AI Agents Are Choosing Denationalized Money Over Stablecoins

Denationalization of money has been a crypto talking point for years — but Sylvia To, vice president at Bullish Capital Management, thinks we've been looking at the wrong end-user. In a March 2026 institutional analysis, To argues that autonomous AI agents — not ideologically motivated humans — may be the most likely adopters of genuinely decentralized money, and that Bitcoin, for all its price volatility, is the only serious candidate on the table.

Hayek, Satoshi, and the Triad Nobody Connects

The argument starts in 1976. That's when Friedrich Hayek published his case for competitive private currencies in 'Denationalisation of Money' — the idea that money, like any product, should be discovered and adopted through market choice rather than decreed by states. His checklist for "good money" reads like a Bitcoin spec sheet: non-state issuance, rule-based supply schedules, voluntary global adoption, resistance to capture, and settlement without institutional permission.

To connects this directly to Satoshi's original design philosophy — privacy, decentralization, censorship resistance — and to Vitalik Buterin's March 2026 observations on X about building "sanctuary technologies" that create shared digital spaces with no owner, enabling interdependence that cannot be weaponized. The denationalization of money thesis isn't abstract anymore. It has a live experiment running on-chain.

Bitcoin, To argues, sits in a special category. Not because it's flawless in its current form, but because it's the first monetary network to plausibly meet Hayek's central requirement: money introduced through a pathway governments cannot easily stop. The price volatility? That's the market deciding what a credibly scarce, ungoverned asset is worth. Birth is messy.

Bitcoin's value may fluctuate, but its rule set is unusually legible. Its issuance is not negotiated. Its core properties do not depend on a board decision, a regulator's discretion or the solvency of a nation.

Why Stablecoins Are Not the Answer

Here's where To gets uncomfortable for a lot of people in this industry. Stablecoins are crypto's most successful product by most measures — fast, programmable, cheap to move across borders. But they don't denationalize money. They do the opposite.

A Bitcoin stablecoin pegged to the dollar imports the dollar's inflation rate, its surveillance infrastructure, its sanction regime, its banking chokepoints, and its regulatory priorities. The network may be open, but the reference asset is the same old sovereign instrument. As To puts it, stablecoins move on open rails but carry a state payload.

The framing is sharp: "If bitcoin is denationalization, stablecoins are nationalization with better UI." Call it provocative if you want, but the logic holds. A stablecoin issuer is a centralized entity subject to government pressure. That entity can freeze funds, censor transactions, or collapse under regulatory action. The architecture looks decentralized; the trust model is anything but.

That said, To isn't dismissing stablecoins as useless. They have real utility today. The concern is that they may inadvertently build the infrastructure for tighter financial control — a perfect on-ramp that becomes a cage.

What Happens When the Economy's Main Participants Aren't Human?

This is the part that most crypto coverage missed entirely. The real thesis isn't about Hayek or cypherpunks — it's about the coming shift in who is actually transacting.

Humans are emotional, politically driven, and short-term oriented. Our monetary systems reflect that perfectly. We trade long-term stability for short-term relief constantly, then express shock when crises compound. But autonomous AI agents — software that purchases services, data, compute, API calls, and inference tools through continuous micropayments — don't have that problem.

Apps are increasingly being designed for agents using frameworks like Model Context Protocol, and the near-term future To describes is one where agents are the dominant transacting parties in significant portions of the economy. What will those agents optimize for? Not brand narratives. Not political alignment. They'll want machine-readable transaction metadata, instant programmable finality, low overhead, composability, censorship resistance — because uptime is a feature — and predictable monetary rules that models can optimize against.

In other words: agents will gravitate toward money that behaves like infrastructure. And when they evaluate a stablecoin, the question they'll ask is clinical: what is the failure mode of the issuer? What's the censorship risk under stress? What's the settlement risk if the issuer faces regulatory action? Bitcoin answers those questions differently than any dollar-pegged token does.

Maybe Hayek's 'new money' was never meant for humans — at least not first. Maybe the pathway that governments 'can't stop' isn't a mass political movement. Maybe it's AI agents who operate at machine speed, indifferent to national identity, optimizing for reliability.

Is This the Pathway Governments Can't Stop?

The most provocative idea in To's analysis is also the one that gets buried in a single paragraph near the end. Crypto adoption has always been framed as a human story — retail investors, institutions, nation-states adding Bitcoin to reserves. But what if the decisive adopters are machines?

AI agents operating at machine speed have no national identity to protect, no political habit to break, no fear to overcome. They don't care about the narrative around Bitcoin being digital gold or a store of value. They care whether the monetary rails are reliable, legible, and resistant to interference. Bitcoin currently provides that more credibly than any alternative.

To's conclusion: when that tipping point arrives — when agents become dominant transacting parties — denationalization of money won't feel like a philosophical triumph. It will be an inevitable engineering outcome. Not propelled by ideology. By raw machine necessity.

Meanwhile, traditional finance continues its own quiet integration. Institutional players including the NYSE owner ICE and Morgan Stanley have been making strategic moves in crypto, and Kraken's reported access to Federal Reserve infrastructure signals an industry that is, simultaneously, both decentralizing and institutionalizing. The contradiction is real. Whether it resolves in favor of Hayek's vision or against it may depend on which type of participant — human or machine — tips the balance first.

Chart of the Week: Kamino's OnRe Market

Kamino's OnRe market grew 80% to nearly $90M in total value locked over 30 days, cementing its position as the primary liquidity layer for OnRe's on-chain reinsurance protocol. The growth gives users exposure to a $480 billion real-world insurance vertical through $ONyc — a tokenized insurance asset used as collateral.

There's a divergence worth watching though. The fundamental growth in OnRe is sharp and real, but the native $KMNO token tells a different story — the KMNO/SOL pair has dropped 16% over six months, pressured by broad market weakness and 13 million monthly token unlocks representing 0.13% of total supply. Protocol growth and token price are running in opposite directions. That's a dynamic that tends to resolve eventually — just not always in the direction TVL bulls expect.