Both USDC and DAI trade at $1.00. They both show up in the same DeFi protocols, denominate the same yield positions, and settle the same trades. From the outside, they look like variations on the same product.
They are not the same product. They achieve the same price target through completely different mechanisms — and those mechanisms carry different trust assumptions, different failure modes, and different exposure to regulatory action. Understanding the difference is less about picking a "better" stablecoin and more about knowing what you're actually holding when you hold one.
The Two Mechanisms
Centralized stablecoins: fiat-backed issuance
A centralized stablecoin like USDC or USDT works the same way a bank deposit does: an issuer holds reserves and issues claims against them. You send $1 to Circle; Circle mints 1 USDC. The peg is enforced through redemption arbitrage — if USDC trades at $0.99, arbitrageurs buy it cheaply and redeem at the $1.00 floor directly from Circle, pushing the price back up.
The mechanism is simple and robust under normal conditions. The reserve is real money, accessible through a regulated entity.
What varies across centralized stablecoins is reserve composition and transparency. USDT has historically held a mix of cash, commercial paper, secured loans, and other assets — its 2021 CFTC settlement included a $41M fine for misrepresenting reserve composition. USDC publishes monthly attestations by a Big Four accounting firm and holds primarily cash and short-term US Treasury bills. The peg mechanism is identical; the reserve quality is not.
The March 2023 SVB collapse illustrates the constraint in practice. Circle held $3.3B of USDC reserves at Silicon Valley Bank — roughly 8% of USDC's then-backing. When SVB failed, USDC briefly traded at $0.87 on secondary markets before Circle confirmed the funds would be made whole via the FDIC backstop. The peg mechanism held, but only because of a government intervention that Circle could not have guaranteed in advance.
Centralized stablecoins also retain the ability to blacklist addresses. Circle has frozen hundreds of millions of USDC on demand — primarily in response to sanctions compliance and law enforcement requests. USDT has done the same. This is a feature for regulatory compliance and a constraint for permissionless use cases.
Decentralized stablecoins: protocol-enforced overcollateralization
DAI (now managed by Sky Protocol, formerly MakerDAO) works differently. To mint 100 DAI, you lock collateral — ETH, WBTC, or other approved assets — worth significantly more: typically 150% or higher. The protocol enforces a minimum collateralization ratio. If the value of your locked ETH falls below the liquidation threshold, a keeper bot liquidates your position, sells the collateral to cover the DAI debt, and the system remains solvent.
There is no custodian. The protocol governs the peg through several mechanisms:
- If DAI trades above $1.00, opening new vaults becomes attractive — users lock collateral and sell freshly minted DAI, increasing supply and pushing the price down
- If DAI trades below $1.00, repaying DAI debt at a discount becomes attractive — users buy cheap DAI, repay their vault, and release their collateral
- The protocol's Stability Fee (interest charged on minted DAI) and DSR (DAI Savings Rate, interest paid for locking DAI) are governance-adjustable levers that influence demand
LUSD (Liquidity Protocol) operates similarly but with stricter constraints: ETH-only collateral, no governance tokens, hardcoded parameters — sacrificing flexibility for resilience.
A critical distinction: algorithmic vs decentralized
UST/LUNA was not a decentralized stablecoin in the DAI sense — it was an algorithmic stablecoin with no hard collateral backing. Its peg relied on an incentive mechanism: burn $1 of LUNA to mint 1 UST, and vice versa. When confidence in the system broke in May 2022, the mechanism went into a death spiral — collapsing $40B in value in roughly 72 hours. This failure does not characterize overcollateralized decentralized stablecoins; it characterizes a system with no underlying assets.
Where the Constraints Live
Centralized stablecoins:
- Issuer risk — the organization holding reserves must remain solvent and operational
- Banking risk — reserves held at banks are subject to bank failure (SVB precedent)
- Regulatory risk — the issuer operates under financial regulation; licensing requirements, asset freezes, and jurisdiction-specific restrictions apply
- Blacklisting — addresses can be frozen at issuer discretion
Decentralized stablecoins:
- Collateral quality risk — if the primary collateral (ETH) experiences a severe, rapid decline, the protocol may not liquidate positions fast enough to avoid undercollateralization
- Oracle risk — the protocol relies on price feeds to trigger liquidations; a manipulated or delayed oracle feed is an attack surface
- Smart contract risk — the MakerDAO/Sky contract stack has been audited repeatedly, but contract risk is never zero
- Capital inefficiency — overcollateralization requires users to lock $150+ of collateral to access $100 of stablecoin; this capital constraint limits scale relative to fiat-backed alternatives
What's Changing
DAI's collateral has shifted toward centralized assets. MakerDAO's transition to Sky Protocol included expanding DAI collateral to include USDC, Coinbase custodied ETH (cbETH), and US Treasury Bills via real-world asset (RWA) vaults. This improves capital efficiency but introduces centralized dependencies — DAI's decentralization claim is now partial. The protocol acknowledges this tension explicitly.
Ethena's USDe is a distinct third model. It holds ETH and uses short perpetual futures positions to delta-hedge, targeting price stability through funding rate mechanics rather than overcollateralization or fiat reserves. At $5B+ supply by early 2025, it's a real alternative — but its risk profile differs from both: funding rate dependency, perpetual market liveness, and counterparty exposure to centralized exchanges.
Regulated centralized stablecoin expansion. PayPal's PYUSD (issued by Paxos), Ripple's RLUSD, and the potential passage of the US GENIUS Act (stablecoin legislation) are pushing centralized stablecoins toward explicit regulatory licensing. This increases clarity and custody standards but also formalizes the state's authority over them.
What Would Confirm This Direction
- Sky Protocol's RWA vaults continuing to scale without triggering DAI peg instability — demonstrating that hybrid collateral can work at size
- GENIUS Act or equivalent legislation passing and major issuers operating under it — formalizing the centralized model's regulatory status
- Ethena USDe surviving a sustained period of negative funding rates without depegging — validating the model under stress
What Would Break or Invalidate It
- An ETH flash crash causing a wave of DAI/LUSD undercollateralization before liquidations clear — would impair the overcollateralized decentralized model's core safety claim
- Tether insolvency or a reserve revelation materially inconsistent with published attestations — would damage the centralized model's credibility broadly, not just USDT
- Regulatory prohibition on fiat-backed stablecoin issuance in a major jurisdiction without adequate domestic alternatives
Timing Perspective
Now: The centralized/decentralized distinction matters for DeFi protocol risk assessment today. Centralized stablecoin exposure carries blacklist and issuer risk. Decentralized stablecoin exposure carries oracle and collateral risk. Neither is inherently safer — they carry different risks.
Next: The regulatory landscape for centralized stablecoins is clarifying in the US and EU. Clarity could increase institutional adoption of centralized stablecoins and put pressure on hybrid models like DAI to define their regulatory posture.
Later: Whether genuinely decentralized stablecoins can scale to compete with fiat-backed issuers at institutional size without compromising their collateral model remains unresolved.
Boundary Statement
This post explains the mechanism of centralized and decentralized stablecoins and the structural constraints each carries. It does not recommend holding or avoiding any specific stablecoin, nor does it assess current reserve composition for any issuer. Reserve quality, audit coverage, and regulatory status change — verify current conditions before making decisions.
The mechanism works as described. Whether any particular stablecoin is appropriate depends on factors outside this scope.
