Apex is a non-custodial lending market: depositors supply assets to earn yield; borrowers post collateral and pay interest on what they borrow. Total value locked reached $1.2B at the data cutoff^[01], up 240% over twelve months^[02]. The growth is almost entirely supply-side: utilization (the share of deposits actively borrowed) sits at 34%^[03], against the 60–75% range typical of mature lending markets such as Aave V3 and Compound V3. Apex funds the 7.3-point spread between 2.1% base yield and 9.4% total depositor yield by distributing APX tokens at 4.2M per month^[05] (approximately $19M per year in native token dilution at spot price). Data cutoff: 2026-05-18.

An estimated 64% of deposits are there for the token rewards, not the lending product: deposits above the point where borrower interest alone would retain them^[06]. Comparable episodes (Compound's COMP farming epoch, November 2021, and Anchor Protocol's 20% yield era) produced 55–80% TVL loss within two quarters of similar reward cuts^[07]. In Q3 2026, monthly APX emissions fall from 4.2M to 1.9M tokens, a 54% cut^[08]. At current utilization, deposit yield falls to approximately 3.7% after the cut^[09] — below every major lending comparable and below the risk-free rate, the threshold at which incentive-seeking capital has no reason to stay.

Sustained utilization above 55% for two consecutive quarters after the cut would signal genuine borrower demand capable of sustaining competitive yields without the subsidy^[10]. APX token price is a secondary variable: the yield calculation assumes the current APX price; a 40% APX decline pushes effective total yield below the base rate even before the cut takes effect^[11]. The borrow side is also concentrated: 71% of interest revenue comes from a single collateral asset^[10], so the recovery depends on one asset holding up.

We do not yet know whether borrowers are net beneficiaries of the same APX emissions they help fund^[11]. If borrowers offset their 7.2% gross borrowing cost with APX rewards, removing the subsidy may compress both sides of the balance sheet simultaneously — a utilization collapse, not a partial deposit exit. Apex's Q3 documentation contains no published utilization targets, leaving the key variable unquantified.