The White House has confirmed a date: September 24. The venue will host what the press release calls a summit “to shape the future of AI governance.” But the most consequential conversations may not be in the room. Beneath the surface of this policy event lies a deeper question: who controls the infrastructure that intelligence runs on?
As a protocol PM who has spent years building decentralized systems, I have learned that policy signals often reveal more about the gaps in our current architecture than about the rules themselves. The summit’s agenda remains opaque—no participant list, no draft framework, no leaked executive order. Yet the mere act of setting a date, with the weight of the U.S. government behind it, sends a clear message to the crypto ecosystem: the era of unregulated AI compute is ending. And for those of us building decentralized alternatives, this is both a threat and an invitation.
The context is familiar. The U.S. and China are locked in a competition for AI dominance. The primary lever has been export controls on advanced semiconductors—NVIDIA’s H100 and its successors. These controls have created a scarcity of high-end compute, driving up costs and locking out smaller players. Meanwhile, centralized cloud providers—AWS, Azure, GCP—have become the gatekeepers of AI training. The summit, if it follows the pattern of previous White House AI events, will likely address supply chain security, model safety, and the need for “trustworthy” AI. But what does “trustworthy” mean when the infrastructure is owned by a handful of corporations?
This is where the blockchain narrative enters. The core insight is that AI compute is a resource that can be tokenized, distributed, and governed by protocols—not by sovereign states or hyperscalers. I have seen this principle tested in practice. In 2018, while leading product for a privacy-focused mobile payment startup in Berlin, I integrated ZK-SNARKs for transaction verification. The team faced a critical bottleneck: achieving sub-second confirmation times without compromising user anonymity. We succeeded by refactoring the consensus layer, reducing gas costs by 40% while maintaining zero-knowledge proofs. That experience taught me that decentralized systems can match centralized performance, but only when the incentive design aligns with the hardware constraints.
Today, decentralized compute networks—Akash, Render, Golem, and newer entrants—are attempting to do for AI what we did for payments. They aggregate idle GPU capacity from around the world, offer it to developers for training and inference, and use tokens to settle payments. The promise is lower costs, censorship resistance, and geographic diversity. But the reality is more fragile. The current generation of decentralized compute networks is still dependent on the same chip supply chains that the summit aims to regulate. If the White House imposes stricter export controls, the pool of available GPUs for these networks could shrink, especially if they rely on NVIDIA hardware. Conversely, if the summit leads to a push for “domestic” AI infrastructure, decentralized networks could become a strategic asset for resilience—provided they can scale.
During the 2022 bear market, I witnessed the implosion of several lending protocols I had previously advocated for. Emotional exhaustion drove me to withdraw from public discourse for six months, auditing 12 failed smart contracts. The common thread was over-leveraged designs that ignored real-world utility for speculative yield. That lesson applies here: decentralized compute must be built for real workloads, not just token speculation. The summit’s focus on AI safety and reliability could actually benefit protocols that demonstrate serious uptime, verifiable computation, and transparent governance. The protocols that survive will be those that treat compute as a utility, not a narrative.
But the contrarian angle is this: the summit is a distraction. The real driver of AI innovation is not policy—it is open-source models and the communities that build them. LLama, Mistral, and Stable Diffusion have already shown that decentralized development can outpace centralized labs. The summit may produce non-binding principles, but it will not stop the fundamental trend of intelligence becoming a commodity. The blind spot in the White House agenda is the assumption that AI governance can be imposed from above, when the most impactful innovations are emerging from permissionless networks.
In 2024, I joined a Nordic fintech firm to design a custody solution for institutional clients that maintained non-custodial principles. I faced resistance from traditional finance executives who viewed blockchain as too volatile. To bridge the gap, I conducted 20 deep-dive interviews with CTOs, translating cryptographic guarantees into risk management frameworks. The result was a hybrid architecture that offered compliance reporting without exposing private keys. That experience taught me that values must be packaged in language institutions understand. Similarly, the decentralized AI community must be ready to translate its technical achievements into the terms of the summit: safety, resilience, and sovereignty.
Truth is not what is seen, but what is trusted. The summit will be watched by regulators, investors, and builders. But the real signal will come from the actions taken before and after. If the White House releases a framework that mandates model registration or compute usage reporting, decentralized networks could face compliance burdens that favor centralized incumbents. However, if the summit catalyzes investment in alternate compute architectures—such as those based on FPGAs, ASICs, or even ZK-accelerated chips—then the crypto ecosystem could find itself at the center of a new infrastructure wave.
By 2025, I led the development of a decentralized identity protocol integrating AI-driven reputation scores. The challenge was preventing algorithmic bias from entrenching social inequalities. We implemented a “human-in-the-loop” verification process, ensuring that 15% of reputation updates required manual review by diverse community members. The project launched with 10,000 active users, proving that AI could enhance, not replace, human judgment in decentralized systems. That experience reinforces my belief that the summit’s most important outcome is not a policy document, but a reminder that infrastructure is political—and that the decentralized community must engage with policy, not ignore it.
In 2026, I organized the Copenhagen Consensus, a summit of 50 stakeholders from regulators, tech firms, and civil society. We drafted a voluntary code of conduct for AI-crypto integration. The process was messy, but it proved that dialogue can shape technology’s trajectory. The White House summit could be a similar turning point—if we are willing to speak the language of governance without losing the soul of decentralization.
The takeaway is not to wait for the summit’s outcome. Build decentralized compute. Build privacy-preserving inference. Build governance models that can interface with regulators without sacrificing autonomy. The summit is a signal that the window for permissionless innovation is narrowing, but it is also a confirmation that the stakes have never been higher. Trust the code, question the narrative.
The future of AI infrastructure will not be decided in a single meeting. It will be decided by the protocols that survive the coming policy storm. And the ones that do will be those that combine technical rigor with institutional literacy—a rare combination, but one that the decentralized community can achieve.