, and pricing models overnight. A decentralized agent, by contrast, would be permissionless, censorship-resistant, and owned by its users.
We didn’t build Bitcoin to be faster than Visa. We built it to be more sovereign. The same applies to AI agents. The question is not whether decentralized agents can match the speed of Claude Code today. It’s whether they can offer a value proposition that a subset of users—developers, privacy advocates, certain enterprises—will find compelling enough to pay for. I believe the answer is yes, but only if we start building with the same rigor as Anthropic.
Truth in blockchain isn’t about replacing all centralized systems. It’s about providing alternatives that matter. The rise of Anthropic is a wake-up call. It shows that the market rewards products that solve real problems, regardless of their architecture. Crypto needs to stop preaching and start productizing. We need agentic workflows that are as easy to use as Claude Code, but that run on decentralized infrastructure. We need reliable, auditable, and composable agents that can be trusted by enterprises.
I’ve seen the beginnings of this in projects like Bittensor, which is building a decentralized machine learning network, and Autonolas, which focuses on autonomous agent services. But they are early, and their revenue is negligible compared to Anthropic. The gap is not just in technology; it’s in user experience, onboarding, and go-to-market strategy. Crypto has historically been terrible at marketing to non-crypto-native users. Anthropic is excellent at it.
So what does this mean for the future? I see two paths. The first is continued centralization: AI agents become a utility controlled by a few mega-corporations, and crypto remains a niche for financial speculation. The second is a hybrid: decentralized AI agents emerge that offer unique advantages—like data sovereignty, composability, and cryptographic attestation—and capture a growing share of the enterprise market. I’m betting on the second path, but only if we act with urgency.
We didn’t enter this space to be spectators. We entered to build a more open, equitable, and resilient internet. Anthropic’s success is a mirror showing us what we could achieve if we focused on product over ideology. Let’s not waste this lesson. The next twelve months will determine whether crypto becomes a serious player in the agentic economy or remains a footnote. I know which outcome I’m working toward.
