HSM: Delist sUSDS & sUSDe, consolidate into aUSDT/aUSDC, realize ~$120k of HSM profits to Treasury

Summary

This referendum winds down sUSDS and sUSDe as Hollar Stability Module (HSM) collaterals, consolidates HSM liquidity into the remaining two collaterals (aUSDT and aUSDC), and realizes the majority of the profits the HSM has accrued (~$120k of ~$130k) to the Treasury. The entire lifecycle — freeze, conversion, profit skim, and final removal — executes from this single referendum; no follow-up proposal is required.

It is part of the broader wind-down of sUSDS/sUSDe on Hydration (the HUSDS/HUSDe money-market pools will be delisted in a separate proposal).

HSM profits

The HSM mints HOLLAR 1:1 (in USD terms) against collateral at the moment of purchase. Its HOLLAR liability is fixed at mint, but sUSDS and sUSDe are yield-bearing — their USD value has appreciated since the HSM acquired them. That difference is protocol profit.

Current state (17 July 2026):

	Amount	Value (USD)
sUSDS holdings	74,264.6	~$81,933 (rate 1.1033)
sUSDe holdings	191,212.4	~$236,738 (rate 1.2381)
aUSDC holdings	174,517.1	~$174,517
Total backing		~$493,188
HOLLAR minted by HSM (facilitator bucket level)		363,407
Surplus (profit)		~$129,781

This referendum transfers 97,000 sUSDe (~$120,095, ≈92% of the surplus) to the Treasury. The remaining ~$10k surplus stays in the HSM as an over-collateralization buffer.

What the referendum does

At enactment

1. Freezes HOLLAR minting against sUSDS and sUSDe (max_in_holding = 0). Redemptions (HOLLAR → sUSDS/sUSDe) remain open.

2. Transfers 97,000 sUSDe from the HSM to the Treasury.

3. Starts three DCA schedules, each trading once every 10 minutes (~3 days total, ~432 trades each), sized small (~$190–280 per trade) to minimize price impact:

   • HSM: remaining sUSDe (~94,200) → aUSDT
   • HSM: sUSDS (~74,265) → aUSDC
   • Treasury: 97,000 sUSDe → aUSDT

   The HSM schedules are open-ended and self-terminate when the balances are exhausted, so any balance drift between now and enactment is handled automatically. Each trade is protected by a 1% max slippage vs oracle and an absolute floor at 99% of the current peg.

Day 5 (scheduled on-chain by the same referendum)

4. The Treasury converts its accumulated aUSDT (~120k) into HOLLAR through the HSM at 1:1 — the aUSDT becomes HSM backing, and the Treasury books its profit in HOLLAR.
5. Residual sUSDS/sUSDe dust in the HSM is swept to the Treasury, and both assets are removed from the HSM collateral list.

Resulting state

• HSM collaterals: aUSDT and aUSDC only — ~$493k of backing against ~483.5k HOLLAR minted, keeping a ~$10k buffer.
• Treasury: ~120k HOLLAR of realized HSM profit.
• HOLLAR redemption remains available throughout the transition: against sUSDS/sUSDe until those holdings drain, against the existing ~175k aUSDC buffer from day one, and against aUSDT as the conversions accumulate.

Execution safety

• The full preimage was dry-run on a chopsticks fork of mainnet: all 8 calls execute successfully.
• Each DCA route was executed against forked live liquidity to confirm pricing (sUSDe → aUSDT at 1.2383, sUSDS → aUSDC at 1.1019, both within a few bp of peg).
• The day-5 cleanup call was separately dry-run against a simulated post-DCA state: dust sweep and both collateral removals succeed.
• The cleanup uses force_batch, so no single failing item can block the rest; in the worst case (a DCA still running at day 5) the removals simply wait for a trivial follow-up.

Preimage hash: 0x6537b64f5cb16b2b558e42cd68f4c870a5e290b77ff45331991cf756e8497afd (904 bytes, Root track)