Hook
Last week, a report from Ramp—a corporate expense management platform—landed with a thud in the crypto and AI press. According to Ramp, Anthropic has surpassed OpenAI in enterprise AI adoption across the United States. The claim, amplified by outlets like Crypto Briefing, sent ripples through valuation discussions: if true, Anthropic’s market standing could justify a higher price tag in its next funding round. But as a DAO Governance Architect who has spent years auditing whitepapers and governance proposals, I’ve learned that the most dangerous data is the one that sounds right but can’t be audited. Who verifies the verifier? Ramp is a centralized fintech company, not a neutral research institute. Before we celebrate or invest, we need to ask: where is the proof?
Context
Enterprise AI adoption is the new gold rush. From code generation to financial document processing, large language models are being embedded into the core workflows of thousands of companies. The race between Anthropic and OpenAI is not just about model quality—it’s about who owns the enterprise customer relationship. For the crypto ecosystem, this matters immensely. AI agents that interact with smart contracts, DAO governance, and DeFi protocols rely on models that are both capable and trustworthy. If Anthropic truly leads in enterprise adoption, it could shape the infrastructure of the next generation of decentralized applications. Yet the data behind this claim comes from a single source: Ramp’s internal billing records, covering only its own customer base. No methodology, no sample size, no time window—just a headline. The crypto community, built on principles of transparency and verifiability, cannot afford to accept such opaque signals. We need to dissect the report, understand its biases, and demand a better standard for market intelligence.
Core
Let’s start with what Ramp’s data actually captures. Ramp is a spend management platform that tracks corporate expenses—API subscriptions, SaaS fees, and cloud marketplace purchases. If a company buys Anthropic’s Claude API directly or pays for Claude Pro, that transaction appears in Ramp’s system. But here’s the catch: a significant portion of OpenAI’s enterprise sales flow through Microsoft Azure, as part of bundled Azure OpenAI Service offerings. Those costs may be recorded under “Azure cloud infrastructure” rather than a separate line item for “AI tools.” In other words, Ramp’s data might systematically undercount OpenAI spending, while overcounting Anthropic spending because Claude is often purchased as a standalone line item by developer teams. This is a classic measurement bias—one that I’ve seen in countless DAO treasury audits. The same flaw can turn a leading indicator into a misleading mirage.
During my time auditing DeFi protocols, I encountered a similar situation in 2020. A popular analytics dashboard claimed that a certain DEX had captured 80% of liquidity in a particular token pair. The data came from the DEX’s own API. When I cross-referenced with on-chain data, I found that the DEX was counting only its own pools, while ignoring the liquidity on competing platforms that used different routing. The 80% claim was true within its own universe, but false in the broader market. Ramp’s report may suffer from the same tunnel vision. The enterprise AI market is not just about direct API purchases; it includes embedded models through platforms like Copilot, Vertex AI, and Bedrock. If Anthropic is leading only in the “standalone AI tool” segment, while losing in the “bundled enterprise suite” segment, the narrative changes entirely.
Now, let’s examine the competitive dynamics. The parsed analysis of the original article highlights that Anthropic’s Claude 3.5 Sonnet and Claude 4 have received strong developer praise for code generation, long-context understanding, and safety features. This aligns with my own observations from the DAO community: several projects I’ve collaborated with have switched from GPT-4 to Claude for governance document analysis and smart contract auditing. The reason is not just capability—it’s the perception of reliability. Anthropic has positioned itself as the “safe” AI, a brand that resonates with compliance-conscious enterprises. That reputational advantage can translate into faster adoption among mid-sized tech companies, which are precisely the customers that Ramp serves. However, OpenAI still holds the largest enterprise customer base globally, thanks to Microsoft’s sales force and the ubiquity of Office 365. The question is not who is ahead, but how the gap evolves.
From a valuation perspective, the report’s claim could serve as a catalyst for Anthropic’s next funding round. If investors believe that Anthropic is winning the enterprise race, they may bid up the company’s valuation from the reported $60 billion in 2024 to over $100 billion. But here’s the contrarian truth: valuation is a story, and stories need evidence. The article itself admits that the report lacks verification, sample details, and competitive benchmarks. As someone who has written about the ethics of empty vests, I know that a narrative without substance is a house of cards. The crypto market has seen too many projects pump on unverified metrics—total value locked (TVL) that was double-counted, user counts that included bots, and now, AI adoption claims that may be selectively sampled. We must resist the urge to amplify headlines without demanding the underlying data.
Code is law, but people are the soul. This is one of my core signatures. It reminds us that trust in code is not enough; we must also trust the people who produce the data. Ramp may have the best intentions, but its report is not a smart contract—it cannot be audited on-chain. The solution is not to dismiss the report, but to build a better system. Imagine a future where enterprise AI adoption metrics are published as verifiable on-chain proofs, using zero-knowledge computations to aggregate spending data from multiple sources without revealing proprietary information. DAOs could vote on which metrics to trust, and the community could collectively validate the numbers. This is the kind of governance architecture I believe in: not top-down announcements, but bottom-up verification.
Don’t govern the exit, govern the entrance. This signature applies here too. Instead of waiting for reports to influence valuations, we should govern how such reports enter the market. If Ramp wants its data to be taken seriously by the crypto community, it should open-source its methodology, share the raw aggregated data (within privacy bounds), and allow independent auditors to run the same queries. Until then, every interpretation is provisional. I have seen too many projects rush to judgment based on a single data point, only to be burned when the full picture emerges. Enterprise AI adoption is a marathon, not a sprint. The winner will be determined by sustained revenue growth, customer retention, and real-world impact—not a single report from a single platform.
Contrarian
Let me now offer a counter-intuitive angle: what if the report is actually correct, and Anthropic is genuinely leading? Even then, the way this information is presented does a disservice to the ecosystem. By framing the data as a definitive “leading position,” the report discourages deeper investigation. It creates a false sense of certainty that can lead to over-allocation of resources—both financial and intellectual. The real risk is not that the report is wrong, but that it’s incomplete. The crypto community prides itself on being skeptical of centralized authorities. Yet here we are, accepting a corporate expense report as gospel. The greatest threat to decentralization isn’t censorship; it’s the illusion of authority. We should treat every piece of market intelligence as a hypothesis, test it against multiple sources, and only then let it inform our decisions.
Moreover, the report’s timing is suspicious. Ramp is actively building its own AI agent, Ramp Intelligence. By publishing a report that highlights the leading AI provider, Ramp positions itself as a neutral observer while simultaneously riding the AI hype wave. This is not a conflict of interest per se, but it is a pattern I’ve seen before: a company uses its data to shape the narrative in a way that benefits its own brand. The crypto community should be particularly sensitive to this, given our history of influencer-driven markets. We must demand separation between data providers and narrative builders. The ideal solution is a decentralized network of data oracles, each contributing a piece of the puzzle, with no single entity controlling the story.
Takeaway
The Ramp report is a useful signal, but it is not a verdict. It tells us that enterprise AI adoption is real, and that Anthropic has carved out a meaningful niche. However, the path forward requires us to rebuild the metrics infrastructure from first principles. I envision a future where every important market metric—from AI adoption to DeFi TVL—is published on-chain, with cryptographic proofs, governance tokens, and community verification. Until that day comes, we must remain vigilant. The next time you see a headline claiming someone is “leading,” ask yourself: who measured it, how, and can I verify it? Code is law, but people are the soul. Let’s ensure that the people governing our data are as transparent as the code they use.
Listen more than you code. This is my final reminder. The data is speaking, but we must listen carefully—not just to the headlines, but to the hidden assumptions, the missing variables, and the silent biases. Will we build the tools to verify, or will we keep trusting the gatekeepers?