TL;DR
- What it is: Leveraged stablecoin carry on Solana — supply USDe (Ethena) on a Sentora-curated Kamino market or Bitwise-curated Jupiter Lend market, borrow USDG (Paxos), swap back to USDe, repeat up to ~12.5× at 92% LTV.
- Yield source: sUSDe yield (now ~4%, down from 15%+ in Sept 2025 as Ethena rotated from perps to RWA/credit) minus an admin-set USDG borrow rate (~2% realized), times leverage — headline ~19% APY at 9.5×.
- Biggest risk: USDe is short-vol carry — negative funding regimes drain cashflow and can de-peg the collateral, while a hardcoded USDT=$1 oracle means the market never liquidates, only amplifies (Risk #1).
- For detail: Ethena Solana lending proposal, RedStone Solana Lending Markets 2025-12-11, Silvio Busonero 2026-05-21.
How it works
The looper deposits USDe as collateral on either the Sentora-curated Kamino V2 isolated market or the Bitwise-curated Jupiter Lend (Fluid framework) market, borrows USDG at a low admin-capped rate, swaps USDG→USDe on a Solana DEX, re-deposits, and repeats. At 92% effective LTV the convergent leverage is 1/(1−0.92) = 12.5×; realized average is ~9.5×. The net USDe is then staked into sUSDe to harvest Ethena’s blended portfolio yield. Both markets use a hardcoded USDT=$1 oracle on both legs, so intra-market USDe or USDG depegs never trigger liquidation.
Ethena is the sole lender of USDG into both markets, using USDG it holds on its own balance sheet ahead of formal backing-asset designation. The USDG borrow rate is therefore an admin parameter, not a clearing price — Ethena can re-pin it unilaterally and invert the carry overnight. The Kamino market is at its $200M USDG borrow cap and fully utilized; entry is queued for the next curator-approved cap raise.
Both stablecoins reach Solana via LayerZero OFT. USDe is not natively issued on Solana — the ~$567M Solana supply exists only as a LayerZero OFT representation, with mint authority held by the OFT manager program 9S6fwN...S2y9. USDG cross-chain flows route through LayerZero as well. Ethena’s own re-rate-down triggers explicitly include “LayerZero OFT pause.” Exit requires unwinding the loop into Solana DEX depth: Ethena’s risk-committee analysis shows the USDG/USDC route clears $10M at ~0.2% slippage but cliffs to 60.9% slippage at $25M, with $31.85M aggregate USDG depth across 12 pools — against $200M+ borrowed on Kamino alone.
Risks
| # | Risk | Mechanism | Source + our delta |
|---|---|---|---|
| 1 | USDe is short-vol carry that can de-peg | Funding-rate-dependent collateral. Negative funding regimes drain sUSDe cashflow; sustained drawdown reduces USDe backing value. Hardcoded USDT oracle means the market never liquidates — borrowers exercise asymmetric-default option, Ethena absorbs magnified loss. Oct 2025 Binance flash-depeg printed $0.65 during $19B cascade. | Ethena gov reserve overhaul (Apr 2026, perps 93%→11% of backing); CoinDesk 2025-10-11. We treat this as design-level disqualifying, not parameter-tunable. |
| 2 | LayerZero OFT bridge surface on both legs | USDe-Solana is entirely an OFT representation of Ethereum USDe; USDG cross-chain also routes via LayerZero. OFT manager 9S6fwN...S2y9 is upgradeable; DVN/executor compromise mints phantom USDe-Solana or freezes the representation. Ethena’s own re-rate-down triggers include OFT pause. | On-chain primary: tx uqXry...J4LR shows Send → Endpoint 76y77p...gjEn6 + QuoteDvn; cross-referenced against LayerZero’s deployment registry. Issuer treats bridge pause as a live operational lever; we treat it as a third-party-bridge concentration we won’t underwrite. |
| 3 | Carry inverts on a single admin change | Ethena is sole lender; USDG borrow rate is admin-capped (realized ~2%, governance schedule 3.3% @ 90% util → 10% @ 99.9%). At 9.5× leverage and the governance-proposed 99.9%-util rate, net carry is (4% − 10%) × 9.5 = −57%. No advance signal — discrete-jump risk at full utilization. Exit if borrow rate rises above sUSDe yield. | Ethena gov proposal IRM schedule vs Kamino realized rate (2026-05). Article claims “capped at 2%”; the gov schedule does not support this past launch window — we flag the discrepancy. |
| 4 | USDG/USDC Solana DEX depth cliffs at $25M | $10M clears at <0.2% slippage; $25M at 60.9%; aggregate $31.85M across 12 pools. Against $200M+ borrowed on Kamino at launch, unwind is structurally insufficient by an order of magnitude. Exit requires Ethena’s cooperation (raising IRM), cross-chain bridge-out (LayerZero risk), or accepting catastrophic slippage. | Ethena risk-committee proposal documents the depth analysis itself; we read it as a self-disclosed exit-liquidity gap. |
| 5 | USDG-Solana single-key seize authority | permanentDelegate + freeze authority is 2apB...YJjk, a System-Program-owned account whose recent txs do not route through Squads — most likely a plain ed25519 key. Same address is one of four signers on the 1-of-4 SPL mint multisig — can mint AND seize. Affects every USDG holder, including the borrowed leg. | On-chain getAccountInfo 2026-05-25 — primary verification. Paxos is MAS-regulated; risk is operational/compelled-action, not rug. |
| 6 | Reflexive system at the strategy level | Ethena lends its own backing asset (USDG) to enable leverage on its own collateral (USDe). Up to 12.5× recursive leverage. ~99% of Solana USDe and ~75% of Solana USDG sit inside these two markets — the markets nearly are the asset on Solana. Any disruption is co-extensive with the entire Solana presence. | On-chain supply reads 2026-05-25 vs article figures; LlamaRisk supports phased rollout from inside Ethena’s Risk Committee (conflict-of-interest discount applies). |
Liquidity & exit
Primary unwind requires USDe→USDG swaps on Solana DEXs to repay debt at each loop iteration. Ethena’s own risk-committee documents USDG/USDC route at $10M = <0.2% slippage (2026-05), USDG/USDC route at $25M = 60.9% slippage (2026-05), Direct USDG/USDT at $5M = 31.9% slippage (2026-05), Aggregate USDG Solana depth across 12 pools = $31.85M (2026-05). Backup path is cross-chain unwind via LayerZero OFT back to Ethereum to sell into deeper USDC liquidity — adds bridge risk and multi-day execution, and Ethena’s own re-rate-down trigger fires on OFT pause. Worst-case latency is unbounded: at 100% utilization with Ethena as sole lender, there is no FIFO queue to wait in — exit depends on Ethena either raising the IRM curve to dis-incentivize borrowers or unwinding its own lender position. Defensible size at current liquidity (post-haircut): $0 — the $25M slippage cliff (3× current haircut → effective ~180% slippage well past 500bps) breaks any institutional sleeve, and the strategy ceiling sits below the strategy’s monthly yield once the round-trip is priced in. Closed venue, no stress scenario needed.
External reviews
Source links for the third-party takes in the Risks table. Per rubric.md, reviews older than 90d (⚠️) are stale — refresh before relying.
- LlamaRisk — USDG-as-backing assessment · 2026-05-07 · on Ethena’s Risk Committee (discount)
- Ethena governance — Solana lending proposal (Kamino + Jupiter) · 2026-05-07 · issuer
- RedStone — Solana lending markets report · 2025-12-11 · ⚠️ stale
- CoinDesk — Oct-2025 USDe flash-depeg coverage · 2025-10-11 · ⚠️ stale
- Silvio Busonero — structural analysis of the loop · 2026-05-21 · independent, not quantitative