The Unzoned City: What Praxis's Uruguay Announcement Refuses to Disclose

On a gray Seattle morning I read the Crypto Briefing item twice, certain I had skipped a paragraph. Praxis, it said, intends to build an AI-driven city in Uruguay. The first residents are expected in 2027. Two sentences. No whitepaper. No contract address. No land deed, no municipal filing, no funding round, no named team, no repository. I have spent enough years auditing code to learn that absence is more honest than assertion — and here the absence was the loudest thing on the page. I do not write this to mock the ambition. I write it because the genre has a body count of promises, and because the people who will eventually move there deserve a ledger they can actually read. The idea is not new. Sidewalk Labs tried to instrument a Toronto waterfront and withdrew in 2020 after a fight over data governance — the sensors were never the problem; the consent was. Neom promised a linear city backed by sovereign capital and has since been publicly rescaled. Telosa remains largely a rendering. Prospera operated as a special economic zone on the Honduran island of Roatán until the Honduran congress repealed the ZEDE framework in 2022, a reminder that a jurisdiction can be legislated out from under a community overnight. Praxis arrives into that graveyard carrying a different phrase: AI-driven. Where the earlier experiments promised libertarian governance or sustainable urbanism, this one promises intelligence — the city as a system that senses, decides, and adapts. That phrase is why the item landed on a crypto wire and not in an architecture journal. The announcement is not a construction update; it is a capital-formation gesture aimed at a community that buys narrative faster than it buys permits. I say that without cynicism. I have felt the same reflex in myself, in 2021, standing in a cabin outside Seattle calculating contagion risk in leveraged stablecoin positions while everyone around me chased yield. An instrumented city needs three layers, and the phrase dissolves the moment you try to audit it. Sensing: distributed infrastructure — energy telemetry, water, transit, occupancy — with defined data ownership. Decision: the models that allocate resources and, crucially, the objective function they optimize. Actuation: the mechanism by which a model's output becomes a physical service, whether a bus route or a water tariff. None of these are disclosed. There is no architecture diagram, no model card, no open-source repository, and no description of who holds the kill switch when a model misallocates water to the west district and leaves the east dry. I have audited smart contracts with less at stake. In 2017 I spent six months inside MakerDAO's early governance contracts and found a logic flaw in the stability-fee calculation that threatened user solvency. I reported it anonymously on GitHub and the team fixed it, but the experience taught me something I have never stopped using: the specification is the ethics. When a project cannot state its architecture, it is often not being secretive — it simply has not decided yet, and the announcement is doing the deciding on its behalf. Then there is the calendar. A city, even a small one, from land acquisition through permitting, utilities, communications, housing, and occupancy, typically takes five to ten years. Modular construction can compress the build; it cannot compress jurisdiction. You still need a land title, an environmental review, a utility interconnection agreement, and a migration pathway for the people you expect to live there. None are confirmed. If 2027 means anything, it most likely means a small, modular first settlement on land already controlled by the developer — not a city. It is a founding camp with a marketing department. And "residents" is doing quiet work. The first cohort is unlikely to be families with school calendars and fixed mortgages. It is more plausibly remote workers, crypto natives, and the seasonal migration pattern this industry has normalized — people who can leave. A city whose first citizens are people who can leave is a city with an exit built into its constitution. That is worth naming before anyone mints anything. Which brings me to the token. The source discloses none. That is not reassurance; it is a blank. In this genre, the blank is usually filled later — with a land NFT, a membership pass, a governance token, or a staking instrument tied to future rents. I have watched this movie since the ICO era. The pattern is consistent: a physical asset with long horizons and low liquidity gets wrapped in a digital asset with short horizons and high liquidity, and the mismatch is the product until it becomes the loss. After the 2022 collapse I withdrew from public discourse for three months and audited fifty failed protocol post-mortems. The common thread was never the cryptography. It was the absence of governance structures with the standing to say no. Cities are the same organism at a slower tempo. When I worked with three indigenous artists on Tezos in 2021, we raised only fifteen thousand dollars and coded the contracts to guarantee permanent, royalty-free community access, rejecting the standard speculation model. We minted souls, not just tokens — and the trust that project built outlasted every yield farm I watched die that year. Here is where I part from the chorus of skeptics. Almost everyone will attack the AI framing — the vagueness, the buzzword, the missing model cards. Fair. But the more interesting signal is the map pin. Praxis chose Uruguay. Not a charter city carved out of a failed state, not a special economic zone with its own courts, not a libertarian enclave in international waters. Uruguay: stable, rule-of-law, high institutional trust, unremarkable taxes, genuinely boring governance. For a project whose entire pitch is escape from regulation, that is a strange choice. I think it is the most honest thing in the announcement. Choosing a jurisdiction you cannot capture is an admission that you need enforceable contracts — that you intend to be a tenant of the law rather than a sovereign above it. In Honduras, the ZEDEs were repealed and a community that had imagined itself autonomous discovered it had no standing. Sovereignty, it turns out, is not a feature you can ship; it is a relationship you inherit. The thing that makes Praxis credible is the same thing that makes it killable. Openness is not a feature; it is a philosophy — and so is jurisdiction. A project that can be destroyed by a democratic vote has told the truth about where power actually lives. Most of the crypto-city canon has not. That reframes the 2027 date entirely. It is no longer a promise to residents. It is a bet on how quickly a patient institution will move, and on whether the developer can hold land long enough for institutions to catch up. To build in public is to trust the void; to build in a democracy is to trust a court. Both are wagers. Only one of them can be graded. Praxis may build something in Uruguay. A campus, then a district, then — decades on — something that earns the word city. What it has built today is a ledger of absences, and those absences are legible if you know how to read them. Code is poetry, but community is the chorus; a chorus does not move to a place that has not been zoned. The question is not whether the first residents arrive in 2027. The question is whether we will keep grading these projects on the beauty of what they promise, or on the transparency of what they disclose. Truth emerges when the ledger is transparent. And humanity remains the only non-fungible asset — the one collateral no protocol has learned to mint.