Not all stablecoins are stable. The collapse of UST in 2022, the USDC depeg in March 2023, and the persistent fragility of algorithmic stablecoins have made one thing clear: “stable” is a spectrum, not a binary.
TrustGrade scores every major stablecoin using the same Bayesian engine that powers our DeFi protocol rankings — but with stablecoin-specific signal weights. The results are stark. USDC scores 89 (A). Tether scores 76 (B). DAI scores 85 (A−). And every algorithmic stablecoin we’ve tracked? F grade.
Here’s why.
What Makes a Stablecoin “Stable”
A stablecoin’s job is simple: maintain its peg to a reference asset (usually USD). TrustGrade measures four core signals to evaluate how reliably a stablecoin does this:
| Signal | Weight | What It Measures |
|---|---|---|
| Peg Stability (Wilson) | 45% | Historical peg adherence rate |
| Collateral Quality (Bayesian) | 20% | Reserve composition transparency |
| Depeg History (Penalty) | 20% | Severity and frequency of depeg events |
| Market Depth (Volume) | 15% | Liquidity depth across major DEXs/CEXs |
Notice the emphasis: 45% weight on peg stability. This is the single most important metric. If a coin can’t hold its peg, nothing else matters.
USDC: The Gold Standard (Score: 89, Grade: A)
USDC’s trust profile is strong across every dimension:
- Peg stability: 99.94% of trading days within ±0.3% of $1.00 over 4+ years
- Collateral: Cash + short-term US Treasuries, attested monthly by Deloitte
- Depeg events: One significant event (March 2023, SVB collapse), recovered within 6 days
- Daily volume: $8B+ across all venues
The Wilson score for peg stability is 0.9986 — extraordinarily high. Even the SVC depeg wasn’t a USDC failure; it was a banking failure that temporarily disrupted reserves. The peg recovered fully.
curl -X POST https://api.trustgrade.ai/v1/score \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"entityType": "stablecoin",
"entityId": "USDC",
"data": {
"pegStability": 0.9994,
"depegEvents": [{"date": "2023-03-11", "severity": 0.3, "duration": "6d"}],
"collateralRatio": 1.0,
"collateralQuality": "cash+treasuries",
"attestation": "monthly-Deloitte",
"dailyVolume": 8000000000
}
}'
# Response
{
"score": 89,
"grade": "A",
"confidence": 0.97,
"breakdown": {
"wilson": 0.9986,
"bayesian": 0.92,
"volumeBonus": 2.4,
"penalty": 0.09,
"optionals": { "collateralTransparency": 0.95, "regulatory": 0.85 }
}
}
Why not A+? The March 2023 depeg, while fully resolved, activates the compound penalty. Regulatory uncertainty around stablecoin legislation creates a small Bayesian discount. And Circle’s reliance on US banking partners is a centralized risk vector.
USDT: Strong but Opaque (Score: 76, Grade: B)
Tether is fascinating from a scoring perspective. Its peg stability is excellent (99.7% of days within ±0.5%), and daily volume ($45B+) is unmatched. But it loses points on two dimensions:
-
Collateral transparency: Despite improving dramatically (Tether now publishes quarterly attestations from BDO), the reserve composition includes commercial paper, money market funds, bitcoin, and “other investments.” The Bayesian prior for mixed-quality collateral is lower than for pure cash + Treasuries.
-
Historical opacity penalty: From 2017–2021, Tether operated with minimal transparency. The compound penalty accounts for this history even though current practices have improved.
The score reflects current reality, not past reputation. If Tether maintains its current transparency standards for another 18 months, expect the score to rise to B+.
DAI: Decentralized but Complex (Score: 85, Grade: A−)
DAI presents an interesting case. It’s over-collateralized (collateral ratio typically 1.5x+), has maintained peg through extreme conditions for 8+ years, and is fully decentralized. But:
- Collateral complexity: DAI is backed by a mix of ETH, USDC, RWA (real-world assets), and other tokens. The RWA component (~40% of backing) introduces legal and counterparty risk.
- Depeg history: DAI traded below $0.95 during the March 2020 COVID crash and briefly above $1.05 during the UST collapse. These events, while resolved, trigger the compound penalty.
- Governance risk: MakerDAO governance can change collateral parameters. While this hasn’t been abused, it creates theoretical risk.
DAI scores 85 — excellent, but the complexity of its collateral framework prevents it from matching USDC’s simpler, cleaner profile.
The F Grade: Why Algorithmic Stablecoins Fail
Every algorithmic stablecoin we’ve scored earns an F. This isn’t ideology — it’s mathematics.
The Structural Problem
Algorithmic stablecoins maintain their peg through incentive mechanisms rather than collateral. When demand drops, the protocol creates/mints/burns governance tokens to absorb selling pressure. This creates a fundamental problem: the stabilization mechanism depends on market confidence in the governance token — the very thing that’s collapsing when a depeg starts.
It’s a reflexive death spiral waiting to happen.
Case Study: UST (Terra)
TrustGrade’s historical scoring of UST (before its collapse):
| Date | Score | Grade | Peg Deviation | Signal |
|---|---|---|---|---|
| Jan 2022 | 62 | C | ±0.5% | Thin collateral (3% BTC reserve) |
| Apr 2022 | 55 | D | ±1.2% | Anchor Protocol withdrawals accelerating |
| May 7, 2022 | 38 | F | -3.2% | Death spiral initiated |
| May 9, 2022 | 12 | F | -30% | Terminal collapse |
The Bayesian model detected the deterioration months before the collapse. As Anchor Protocol withdrawals accelerated (the “bank run” precursor), the Wilson score on peg stability dropped. The compound penalty amplified each day of deviation.
If an AI agent had checked UST’s TrustGrade on May 1, 2022, it would have seen a D-grade asset and blocked all interactions.
Current Algorithmic Stablecoins (July 2026)
| Stablecoin | Score | Grade | Notes |
|---|---|---|---|
| USDD (Tron) | 42 | F | Limited collateral, centralized control |
| USTC (Terra Classic) | 8 | F | Terminal — peg never recovered |
| FRAX (V2) | 58 | D | Partial algorithmic, transitioning to fully collateralized |
FRAX is the highest-scoring “algorithmic” stablecoin because it’s been migrating toward full collateralization. As its collateral ratio increases, its score improves. But the algorithmic component keeps it in D territory.
The TrustGrade Stablecoin Trust Matrix
| Stablecoin | Score | Grade | Backing | Peg (90d) | Daily Vol |
|---|---|---|---|---|---|
| USDC | 89 | A | Cash + Treasuries | ±0.02% | $8B |
| DAI | 85 | A− | Over-collateralized (mixed) | ±0.05% | $2B |
| USDT | 76 | B | Mixed reserves | ±0.04% | $45B |
| BUSD | — | — | Discontinued (Paxos) | — | — |
| FRAX | 58 | D | Partially algorithmic | ±0.15% | $200M |
| USDD | 42 | F | Limited collateral | ±0.30% | $80M |
| USTC | 8 | F | Un collateralized | ±85% | <$1M |
How to Use These Scores
For Developers and AI Agents
Integrate TrustGrade stablecoin scores to automatically filter which tokens your application accepts:
# MCP query: only accept A-grade stablecoins
result = await mcp_client.call_tool("check_trust", {
"entity_type": "stablecoin",
"entity_id": "USDC"
})
accepted_stablecoins = []
for coin in ["USDC", "USDT", "DAI", "FRAX", "USDD"]:
score = await get_trustgrade_score(coin)
if score["grade"] in ("A", "A-", "B+"):
accepted_stablecoins.append(coin)
else:
log.warning(f"Rejected {coin}: grade {score['grade']}")
# Result: ["USDC", "DAI"] (USDT may pass if threshold is B)
For Traders and DeFi Users
Use the score as a first filter, not the final decision. A stablecoin scoring B might be acceptable for short-term yield farming but not for long-term holding. F-grade stablecoins should never be held.
For Protocols
If your protocol accepts multiple stablecoins as collateral, use TrustGrade scores to set dynamic loan-to-value ratios — higher LTV for A-grade coins, lower for B-grade, zero for D and F.
Conclusion
The data is unambiguous: collateralized stablecoins are safer than algorithmic ones. This isn’t a controversial opinion — it’s what the Bayesian math says when you feed it peg data, collateral composition, and depeg history.
USDC earns its A through transparent backing, excellent peg stability, and rapid recovery from the one significant depeg event. Algorithmic stablecoins earn their F through structural fragility — the stabilization mechanism itself becomes the attack vector.
When you see a TrustGrade score on a stablecoin, you’re seeing the output of months of onchain data, processed through a methodology that doesn’t care about marketing, team reputation, or community sentiment. Just the numbers.
Build trust-aware applications with TrustGrade. Get your free API key — score any stablecoin, DeFi protocol, or smart contract in under 50ms. MCP-native for AI agent integration.