Aave's 4.5% GHO borrow rate aims to help refill its depleted USDC module as borrowers swap in stablecoins

TokenLogic, the DAO service provider behind the change, lists two places a borrower can get the GHO needed to repay: the secondary market, or a GHO Stability Module that swaps USDC or USDT for it ^[4]. Only the swap leaves stablecoins in a reserve that another GHO holder can later redeem against ^[5]. A market purchase can support GHO's price, and the loan still gets repaid without any USDC reaching a module ^[5].

Holders of sGHO are on the other side of that reserve. Aave's documentation says the vault redeems shares for GHO with no cooldown and does not rehypothecate deposits ^[10], subject to a pause state and per-user withdrawal limits ^[11]. A saver who wants USDC needs a separate conversion, and it works only against inventory and liquidity the module actually has ^[12]. Aave Labs left USDC modules out of its Sept. 24 inventory count because their redeemable balances were negligible ^[17].

The counted dollars were USDT, and most of them were not on Ethereum. Aave Labs put 19.2 million USDT on Ethereum and 40.7 million on Plasma, 59.9 million in all ^[14], so about 68% sat on Plasma ^[20]. Using that inventory means bridging GHO over Chainlink CCIP and turning the module's assets into usable stablecoins, and the time both steps take decides how useful the route is ^[15]. Kairos Research, working from Sept. 8 readings, counted 40.6 million of nominal Plasma redemption inventory against 38.6 million in the underlying lending pool ^[16], a 2 million difference ^[21]. TokenLogic's Oct. 2 update cited about 22.5 million USDT in a USDT module without naming the network ^[6]. Neither TokenLogic nor Aave Labs supplied matched Oct. 5 balances ^[7].

The levers Aave controls set room and price. Under the RemoteGSM design TokenLogic described in March, facilitators supply preminted GHO to a GhoReserve and modules draw on it within assigned limits; a higher limit permits more incoming swaps, but users still have to deliver the USDC or USDT ^[13]. The rate is the same kind of lever. TokenLogic's own expectation of a refill is conditional on borrowers getting their repayment GHO through the modules ^[1].

Net Core debt fell 0.2 million GHO between Oct. 2 and Oct. 5, about 0.17% of the book ^[19], and the new rate had been live for between one and two days at the last reading ^[22]. The figure is net, so gross repayment could be larger if new loans offset it.

If borrowers swap USDC into the module to repay, savers get a dollar exit back. Buying on the market instead supports GHO's price and leaves the module depleted ^[5]. Borrowers who ran the old loop, paying 4.25% to earn 4.5% ^[2], have a third route: redeem their own sGHO and repay with the GHO it returns ^[10]. Debt and deposits then shrink together and no stablecoin moves.

I think that third route is the likeliest for whatever share of Core the old spread attracted, because it is the cheapest unwind and needs no outside dollars. The spread was small money for the DAO either way. If all 116 million of Core debt had run the loop, 25 basis points comes to about 0.29 million GHO a year ^[18]. The counter-case is that borrowers who need fresh GHO and already hold USDC find the module the most direct source, and even a handful of those swaps would refill a balance TokenLogic called depleted ^[1]. A rising USDC module balance alongside falling Core debt would show that, and would make this view wrong.