1inch Aqua Incubator 25% to the DAO · from the first dollar

A lending position
becomes AMM inventory.

Earning LP fees on top of lending yield — without the position leaving the LP’s wallet.

1delta CreditMaker · Aqua Strategy01
What it unlocks

One position, three things it could never do at once.

01 · Yield while quoting
Capital does both jobs
Aqua’s thesis is capital in the LP’s wallet backing more than one strategy. This is that thesis taken to lending: the same capital keeps its yield and quotes a spread.
02 · RWA & illiquid collateral
Credit becomes the exit
Whole asset classes have deep borrowing markets and almost no spot market. The asset’s own borrowing capacity becomes its secondary market — no pre-funded inventory required.
03 · Strategy flexibility
One mechanism, many books
Deposits, debt, or collateral against debt — the same position quotes whatever a lender’s configuration permits, without a new adapter per market.
$3.35BFluid's 30-day volume on $283M of lending-backed TVL — 11.8× turnover. We model 5×.
$24.9BHeld by Aave, Morpho, Compound, Euler and Spark on our launch chains.
4.5%Of flow that already trades on Base is all our base case needs.
Fluid proves the model — but quotes only pairs Fluid itself lends against02
The mechanism

The LP's own account is the maker.

01ship() runs no solvency check — the strategy ships a virtual balance for buffer + withdrawable + borrowable, not tokens held.
02pull() has no maker-side hook — one shared posManager sources the token inside the swap.
03Out: buffer → withdraw → borrow.   In: repay → supply → buffer.
04Nothing is wrapped or migrated. Reach needs authorization and opt-in and an active strategy — any one revocable.

No owner, no upgrade path, no generic execute. The LP delegates an action, never the position.

The position stays where the LP put it03
What a position can quote

The lender's own configuration decides which pairs are tradeable.

What tradesCollateral sideDebt sideLive market
spark
Deposits move
supply rotates between collaterals
wstETH⇄WETH⇄cbBTC+3
none
Spark · standard
85% max LTV
aave
Debt moves
one collateral, two loan assets
syrupUSDT
USDT⇄GHO
Aave v3 · syrupUSDT e-mode
90% max LTV
aave
Collateral moves
seven LSTs, one loan asset
wstETH⇄weETH⇄rETH+4
WETH
Aave v3 · ETH-correlated
93% max LTV
aave
Both move
both sides rotate
USDe⇄sUSDe
USDT⇄USDC
Aave v3 · sUSDe stables
90% max LTV
morpho
Collateral against debt
taker sells → we supply + borrow
taker buys → we repay + withdraw
AA_FalconXUSDC
⇄USDC
Morpho · isolated market
77% max LTV

All five are live today, and we did not design any of them — the lenders did. The quoter reads each market's LTV and free liquidity and offers exactly what that configuration permits. How much depth sits behind them comes next.

Same mechanism · the lender's config decides the tradeable set04
The pricing domain

Any curve can price it. Solvency decides where.

A curve priced off pool reserves has no notion of solvency — it will trade a leveraged position toward liquidation without noticing. Priced off the lender’s own health factor, it cannot.

borrowable  = C·P·LLTV_safe − Dmore debt it can take
withdrawable = C − D/(P·LLTV_safe)collateral that can leave
freeboard   = 1 − D/D_max1 with no debt · 0 on the ceiling

Both are their own maximum × freeboard, so freeboard is the whole state. Any bonding shape can sit on it — flat and tight for correlated pairs, wider and convex for uncorrelated — but the health factor bounds where any of them may operate.

Solvency is a property of the state space, not a check bolted on05
The five rules

Any curve over that domain must obey these.

  • Quote zero at the bound. Both capacities hit zero together — it stops quoting rather than sell into liquidation.
  • Deepest when safest. Size on offer shrinks as the position turns directional.
  • Deleveraging is cheaper. Unwinding prices better than extending, so it mean-reverts with no keeper.
  • Steep near the edge. Curvature makes pushing toward liquidation unprofitable, not merely blocked.
  • Spread covers cost. Freeboard consumed, borrow rate, oracle staleness.
Not constant-product · not concentrated-liquidity · not stableswap · not a fee mechanism · no licensed math06
The domain, drawn

Take Uniswap v3’s CLAMM. Change only the domain.

0% 25% 50% 75% 100% mid −60 −30 +30 +60 price offset from mid (bps) · depth as % of the position’s capacity deleveraging — repay + withdraw levering — supply + borrow what a plain CLAMM would still quote freeboard bound — no size beyond

The resulting book, drawn the way a v3 pool is drawn. A plain CLAMM is symmetric about mid — the dashed outline is the depth it would still be quoting on the right. Over credit capacity the book is asymmetric: deeper where the trade deleverages the position, thinner where it levers, and zero five ticks out where freeboard runs out. Three of the five rules, drawn rather than asserted.

Stylised CL shape · the curve families are chosen in M1 and backtested in M307
The evidence

Exit cost is a cliff, not a slope.

100% 75% 50% 25% 0% $100k $250k $1M $2M $5M $10M syrupUSDC · Base USDe · Base BOLD · Ethereum

Quoted against a live aggregator, both directions. Small trades route fine — what fails is leaving at size.

Measured by quoting exits, not by reading pool TVL · 28 Jul 202608
The gap

Lending depth dwarfs the venue a holder can leave through.

Base & EthereumLending collateralLargest single exit
USDe · Base92% LLTV on Morpho · $26.1M free to borrow · spot trades $6.2k/day $265.1M ~$1.2M  ·  221×
BOLD · EthereumLiquity v2 — mint capacity is not rationed by other depositors $71.3M ~$7.9M  ·  9×
syrupUSDC · Base90% LTV in its USDC/GHO E-Mode · spot trades $454/day $12.4M ~$140k  ·  89×

Bridging and issuer redemption exist — but take minutes to days and cannot be reached from inside a swap.

The asset's own borrowing market becomes its exit venue09
Why 1delta

The hard part is the lending integration — and it already runs.

1inch has already judged this
1delta Unite won first place in 1inch's "Expand Limit Order Protocol" track at ETHGlobal Unite DeFi — opening margin positions against Aave v3 and Morpho Blue, gasless, fillers holding no inventory. The same composition, moved from order flow to a continuous quoting surface.
Production since 2023
40+ chains · 300+ lending markets · 150+ deployments · $200M+ processed. Aave v2/v3/v4, Compound V2/V3, Morpho Blue, Euler v2. Grants from Aave, Compound, Polygon and Mantle; Compound integration audited by Certora.

A team without this spends most of a $50k grant building adapters. We start at the curve.

ethglobal.com/showcase/1delta-unite-kyoz410
Milestones

Twenty weeks, gated on criteria you can check.

wk 0 wk 5 wk 10 wk 15 wk 20 M1 Idea verification $2,500 M2 Proof of concept $5,000 M3 Working implementation $17,500 M4 1inch integration & mainnet $25,000

Every gate is a number or a binary: an auditor accepts the spec · no swap fails · every decision point tested both ways ≥90% · no fill worse than quoted.

M4's seven weeks exclude audit queue time11
Economics & ask

25% of gross fee revenue, from the first dollar.

Six months after M4 · reported, not gatedBaseUpside
Monthly volume$15M$40M
Gross fee revenue$11,100$32,200
To the 1inch DAOa percentage of what is actually earned — not a minimum$2,775$8,050
Grant payback18 mo6 mo

$50,000, milestone-gated — half releases only against a live, audited, aggregator-routing deployment, and the DAO’s share carries the standard three-year buyout at fair market value. The return that matters is $15–40M/month routed through 1inch that does not exist today.

1delta · achim@1delta.io · 1delta.io12
← → move · N notes · F fullscreen