I was three hours into a contract read when I found it. A single function, buried in an upgradeable proxy, that let an admin freeze any balance, at any time, without a vote. No timelock. No published multisig threshold. Just a key — and a promise that whoever held it would behave.
That was 2023. By then I'd already learned the hard way what unsecured code does to a community. During the DeFi Summer of 2020 I ran three yield aggregators at once — $2 million in total value locked, zero audits. A minor exploit drained 15% of the liquidity, and the people who'd trusted me wanted answers I didn't have. I wrote a post-mortem instead. I called it "Imperfect Innovation," because that was the honest name for what we'd built. Vulnerability, it turns out, travels further than polish.
So when a headline crossed my feed last week — USDG at $3.2 billion, up 340% year over year — I didn't feel FOMO. I felt the old locksmith instinct. Don't look at the paint. Look at the hinge.
Because here's the thing the bull market keeps skipping past: a fiat-backed stablecoin's competitive advantage has almost nothing to do with cryptography and almost everything to do with who holds the keys, who holds the reserves, and who gets paid when the reserves earn interest. USDG's 340% is a real number. It's also a number that tells you almost nothing about whether the thing underneath it is durable.
I want to take that apart — slowly, at the contract level, the way I'd want someone to take apart my own failures.
The Least Glamorous, Most Consequential Layer
We didn't come into crypto to argue about dollar-pegged settlement instruments. We came for sovereignty, for censorship resistance, for the idea that code could be law. And yet the part of the stack that actually moves the world's money is the part that looks least like the dream.
I learned that in 2017, in a cryptography lecture hall in Tallinn, when I first understood that a transaction could be unstoppable. I dropped the Java projects I'd planned and wrote a forty-page manifesto called "The Freedom Stack," printed 500 copies, and handed them out at the local hacker space until 200 people had subscribed to a newsletter I didn't know how to run. I believed then — and mostly still do — that technology should serve human autonomy. But the years since have taught me that autonomy at the edges often depends on trust at the center, and stablecoins are where that tension becomes unavoidable.
Stablecoins are where the abstraction meets the bank account. When you hear that a stablecoin "settles" a payment, what you're really hearing is that somewhere a trust company is holding Treasury bills and a smart contract is minting a receipt. The philosophy we fight about at the consensus layer — who validates, who can be censored, who decides — gets compressed at the settlement layer into a much blunter question: who can freeze, and under what authority?
USDG is issued by Paxos, a New York-chartered trust company that has spent the better part of a decade inside the regulatory perimeter. It holds a trust charter from the New York Department of Financial Services. It has issued regulated assets before — USDP, and the Binance-branded BUSD that the SEC once threatened and then backed away from. If you're ranking issuers by how boring their legal posture is, Paxos sits near the top. That matters more than any technical spec, because in this corner of the market, boring is the product.
The backdrop is what made boring valuable. Over the past few years the regulatory picture stopped being a fog and became a map. Europe wrote its rules down in MiCA. The United States spent years arguing about stablecoin legislation before moving toward a framework. Every one of those moves punished opacity and rewarded the issuers who had already built a compliance department. That's the wind at USDG's back — not a technical breakthrough, but a legal one. The same forces that make life hard for a gray-market dollar token make life easier for a chartered one.
The interesting part isn't the issuer, though. It's the distribution model. USDG is tied to something called the Global Dollar Network — an alliance in which partners distribute the stablecoin and, in exchange, share in the yield the reserves generate. This is the actual innovation, and it isn't a technical one. It's an economic one. The reserve yield, not the token, is the product.
Let me put a number on that, because numbers cut through narrative. At $3.2 billion in circulation, assuming a 4–5% yield on short-term Treasuries, the reserves throw off somewhere between $128 million and $160 million a year. That's the pie. The question that determines USDG's future is not "how big is the pie" — it's "how is the pie split, and with whom."
And that's where the story gets fragile.
The source material here is thin — a two-data-point brief, no primary sourcing, no date. I want to be honest about that, because the version of me that ran those unsecured aggregators learned that thin information is exactly where people project their hopes. So let me separate what we know from what we're inferring. We know: $3.2 billion, 340% growth, Paxos issues it. We're inferring: multi-chain deployment, freeze and blacklist functions in the contract, a yield-sharing arrangement with a small set of partners. Those inferences are load-bearing, and I'll flag them as such as we go.
Reading the Hinge
Here's what a stablecoin actually is, stripped of the branding.
On the reserve side, you have a pile of assets — usually cash and short-term government debt — sitting with a custodian. On the contract side, you have a mint/burn mechanism and, in almost every regulated design, a set of administrative controls: freeze, blacklist, pause, upgrade. The "technology" of a fiat-backed stablecoin is not consensus, not cryptography, not some novel signature scheme. It's the engineering of trust boundaries — where you draw the line between what the code enforces and what a human can override.
I've read enough of these contracts now to know the pattern. The interesting functions are never in the whitepaper. They're in the proxy admin slot. And the honest answer for nearly every regulated stablecoin is: yes, someone can freeze your balance, and yes, someone can upgrade the contract underneath you. That's not a bug in USDG specifically. It's the price of the regulatory perimeter. But it means the decentralization story is, at the settlement layer, largely decorative.
There's a parallel I keep reaching for. We spent two years being sold "decentralized sequencing" for Layer 2s, and when you actually look at who runs the sequencer, it's a single node with a nice logo. The word "decentralized" did a lot of work that the code didn't. Stablecoins are the same trick at a different layer: the branding says permissionless, the contract says admin key. When the adjective and the architecture disagree, trust the architecture.
Now layer the economics on top.
A stablecoin that shares reserve yield with its distributors is running a customer-acquisition business, not a monetary one. Think about what that actually means. Paxos collects interest on Treasuries. It hands a slice to partners — exchanges, brokers, payment processors — who then push USDG into their user bases. The partner's incentive is the slice. The user's incentive is convenience, and maybe a cut of the yield too, if the model reaches that far.
This is a genuinely clever structure, and I want to give it credit before I stress-test it. It's more sustainable than the old playbook of paying users in a governance token that inflates to zero. Yield sharing uses real revenue — Treasury interest — rather than emitted tokens. That's a meaningful upgrade over the 2020-era liquidity mining I lived through. My aggregators paid in tokens that I had no real cash flow to back. Paxos is paying in dollars it actually earns. I'm not going to pretend that isn't progress.
But — and this is the hinge — a yield-sharing network is only as strong as the loyalty it buys, and loyalty bought with yield is priced continuously by the market. If a competitor offers a better split, the partner has no technical reason to stay. There's no switching cost baked into the code. There's no network effect that compounds on its own. There's a spreadsheet.
The moat is a business development team, not a protocol.
I keep coming back to the three-year storytelling exercise we've been calling "RWA on-chain." We've spent years promising that real-world assets would migrate to public chains, that institutions would discover the magic of permissionless settlement. And the quiet truth is that most institutions don't need our public chains at all. They need settlement, compliance, and a counterparty they can sue. USDG isn't an argument against that truth — it's a demonstration of it. It's a regulated instrument that happens to use blockchain rails because the rails are cheap and fast, not because the rails are sovereign. The chain is a delivery mechanism. The institution is the product.
Which brings me to concentration risk, the one the brief actually names.
The single most important sentence in that whole report is the warning that USDG depends on a small number of partners. That's not a footnote. That's the architecture. If the growth came from one or two large channels — a major exchange, a broker, a payments firm — then 340% is not a measure of organic adoption. It's a measure of how much one partner decided to push. And a partner who can turn the tap on can turn it off. The vulnerability there isn't in the code — Root: The partner concentration, the thing no audit will ever catch.
I've watched this movie before, from the inside. In 2021 I co-founded an NFT collective with 5,000 holders, blending digital art with real-world residency rights. When the market turned in 2022, the floor dropped 80%, and the holders who'd come for the yield — sorry, for the "utility" — vanished. The ones who stayed were there for something the spreadsheet couldn't price. A community built on incentives dissolves the moment the incentives stop; a community built on belief survives the winter. Stablecoins are running that same experiment at institutional scale, and the brief is quietly telling us which side of it USDG is on.
I spent the following year interviewing fifty long-term holders about what kept them in during the drawdown. Almost none of them mentioned price. They mentioned belonging. I don't think that's sentimentality; I think it's the actual mechanism by which anything survives a bear market. And it's the mechanism USDG's growth model doesn't have. A channel partner has no belonging. It has a rate sheet.
Let me get concrete about the risks, because vague warnings are useless.
Contract risk comes first. No audit is mentioned anywhere in the material. For a $3.2 billion instrument, that's not a detail — it's a gap you could drive a truck through. I've seen unaudited upgradeable proxies go sideways, and the failure mode is almost never the cryptography. It's the admin key. It's the pause function used under legal pressure. It's the upgrade that nobody outside the issuer reviewed. In 2024, working inside Estonia's regulatory sandbox on a decentralized identity protocol, I watched how quickly "compliance" can become a reason to add a control, and how rarely anyone asks who the control serves. The paperwork was miserable — I missed deadlines because I kept chasing new integrations instead of filling forms — but the lesson stuck: every compliance feature is a power, and every power has an owner.
Reserve risk comes second. We don't know the composition. If it's short-term Treasuries, USDG is indirectly exposed to US monetary and fiscal policy — which is fine until it isn't. The 2023 USDC depeg wasn't a crypto failure at all; it was a bank failure, a reminder that the reserve is the real chain of custody. A stablecoin is only as stable as the balance sheet behind it, and that balance sheet lives outside the chain, in institutions the chain can't audit. The peg is a promise, and promises live in the world of contracts and courts, not in the world of blocks.
Regulatory risk in the opposite direction comes third, and it's the most underrated. Everyone frames regulation as a tailwind for compliant issuers, and mostly it is. But the same regulators who bless you can also decide that reserve yield should accrue to holders rather than issuers and their partners. If that happens, the entire yield-sharing model — USDG's reason for existing — gets legislated out from under it. The competitive advantage of a compliant stablecoin is also its single point of failure: the rulebook can change.
I keep thinking about the AI agents I've been building with. In 2025 I launched a platform that lets autonomous agents hold wallets and negotiate services, and the whole thing ran aground on a question no one had an answer for: who is legally responsible when an agent acts? We have economic agency without legal personhood, and the system has no slot for it. Stablecoins are the same shape of problem. They have monetary function without monetary sovereignty, and they borrow their legitimacy from institutions that can revoke it. Code may be law in the narrow sense, but it isn't the only law in the room. That's the essay I wrote for Lisbon, and it's the same essay hiding inside every "compliant" stablecoin launch.
Now the comparison that keeps me honest. USDT holds roughly 65% of the market, USDC about 25%. USDG, at $3.2 billion, is somewhere around 1–1.5%. That's not a rounding error, but it's two orders of magnitude from the top. Stablecoins are a network-effect business — liquidity begets liquidity, integrations beget integrations, merchant acceptance compounds. A late entrant doesn't beat that with a better yield split; it beats it, if at all, by borrowing someone else's distribution. Which is exactly what the Global Dollar Network is — and exactly why the concentration risk matters so much.
I want to be fair to the 340% one more time, because it deserves a careful reading rather than a cynical one. A 340% year-over-year increase sounds explosive. But percentages from a small base can flatter almost anything. Going from a few hundred million to $3.2 billion is genuinely impressive growth — and it's also the kind of growth that a single well-motivated partner can manufacture in a quarter. The honest reading of the 340% is that it's a distribution number dressed up as an adoption number. It measures how hard the partners pushed, not how deeply the market pulled. That's the information gain nobody publishing the headline wants to state: growth velocity in stablecoins is a channel metric before it is a demand metric.
There's a version of this that ends well, and I want to name it, because I don't want to be the guy who only sees the hinge. If USDG's partner base broadens — dozens of channels, none dominant, each with a real reason to stay because the yield split is genuinely competitive — then the concentration risk fades and the 340% becomes the floor of something real. Yield sharing could become the standard, forcing even the incumbents to share reserve income with their distributors and, eventually, their users. That would be a real transfer of value from issuers to the network. That's a future worth rooting for. It just isn't the future the current data proves.
The Pragmatism Test
Here's where I want to push against my own instinct, because the decentralization reflex can be lazy.
The lazy version of my argument is: "USDG is centralized, therefore it's bad." That's not analysis. That's a slogan. Bitcoin's Lightning Network has been effectively niche and half-dead for seven years, and I've made that argument, and I believe it — but it doesn't mean every centralized thing fails. USDT is more centralized than USDG and it's the most-used dollar instrument in the world. Centralization is not a verdict. It's a trade.
So the pragmatist's question isn't "is USDG decentralized?" It's "does USDG's particular bundle of trade-offs serve a real need that the alternatives don't?"
And the answer is: maybe, for a specific customer. An institution that finds USDT's reserve opacity intolerable and USDC's integration costs annoying might genuinely prefer a Paxos-issued instrument with a yield kickback. That's a real segment. The brief's own framing — compliance-first plus yield-sharing, aimed at institutional clients skeptical of USDT — is a coherent position. I don't have to like the architecture to admit the market might want it. A settlement medium that a treasurer can defend to a board is worth something the crypto-native crowd routinely undervalues.
But the pragmatism test cuts the other way too, and this is the part the bulls skip. If the value proposition is "we share yield with partners," then USDG is competing in a market where the incumbent can copy that feature tomorrow. Tether could announce a yield-sharing program in a press release. Circle could fold it into USDC's institutional tier. The differentiator isn't patented; it isn't cryptographic; it isn't even hard. It's a contract amendment and a business decision. And the moment it's copied, USDG is back to being a 1% stablecoin with a slightly nicer compliance story. What determines survival in that scenario — Root: The belief, not the yield.
There's a deeper tension I keep circling, and it's the same one I ran into with RWA. We built this industry on the promise that you wouldn't need permission. And the most successful products in it are the ones that turned permission into a feature — a trust charter, a regulator's seal, a partner's handshake. That's not hypocrisy, exactly. It's the market telling us what it actually values. But we should at least be honest that the settlement layer we're celebrating looks a lot more like the financial system we said we were escaping, just with faster rails and a nicer interface.
The RWA dream and the USDG reality are the same story told from two ends. We said real assets would come on-chain. They did — as custodial receipts issued by a trust company, freezeable on command, distributed through a permissioned club. The chain is real. The sovereignty isn't. And if that's the endgame, we should say so out loud rather than dress it up in the vocabulary of a revolution. The most honest thing a stablecoin can tell you is that it isn't trying to replace the banks. It's trying to be a better one.
What to Watch, and What to Ask
The number to watch isn't $3.2 billion. It's the partner count.
If USDG's growth diversifies — if the yield-sharing network spreads across dozens of channels with no single one dominant — then the 340% is the beginning of something durable, and the model deserves the attention it's getting. If it doesn't, then we're watching a distribution deal masquerading as a monetary network, and the same spreadsheet that turned the tap on can turn it off. The signals worth tracking are mundane and public: the reserve attestations, the partner disclosures, the share of supply that moves through any single channel, and the day the first competitor copies the yield split. None of those require a chain explorer. All of them require patience.
So here's the question I'd put to every builder who reads the 340% and feels the pull: when you look at a stablecoin, are you looking at the paint, or the hinge? Because the paint says compliance and yield and growth. The hinge says a key, a spreadsheet, and a rulebook someone else can rewrite.
We didn't come this far to rebuild the old system with new plumbing. But we should at least know when that's what we're doing — and build the next thing with our eyes open.