Axel Quack
Bitcoin rails, Ordinal deeds, and the libertarian urban frontier — where the series left off.
Axel Quack
Citizenship as subscription: the political theory this installment pressure-tests with $525 million of real money.
The series so far: in 2022 I mapped the first wave — city tokens, Wyoming LCAs, DisCOs, and the cooperative experiments (Part 1); in 2025 I followed the Bitcoin turn — Ordinal deeds, Stacks and Rootstock governance, Lugano and Ljubljana (Part 2). Both installments were about retrofitting existing places with crypto rails. The 2026 edition is about the opposite movement: capital that stops asking existing places for permission. One project has raised more than half a billion dollars to found a city from scratch, one earlier attempt died in the Nevada desert, and — the genuinely new variable — artificial intelligence has entered the planning room. This is the state of crypto cities as I see it in 2026.
Praxis: the $525 million network state
Axel Quack
The institutional vocabulary — charter cities, ZEDEs, Próspera — that Praxis is now outbidding.
Praxis — founded by Dryden Brown and Charlie Callinan, with Howard Hughes Corporation founder David Weinreb as vice chairman and the gravitational pull of Peter Thiel's orbit around it — describes itself as an "internet-native nation." The stated goal: a city of roughly 10,000 people, crypto-centric and AI-friendly in its infrastructure, built by a community that forms before the ground is broken. In 2025 the project raised $525 million to make it real, and as of early 2026 its waiting list reportedly stands around 50,000 people — a population five times the size of the city it has not yet sited.
The location history is the most instructive part:
| Phase | Candidate site | Status |
|---|---|---|
| Original pitch | Mediterranean island | Never materialized |
| 2025–26 | Greenland — Brown flew to Nuuk and met officials | Politically infeasible |
| Current | "Atlas" — a defense-focused spaceport city at Vandenberg Space Force Base, California | Active pursuit |
| Ongoing | Caribbean and Latin American options | Referenced repeatedly |
On May 15, 2026, Brown held a press conference at the Praxis "embassy" in SoHo, New York, titled simply: Starting a New Country.
Read through the lens of my Network State essay, Praxis is the first serious capital test of Balaji Srinivasan's sequence: first the aligned community, then the crowdfunded territory, then the recognition. Every city in history has been land-first — soil, then settlers, then charter. Praxis is attempting the inversion: a city whose citizens exist before its land does, whose "constitution" is a brand and a waiting list, and whose diplomacy is conducted from a rented townhouse in Manhattan. Capital before territory is the network-state thesis made literal.
Two critiques deserve standing, because the sovereignty frame cuts both ways. First, the feasibility critique — Wikipedia's editors and much of the urbanist press call the vision unrealistic, and the Greenland episode showed how fast a location search can collide with an actual sovereign's "no." Second, and more interesting for this blog's readers: the governance question is deferred, not answered. A waiting list is not an electorate; a founder is not a constitution. Charter cities from Próspera to Atlas all eventually face the same interrogation — who votes, who exits, and what happens when the developer's roadmap and the residents' rights disagree. Praxis's answer so far is community and momentum. Whether that survives first contact with zoning, water, and elections is the actual experiment.
Sidebar: the city that died of thirst
The cautionary tale the Praxis story needs sits four years in the past. Blockchains LLC bought 67,000 acres in Storey County, Nevada, in 2018 (~$170 million) and proposed a smart city of 36,000 residents and 15,000 homes, wrapped around a radical legal ask: Nevada's "Innovation Zone" concept, which would have let a landowning company form its own local government.
The sequence of the failure is the lesson:
- The county revolted. Storey County passed a resolution opposing "separatist governing control" — local democratic consent, withheld.
- The state couldn't sign off on water rights — the physical constraint no term sheet can waive, in the driest basin in the country.
- The proposal was withdrawn in 2021, and by late 2023 the land was being parceled off — including 2,200 acres to a data-center developer.
Land is buyable; jurisdiction and water are not. Blockchains Nevada sequenced the way cities have always been sequenced — land first, legitimacy later — and discovered that the second item is not for sale. Praxis's inversion (capital and community first, land last) is best understood as a direct response to exactly this failure mode. Whether it clears the same two gates — physical constraints and democratic consent — when it finally picks soil, is the open question of the decade.
The generative city: AI enters the planning room
Here is what is genuinely new in 2026, and what neither earlier installment could have covered: the city is becoming a generative object.
- The research frontier has consolidated. Nature Cities published a 2026 review of generative AI in urban science and practice; an npj Urban Sustainability paper this year demonstrated a generative model that encodes arbitrary 3D city blocks into a unified latent representation and generates urban layouts across 330 North American cities. Scoping reviews now map GenAI onto urban digital twins — the virtual mirrors of whole cities — for autonomous generation of data, scenarios, designs, and 3D models.
- The workflow has shifted from prompts to simulations. Generative design tools take plain-language briefs and return code-compliant layouts; planners then run those candidates through simulation engines modeling traffic, emissions, and social outcomes at neighborhood scale. The practical consequence: approval cycles that took months of manual review can compress into weeks of model iteration.
- The politics are arriving. AI-powered digital twins are being pitched as instruments of participatory planning — citizens simulating proposals before they're built — with the obvious counter-risk: a technocracy where the model's output carries the authority of objectivity, and the assumptions inside it go unaudited.
Notice where this lands in this series' vocabulary. A charter city is a jurisdiction negotiated with a host state. A network state is a community negotiating with geography. A generative city is a negotiation with a model — and models have owners. The sovereignty question of the 2020s ("who holds the keys to the money?") becomes the sovereignty question of the 2030s: who governs the model that governs the city? Praxis, pitching itself explicitly as an AI-friendly city, is the first project where all three layers — network state, charter politics, and generative planning — sit in the same term sheet.
The 2026 map: three paths to a crypto city
Consolidating Parts 1–3, the field now sorts into three strategies:
| Path | Exemplars (2026) | Mechanism | Chief risk |
|---|---|---|---|
| Retrofit | Lugano (Plan B, JAN3), Ljubljana, Miami, Detroit | Crypto rails on existing municipalities — payments, tokens, reserves | Political reversal at the next election |
| Charter | Próspera (suing Honduras for ~$11B), Wyoming DUNA ecosystem | Negotiated legal autonomy inside a host state | Host-state revocation — the $11B lawsuit is the risk, priced |
| Greenfield | Praxis ($525M, Atlas), the late Blockchains Nevada | Capital + community first, territory last | Physics and consent: water, zoning, electorates |
Two context markers from the 2026 political landscape: charter cities appeared in mainstream campaign promises (per the Financial Times on Trump's 2024 platform), and Próspera's arbitration against Honduras has turned the charter model's tail risk into an eleven-digit number. Meanwhile the retrofit path keeps compounding quietly — the Lugano and Ljubljana numbers from Part 2 remain the field's best evidence that crypto integration can show up in GDP rather than headlines.
The selection pressure is legible: retrofit wins on legitimacy, charter wins on speed, greenfield wins on imagination — and only greenfield has yet to produce a single finished block.
Final Thoughts
Three installments, one arc. In 2022 crypto cities were tokens bolted onto municipalities; in 2025 they were Bitcoin rails under serious governance; in 2026 they are capital formations searching for soil, with generative AI redrawing what planning even means. Praxis is the sharpest expression of where this is heading: a city pre-sold to 50,000 citizens who have never seen its land, funded at half a billion dollars, aimed at a spaceport, designed in the latent space of models.
My read, as always through this blog's flag-theory lens: the interesting sovereignty is no longer only about which jurisdiction taxes you — it is about which stack generates you. The retrofit path keeps the stack public. The charter path rents it. The greenfield path tries to found it. Blockchains Nevada shows what happens when you buy the land and skip the consent; Praxis is testing whether you can buy the consent and skip the land. Water, zoning, and electorates will have the final word either way — and that, frankly, is as it should be.
The website and the information contained therein are not intended to be a source of advice or credit analysis with respect to the material presented, and the information and/or documents contained on this website do not constitute investment advice.