Yield Audit, Episode 9: @Neutrl's NUSD & sNUSD 4.73% APY on the staked side. $134M TVL, down from $234M three months ago. No depeg drama. No funding rate flip. Just a steady contraction that reveals more about DeFi capital flows than any bull‑market AUM number ever could. Here's what happened, where the yield comes from, and the one thing in the staking data almost everyone misread. Neutrl's sNUSD is a synthetic dollar that earns yield in an unusual way. It buys locked VC tokens at a discount, instantly hedges the price risk, and captures the spread. You deposit USDC, USDT, or USDe to mint NUSD—a stable, spendable token. Stake it into sNUSD, and you earn a share of the profits. The mechanism is a basis trade at scale. The protocol buys illiquid tokens (SAFTs, treasury bags) at 10-30% below market, then shorts an equivalent amount on perp exchanges. When the tokens unlock, the short is closed, the spread is realized, and the profit flows to stakers. No token emissions. No inflation. Just a classic hedge‑fund strategy inside an ERC‑4626 vault. Where the yield comes from. Three cylinders, one core. - OTC deal discounts are the main engine. Buy cheap, hedge, lock the spread. - Funding rate arbitrage provides a tailwind. The protocol is short, so it often collects payments from longs. Historically that's added 7–10% on major assets. - Idle collateral sits in sUSDe, USDS, and other yield‑bearing stablecoins, earning a base rate. In the private beta, yields hit 26–30%. Then TVL grew to $234M. More money chased the same limited pool of OTC deals, compressing the spread. You can see it in the numbers. The weekly APY fell from 14.4% in December → 10.2% in February → 8.4% in March → 7.5% in April → 4.73% today. Not a protocol failure. The price of success. Safety and risks. Neutrl is over‑collateralized at 102%, with a $4.8M reserve fund. Collateral basket: 66% sUSDe, 33% USDS, a little USDC. Diversified, but heavily reliant on Ethena and Sky. Four risks matter for a staker. Counterparty risk. If a VC or foundation fails to deliver the locked tokens, the protocol has an unhedged short. Mitigation: triparty escrow via Fireblocks, Copper, Ceffu. Exchange and margin risk. Shorts on Binance or Deribit could get liquidated in a spike. Mitigation: off‑exchange settlement (MirrorX, Clearloop) + 24/7 desk. Lockup and exit friction. This is the one you'll feel. Unstaking requires a 10‑day cooldown, then a 48h redemption queue with KYC/KYB. Skip it with a 1% instant fee, or sell NUSD on Curve. But the Curve pool has just $4.86M TVL and $3M effective depth. In a panic, the exit narrows fast. Depeg risk. NUSD is at $0.998, a tiny discount. The peg is backed by redeemability, not an algorithm. Because of the cooldown, arbitrage can't enforce the peg instantly, so it could widen under stress. The biggest dip recorded was a brief 91bp on June 17, quickly absorbed. What @PharosWatch sees. Pharos rates NUSD C+ (63/100). The breakdown is more telling than the grade. Peg stability: 99/100. Held through a 24% supply contraction in a month with only shallow dips. Rare. Exit liquidity: 64/100. One Curve pool, functional but concentrated. Redemption is a whitelisted KYC queue—not permissionless. That friction protects the peg and caps the exit score. - Resilience: 60/100. Over‑collateralized, reserve fund adequate for moderate stress, but concentration in sUSDe and USDS is a real risk. If Ethena or Sky depegs, Neutrl takes the hit. - Decentralization: 48/100. Single admin key controls minting and denylist. KYC‑gated redemptions. Contracts open, control structure not. - Dependency: 75/100 (best metric). 93.5% of collateral is liquid stablecoins, so the peg engine has a deep, reliable pool. The early warning system is flipping between Calm and Watch. Yield anomaly signals are active. The safety grade slipped from B to C+ as supply contracted and yields fell. Peg works. Capital keeps leaving. Both true. What I found: the staking ratio was a mirage. The % of NUSD staked rose even as APY crashed, from 42% in December to 78% in May. That looked like conviction. It wasn't. Here's what really happened. Late 2025, unstaked NUSD was $123M. Only 12% was staked. Early 2026, that hot money fled. Unstaked fell $123M to $42M, while sNUSD surged. The staking ratio jumped because the denominator collapsed, not because more people locked up. From March, stakers started leaving. sNUSD dropped $184M to $46M. But the ratio stayed high because unstaked NUSD was shrinking too, just slower. The ratio was a ghost. Then June. APY slipped below 5%. Stakers capitulated. The ratio crashed 78% to 52%. Total supply: $134M. Lesson: never trust a staking ratio without the absolute numbers. Neutrl is well-engineered. It does something hard: turning illiquid discounts into non-directional yield. Institutional-grade risk framework, strong backers, real-time attestations. But the yield is 4.73%. Barely above T-bills. And it comes with a 10-day cooldown, a 48h queue, KYC gates, and heavy reliance on Ethena and Sky. The peg was the star. 99/100, held through a brutal contraction. The mechanism works. What broke is the yield, not the dollar. For most stablecoin holders, there are simpler, more liquid options paying the same or better. Rating: 6.0 out of 10. It's a sophisticated product with yield that became a victim of its own early success. Infrastructure solid, peg proven under stress, but the return no longer covers the lockup. Watch, don't touch, until double digit APY returns.