USDT vs USDC vs DAI vs FDUSD — the four stablecoins compared on what matters
USDT, USDC, DAI and FDUSD together hold $228bn — but they differ on reserves, freeze risk, MiCA status and on-chain availability. A side-by-side that actually compares what matters.
As of 14 March 2026 the four largest dollar stablecoins — Tether USD, USD Coin, Dai, and First Digital USD — together carried $228.4 billion of float, or roughly 97% of the dollar-pegged stablecoin market. They are routinely treated as interchangeable on exchange interfaces and DeFi front-ends, which is convenient and wrong. USDT, USDC, DAI and FDUSD differ on reserve composition, custodian, freeze powers, MiCA authorisation, on-chain availability, redemption mechanics and counterparty risk. Choosing between them is not a branding decision; it is a treasury decision, and over time the wrong choice has cost users between 50 basis points and 100% of principal depending on the failure mode.
What is at stake is the operational reliability of the unit you actually hold. A trader holding USDT on a centralised exchange faces different risk than a DAO treasury holding USDC at a smart-contract vault, which faces different risk than a DeFi user holding DAI in a non-custodial wallet. In practice that means each token is fit for some purposes and unfit for others. The comparison below is built from each issuer’s most recent attestation, MiCA filings as of Q1 2026, and on-chain data from DefiLlama and Etherscan retrieved on 12 March 2026.
The headline numbers
| Token | Issuer | Float (Mar 2026) | Chains | Year launched |
|---|---|---|---|---|
| USDT | Tether Holdings (BVI) | $143.2bn | 14 (Ethereum, Tron, Solana, Avalanche, etc.) | 2014 |
| USDC | Circle Internet Group (US, Ireland EMI) | $58.7bn | 15 (Ethereum, Solana, Base, Arbitrum, etc.) | 2018 |
| DAI / USDS | Sky / MakerDAO (decentralised) | $11.9bn (DAI + USDS) | 9 (Ethereum native, bridged elsewhere) | 2017 |
| FDUSD | First Digital Labs (Hong Kong) | $14.6bn | 3 (Ethereum, BNB Chain, Sui) | 2023 |
USDT’s dominance is structural — it was first to Binance, first to Bitfinex, first to most Asian venues, and that liquidity head start has not been displaced. USDC has the regulatory mandate (Circle is an EU-authorised EMI and a NYDFS-supervised limited purpose trust) and won the institutional segment in 2023-2024. DAI is the only meaningful decentralised holdout and is in the middle of a brand migration to USDS under the Sky rebrand. FDUSD is a 2023 entrant that grew on Binance’s promotion of zero-fee BTC/FDUSD pairs and remains heavily concentrated on that one venue.
Reserves and counterparty risk
This is where the four diverge most sharply. USDC holds about 80% of reserves in the BlackRock-managed Circle Reserve Fund (CRF), a 2a-7 government money-market fund holding US T-bills and Federal Reserve overnight repo. The remaining ~20% is cash at BNY Mellon, JPMorgan, Citi and a handful of other GSIBs. This is the cleanest reserve profile of the four: short-dated, transparent, custodied at top-tier banks, and verified monthly by Deloitte.
USDT is more heterogeneous. The Q4 2025 BDO Italia attestation listed roughly 84% T-bills and reverse repo, 5% gold, 3% bitcoin, 4% secured loans, and the balance in other investments. Tether’s secured-loan book ($6.4bn at end-2025) and gold/BTC holdings are unique among major stablecoins and represent the bulk of the credit and market risk in the structure. The attestation is quarterly rather than monthly, and Tether has never published a full PCAOB audit. Its primary custodian is Cantor Fitzgerald.
DAI is unique: its collateral is on-chain and continuously verifiable. The Maker system overcollateralises every DAI in circulation with crypto (ETH, wstETH, wBTC) and tokenised real-world assets (Monetalis-managed T-bills, BlockTower credit). The system collateralisation ratio sits at roughly 178% as of March 2026. There is no bank-custody counterparty risk for the crypto collateral; the RWA portion (~$2.1bn) is custodied by US qualified custodians. FDUSD holds cash and short-dated US T-bills custodied with First Digital Trust in Hong Kong, attested monthly by Prescient Assurance. Its concentration in a single Hong Kong trust company is the structural risk.
Freeze powers and censorship resistance
Three of the four have address-freeze capability built into the token contract. USDC’s blacklist function has frozen ~$97m across 252 addresses since 2020, mostly in response to OFAC designations and law-enforcement requests. USDT has frozen ~$1.3bn across roughly 2,100 addresses, the largest counts on Tron rather than Ethereum. FDUSD has a similar freeze function and has used it sparingly — public counts are low but the capability exists. DAI alone has no token-level freeze function. Maker governance can disable a collateral type or change parameters, but cannot blacklist a holder address.
| Token | Freeze function | Addresses frozen (lifetime) | Approx. value frozen | Process |
|---|---|---|---|---|
| USDT | Yes (Treasury multi-sig) | ~2,100 | ~$1.3bn | Law enforcement request, OFAC compliance |
| USDC | Yes (Centre/Circle key) | 252 | ~$97m | OFAC, court order, compliance |
| FDUSD | Yes (issuer key) | Low, undisclosed | Undisclosed | Compliance request |
| DAI | No (contract has no blacklist) | 0 | $0 | N/A — governance can pause modules only |
Regulatory status: who can hold what, where
The MiCA implementation deadline of 30 June 2024 split the market. Circle obtained an EMI licence from the ACPR in France and operates EURC and USDC as compliant EMTs across the EU; both BaFin and the AMF list Circle Mint Europe SAS as authorised. Tether did not apply and was delisted from EU-facing order books on Binance, Coinbase, Kraken, Crypto.com and Bitstamp during H2 2024 and Q1 2025. FDUSD has filed with regulators in Hong Kong under the SFC stablecoin regime that took effect 1 August 2025 but is not MiCA-authorised. DAI sits in a regulatory limbo: it is not an EMT under MiCA because it is not issued by a single legal person, but it is also not clearly outside the regulation. ESMA has so far not pursued enforcement.
In the US the position is the inverse of the EU. The GENIUS Act framework, signed July 2025, brought USDC, USDT and FDUSD into a federal/state dual-licensing regime; Tether is currently in the process of restructuring a US entity to satisfy the new rules but as of March 2026 only USDC and FDUSD are unambiguously compliant for US retail issuance. DAI continues to be treated as a non-issued asset and falls outside the GENIUS perimeter. Our regulatory calendar tracks the implementation deadlines under both regimes.
On-chain liquidity and fee profile
For DeFi use the practical question is where you can swap large size with low slippage and which routing the aggregators prefer. On Ethereum mainnet, USDC has the deepest stable-to-stable liquidity (Curve 3pool, Uniswap v3 USDC/USDT, and the FRAX/USDC pool together hold ~$3.1bn at March 2026 prices). USDT is roughly equal on Tron — TRC-20 USDT is the dominant remittance rail in Latin America, South Asia and parts of Africa, with the cheapest transfer cost of any major stablecoin (median ~$0.30 per transfer in 2025). DAI has progressively narrower DEX liquidity since Maker’s pivot to USDS but remains the canonical settlement asset for many DeFi protocols. FDUSD’s DEX depth is thin — roughly $80m across DEX pools — because its primary use case is CEX trading on Binance, where the BTC/FDUSD book is the deepest BTC market by quote volume.
Chain footprint matters too. USDT issued on Tron (~$72bn) now exceeds USDT on Ethereum (~$56bn) because the per-transfer cost difference (≈$0.30 vs ≈$3.20 at March 2026 gas prices) compounds at remittance volumes. USDC’s distribution is the inverse — heaviest on Ethereum, with Base (~$5.4bn) and Solana (~$4.9bn) the next two largest deployments. The practical implication is that the “same” token has slightly different liquidity, redemption mechanics and bridge risk on each chain. A USDC balance on Polygon bridged via the older PoS bridge is not interchangeable with native Circle-issued USDC on Ethereum in a stress event; the bridge introduces a separate failure mode. Circle’s CCTP burn-and-mint protocol is the canonical solution and has progressively replaced the legacy bridges since 2023.
- Cheapest cross-border transfer: TRC-20 USDT (~$0.30 typical fee).
- Deepest DEX liquidity on Ethereum: USDC.
- Most censorship-resistant: DAI (no token-level freeze).
- Best institutional custody profile: USDC (BNY Mellon, BlackRock CRF).
- Deepest CEX BTC pair: FDUSD on Binance (zero-fee promotion).
Yield and idle-balance economics
None of the four pays yield to plain token holders — that line is enforced under MiCA Article 45 in the EU and presumed under SEC staff guidance in the US. To earn the underlying T-bill rate (~4.3% SOFR as of March 2026) a holder must move the token into a yield-bearing wrapper. For USDC that means the SOFR-linked products from Circle’s institutional programme or third-party tokenised T-bill funds. For DAI the Sky Savings Rate (SSR) currently pays 4.75% directly on chain, the highest rate among native stablecoin savings products. USDT has no native savings rate; users hold T-bill exposure indirectly via lending platforms like Aave and Compound, where the supply APY oscillates between 3.5% and 8% depending on demand. FDUSD has no native yield product. Our stablecoin yield calculator compares the after-fee net return on each token across the major venues.
Which to use, and for what
The decision rule is roughly: USDC for any EU-regulated activity, institutional treasury, or DeFi position requiring regulatory comfort. USDT for cross-border transfer (especially TRC-20), Asia-Pacific exchange settlement, and any venue where USDC is not deeply quoted. DAI/USDS for DeFi positions where on-chain composability and censorship resistance matter, and for capturing the Sky Savings Rate. FDUSD almost exclusively for Binance trading, where the BTC pair depth justifies holding it.
None of these tokens has a clean profile across every dimension; each is a set of trade-offs between regulatory standing, censorship resistance, distribution reach, and reserve quality. Users who hold significant balances should split across at least two issuers to diversify the residual issuer-default risk — the lesson SVB taught in March 2023, and the lesson Terra taught in May 2022 in a more brutal form. Our stablecoin peg monitor tracks the live deviation, mint and burn velocity and 30-day reserve attestation status for all four, and our market dashboard consolidates float and chain-share data weekly.