There is a peculiar moment in every technological cycle when the architecture of the internet stops being about packets and protocols, and becomes about jurisdictions and jail time. Texas Attorney General Ken Paxton just proposed something that, if it lands, will not merely reshape data center operations—it will rewrite the entire legal grammar of how America treats Chinese technology and artificial intelligence. The proposal is a double-barreled weapon: a federal ban on Chinese tech inside U.S. data centers, and criminal liability for 'harmful AI.' Let me be clear about what this actually is: it is not a regulation. It is a declaration of war on a supply chain, wrapped in a constitutional crisis, with a side of criminal law that could send a CTO to prison for an algorithm's mistake.
I have spent the last decade auditing tokenomics and watching narratives harden into law. In 2017, I ran Python simulations on ICO whitepapers and watched hype collapse under the weight of arithmetic. In 2020, I watched DeFi Summer's liquidity fairy tale turn into a liquidity mirage. In 2021, I asked who owns the soul of crypto art while the Punks minted. Now, in 2026, I am watching something far more consequential: the weaponization of legal ambiguity itself. Paxton's proposal is not a technical document. It is a narrative weapon designed to shift the Overton window on both China and AI in a single stroke. And the crypto industry, which prides itself on decentralized resilience, is standing directly in the blast radius.
The context here matters more than the headline. The United States already has an export control framework—the Export Administration Regulations (EAR), the International Emergency Economic Powers Act (IEEPA), and Title VII of the Defense Production Act. These are blunt instruments for specific items: chips, software, certain technologies. What Paxton is proposing is a categorical ban, not a list-based one. He wants to define 'Chinese technology' in the context of data centers as a class of prohibited goods. That is a fundamentally different legal animal. It moves from 'these specific items are restricted' to 'this entire category of origin is banned.' This is the difference between a scalpel and a sledgehammer, and the sledgehammer is aimed at the foundation of the modern internet.
The most dangerous word in this entire proposal is 'harmful AI.' There is no federal statute in the United States that currently criminalizes AI behavior. The White House AI Executive Order was administrative guidance. State-level laws are patchwork. Paxton wants to create a federal crime where the actus reus is not a specific action, but a vague outcome—'harm.' This is the legal equivalent of making 'bad weather' a crime. Who defines harm? Is a trading bot that causes a flash crash 'harmful'? Is an AI that recommends a stock that loses value 'harmful'? If the definition is strict liability, then any AI system that produces a negative outcome—regardless of intent or negligence—could trigger criminal charges. This would not just chill innovation; it would freeze it solid.
Let me take you inside the mechanics of this proposal because the devil is in the definitional details. The 'Chinese technology' ban would presumably cover hardware (servers, chips, storage), software (operating systems, virtualization platforms), and managed services. But where does the supply chain begin and end? If a Taiwanese chip is manufactured in a Chinese-owned fab, is it Chinese tech? If a U.S. company uses an open-source library originally developed by a Chinese developer, does that taint the entire stack? The compliance burden would be astronomical. Every data center operator would need to prove a negative—that their multi-tiered, globally-sourced supply chain contains zero components attributable to China. This is not due diligence; it is a logistical nightmare. I have audited supply chain narratives in crypto, and I can tell you that even simple token flows are nearly impossible to trace perfectly. Physical supply chains are exponentially more complex. The proposal, as written, would require an entirely new industry of forensic supply-chain auditors.
I keep returning to a phrase I wrote in 2021 during the NFT mania: 'Where the code meets the chaotic human heart.' That is precisely what this proposal does—it forces a rigid, binary legal framework onto the messy, emergent reality of AI systems. An AI model is not a static piece of code; it is a dynamic, evolving entity shaped by its training data, its environment, and its users. To assign criminal liability to such an entity, or to its creators, is to misunderstand the nature of the technology. It is like holding a book publisher criminally liable for every interpretation a reader might have. The legal system is not equipped to handle the probabilistic, non-deterministic nature of modern machine learning. And yet, here we are, watching a state Attorney General propose exactly that.
The counter-narrative here is one that the crypto industry, of all communities, should understand deeply: the problem of centralized control. We have spent years arguing that decentralization reduces systemic risk. But this proposal is a stark reminder that the legal system is the ultimate centralized authority. No smart contract can protect you from a federal indictment. No DAO can vote to override a criminal statute. The 'code is law' mantra collapses when the actual law is written by state officials with geopolitical ambitions. This proposal is a direct challenge to the premise that technology can exist outside the purview of nation-state power. And it is a challenge that crypto, for all its ideological posturing, is not prepared to answer.
Let me bring this down to the practical level, because that is where the pain will be felt. Data center operators will face a compliance burden that could increase operating costs by 10-20%. They will need to hire compliance officers, supply chain auditors, and AI safety testers. They will need to implement blockchain-based traceability systems (ironic, isn't it, that the technology we love might become the tool of its own regulatory capture). They will need to restructure their governance to include AI ethics committees and potentially a 'Chief AI Compliance Officer' reporting directly to the board. The costs will not be borne equally. Large enterprises with deep pockets will survive; smaller players will be squeezed out or forced to consolidate. This is not innovation policy; it is industrial consolidation policy disguised as national security.
There is a deeper, more corrosive effect at play. The proposal, if enacted, would accelerate the bifurcation of the global internet. China's Data Security Law and Cybersecurity Law already require critical information infrastructure operators to purchase secure and trustworthy network products and services—a clear preference for domestic technology. The U.S. proposal would create a mirror-image restriction. The result is a 'dual blockade' where the two largest digital economies are building parallel, incompatible technological ecosystems. This is not de-risking; this is decoupling. And decoupling has a cost that extends far beyond the data center. It will affect everything from cross-border data flows to the ability of researchers to collaborate on AI safety. It will make the 'trust layer for AI'—a concept I have been exploring for my 'Autonomous Economies' report—a pipe dream.
What about the 'harmful AI' criminal liability? Let me paint a scenario. Suppose an AI-driven trading algorithm, deployed by a decentralized autonomous organization, inadvertently causes a market disruption that results in losses for thousands of retail investors. Under Paxton's proposal, who goes to jail? The DAO has no legal personhood. The developers who wrote the initial code are scattered across jurisdictions. The node operators who ran the network are anonymous. The legal system, designed for individual actors, will find itself chasing ghosts. The most likely outcome is that the first person indicted will be a scapegoat—a mid-level engineer or a founder with a U.S. passport who happened to be in the wrong place at the wrong time. This is not justice; it is a lottery.
The legal challenges to this proposal would be immediate and severe. The 'Major Questions Doctrine' from West Virginia v. EPA (2022) requires that administrative agencies have clear congressional authorization when addressing issues of vast economic and political significance. If the administration tries to implement this via executive order under IEEPA, it will almost certainly face a legal challenge. The courts will ask: did Congress really intend to grant the executive branch the power to ban an entire category of technology and create a new federal crime? The answer, almost certainly, is no. But that is the genius—or the danger—of Paxton's approach. He is not just proposing a law; he is setting a precedent for how state-level actors can inject their agenda into federal policy. Even if this specific proposal fails, the conversation has been shifted. The Overton window has moved.
Let me also consider the international law dimension, because this is where the proposal becomes truly global. China could challenge a U.S. ban at the WTO as a violation of Most-Favored-Nation treatment. It could invoke its Anti-Foreign Sanctions Law to retaliate against U.S. companies. This would create a cycle of mutual retaliation that harms global supply chains. The United States could also attempt to impose 'secondary sanctions' on third-country companies that use Chinese technology in their data centers—a form of long-arm jurisdiction that would be deeply controversial. We are not just talking about a U.S.-China issue; we are talking about a fundamental reshaping of the global technology trade regime. The stakes are enormous.
But I want to challenge the prevailing narrative on the other side as well. The tech industry, including much of crypto, has reflexively condemned this proposal as protectionist and xenophobic. That is a lazy take. There are legitimate national security concerns about supply chain integrity and the potential for backdoors in critical infrastructure. The SolarWinds attack and various other supply chain compromises have demonstrated that the threat is real. The question is not whether to address these concerns, but how. A blanket ban is the crudest possible instrument. A more sophisticated approach would involve: risk-based assessments, third-party audits, transparency requirements, and international cooperation on supply chain security standards. But that kind of nuanced policymaking does not make headlines, and it does not win elections. A ban does.
Rewriting the ledger, one story at a time. That is what I have tried to do in my career. And this story is one of the most important I have ever covered. Because what we are witnessing is not just a regulatory proposal; it is a fundamental test of whether our legal systems can adapt to the age of AI and globalized supply chains. The answer, so far, is not encouraging. The legal system is designed for a world of clear causation, individual actors, and national borders. AI and blockchain operate in a world of emergent behavior, distributed agency, and global networks. The mismatch is not a minor inconvenience; it is a systemic failure waiting to happen.
Let me give you a concrete example from my own experience. In 2020, I helped build a narrative-tracking bot for liquidity mining rewards during the Berlin ETHGlobal hackathon. It was a crude piece of software, but it worked. Now imagine that same bot, years later, being used by a hedge fund to automate trading decisions. The bot makes a decision based on flawed data, causing a loss. Under Paxton's proposal, the developers of the bot could face criminal charges for 'harmful AI.' But the bot was not designed to be harmful; it was designed to track narratives. The harm came from its application, not its creation. This distinction—between the tool and its use—is fundamental to any reasonable legal framework. Yet the proposal, as described, seems to blur it entirely.
The compliance burden will be particularly acute for the crypto industry. We have built our entire ethos on the idea of permissionless innovation. Now we are facing a proposal that would require permission for everything. Data centers that host blockchain nodes would need to prove they are 'China-free.' AI agents that interact with DeFi protocols would need to pass safety tests that do not yet exist. The regulatory uncertainty alone could drive innovation offshore—not to China, but to Singapore, the UAE, or Switzerland. This would be a self-inflicted wound for American competitiveness.
I want to be clear about what I am not saying. I am not saying that all regulation is bad. I am not saying that AI should be completely unaccountable. I am saying that the specific proposal on the table is dangerously broad, technically naive, and legally fraught. It conflates two separate issues—supply chain security and AI safety—into a single political gesture. It offers a false choice between national security and innovation, when in fact both are achievable with more careful, targeted policymaking.
Let me now address the timing. This proposal comes at a moment when the AI narrative is already at a fever pitch. The convergence of AI and crypto—what I have been calling the 'Autonomous Economies'—is creating new forms of value transfer and decision-making that challenge existing legal categories. AI agents are starting to hold wallets, sign transactions, and even negotiate with each other. This is happening without clear legal frameworks for liability, accountability, or dispute resolution. Paxton's proposal is a response to this vacuum, but it is a response that would fill the vacuum with a sledgehammer. It would criminalize the very experimentation that is necessary to understand these new systems.
There is a deeper, more uncomfortable truth here. The legal system is not the only institution struggling to adapt. The media, too, has a tendency to simplify complex issues into binary narratives. 'China bad, AI scary, America must act' is a compelling story, but it is not an accurate one. The reality is that technology supply chains are deeply intertwined, and AI is a tool that can be used for good or ill. The real challenge is not to ban or criminalize, but to build the institutional capacity to understand, monitor, and govern these systems effectively. That requires investment in regulatory expertise, not just in enforcement powers.
I have been writing about the intersection of code and human chaos for over a decade. This proposal is a case study in that intersection. It is an attempt to impose order on chaos through the bluntest instrument available: the criminal law. But the chaos of AI is not a bug; it is a feature. It is the result of systems that are complex, adaptive, and unpredictable. Trying to criminalize that unpredictability is like trying to criminalize the weather. It is not just futile; it is destructive.
So what should the industry do? The answer is not to retreat into libertarian slogans about 'code is law.' The answer is to engage seriously with the legitimate concerns that underpin this proposal, and to offer better alternatives. The crypto industry should be at the forefront of developing supply chain transparency tools—using the very blockchain technology we champion. We should be building AI safety frameworks that are not just about avoiding harm, but about ensuring accountability and redress. We should be educating policymakers about the technical realities of both supply chains and AI systems. We should be doing this not because we fear the state, but because we understand that the social contract requires us to address legitimate concerns.
There is a window of opportunity here. Paxton's proposal is extreme, and it may well fail in the courts or in Congress. But if the industry simply dismisses it as absurd, we will have missed the lesson. The lesson is that the era of regulatory neglect is over. The era of criminal liability for technology is beginning. And the only way to prevent that from being a disaster is to shape the conversation now, before the laws are written. This is not a time for complacency; it is a time for engagement.
I am reminded of a line from my 2022 series 'Rebuilding from Ashes,' when I interviewed founders who had survived the bear market. The ones who thrived were not the ones who ignored the market; they were the ones who adapted to it. The same principle applies here. The regulatory bear market is coming. The question is not whether we will adapt, but how quickly we can do so.
Let me close with a forward-looking thought. The next 12-18 months will be decisive. We will see whether this proposal enters the congressional legislative process, whether it is implemented via executive order, and whether it survives judicial review. We will see whether the industry mobilizes to offer constructive alternatives, or whether it simply fights a rearguard action. We will see whether the United States chooses decoupling or managed interdependence. And we will see whether the rest of the world follows suit or charts a different path.
I do not know the answers to these questions. But I know that the choices we make now will shape the technological landscape for decades. The ledger is being rewritten. The question is who will hold the pen. The chaotic human heart is still beating, and it is beating inside the machine. We have a choice: we can let fear dictate our response, or we can use our creativity, our curiosity, and our commitment to human flourishing to build a better path. I know which path I choose. Where the code meets the chaotic human heart, there is still room for hope.
Rewriting the ledger, one story at a time. This is the story of a proposal that could define an era. Let us write it wisely.