Last week, I sat in a Lagos coffee shop, scrolling through the same leaked CFO memo that had been making rounds on X. The numbers were staggering: OpenAI’s annualized revenue run rate had jumped 35% in just a few months, hitting an estimated $36.2 billion. Enterprise business was growing at 50% year-over-year. 20 million weekly active users. And then there was that strange data point about Anthropic—$11.6 billion in Q2? That number didn’t sit right. It felt like a typo, or a deliberate disinformation signal. But the real question for me, as someone who has spent years building crypto education in Nigeria, wasn’t about the accuracy of numbers. It was about the underlying trust infrastructure.
I’ve been in the crypto space long enough to know that when a centralized entity reports explosive growth, the first thing I check is not the revenue but the code. OpenAI’s success is a story about centralized AI services—models running on Microsoft Azure, controlled by a single board, with no transparency into how data is used, how models are trained, or how outputs are verified. The 2000 weekly active users? They’re trusting a black box. The enterprise clients paying 50% more? They’re betting on a promise of security that cannot be independently audited.
Here’s the core of what I see: OpenAI’s growth is a perfect case study for why blockchain-based verification layers are not just nice-to-have, but essential for the next phase of AI adoption. The numbers are impressive, but they mask a fundamental vulnerability. Every dollar flowing into OpenAI is a dollar that could be tracked on-chain if the model were verifiable. Every API call could be cryptographically attested. Every training data provenance could be recorded on an immutable ledger. This is not a theoretical exercise—projects like Vana, Gensyn, and the Verifiable Truth Initiative I’m leading are already building these primitives.
Let me pull back the curtain on the technical reality. The 35% acceleration in revenue run rate is largely driven by enterprise customers moving from experimentation to production. But what happens when a banking client in Lagos rotates their entire customer support system to GPT-4o, and the model hallucinates a fraudulent transaction? The bank has no recourse—no audit trail, no way to prove that the output came from OpenAI’s model and not from a corrupted prompt. Self-sovereign identity and zero-knowledge proofs could solve this. I’ve seen the same pattern in DeFi: when users trust a smart contract without verifying the code, they get rekt. The same principle applies to AI.
Now, the contrarian angle. You might think that OpenAI’s closed-source model is its greatest competitive advantage. After all, secrecy protects proprietary training data and prevents model theft. But I argue the opposite: closed-source is the weakest link in the trust chain. As enterprise clients demand regulatory compliance (GDPR, HIPAA, the EU AI Act), they will increasingly require verifiable computation. If OpenAI cannot prove that its model outputs are computed correctly and that data is not leaked, it will lose the next wave of enterprise deals to open-source alternatives that run on decentralized compute networks. The Lightning Network taught us that a half-dead routing protocol can never scale—OpenAI’s current architecture is the Lightning Network of AI: centralized, fragile, and too expensive to trust.
Finally, the takeaway. The next billion-dollar AI company will not be the one with the best model, but the one that can prove its model’s integrity on-chain. OpenAI’s growth is a wake-up call for the crypto community: we have been too focused on trading tokens and not enough on building the infrastructure for AI verification. The numbers are real, but the trust that backs them is not. Trust the process, but verify the code. If we don’t build the verification layer, the AI boom will collapse under the weight of its own opacity.