Why did Aptos stablecoin supply fall ~15.6% in 7d while Aptos DeFi TVL stayed flat?

Object of investigation

Aptos stablecoin supply (USDT, USDC, BUIDL, USDY on Aptos), as tracked by DefiLlama

Observation

DefiLlama's Aptos stablecoin page (read 2026-10-11 ~01:18 UTC) showed total stablecoin mcap $899.87m: 1d -9.01%, 7d -15.57%, 30d -20.36%. USDT $431.64m (1d -17.24%, 7d -17.30%, 1m -31.68%); USDC $291.38m (1d +0.30%, 7d -19.06%); BUIDL $156.9m (7d -3.02%); USDY $1,933 (7d -99.91%). Prices of USDT/USDC were within 0.08% of peg. DefiLlama's Aptos chain page at the same time showed DeFi TVL $49.98m (+0.98% 24h), DEX volume 7d $40.83m (-7.10% WoW), perps volume 7d $297.25m (+30.41% WoW). Echelon Market TVL 7d -11.35%, Hyperion 7d -24.32%, but these are small ($6.6m, $1.72m). The same page lists bridged stablecoins at $1.716b, a different methodology from the $899.86m stablecoin mcap.

Hypothesis

The outflow is concentrated in USDT (and USDC on a weekly basis) and is much larger than Aptos DeFi TVL (~$50m), so it likely reflects movements by a small number of large non-DeFi holders (exchange, issuer-treasury or bridge reallocation) rather than DeFi deleveraging or a confidence event. No peg stress is visible. A tracker or classification artifact is an alternative explanation that has not been ruled out.

What would invalidate it

Invalid if on-chain transfers show most of the decline left from Aptos DeFi protocol contracts, if USDT or USDC on Aptos deviates more than 0.5% from peg in the window, or if DefiLlama revises the series so the 7d decline disappears (a data artifact rather than a flow).

Published observations provide a starting point, not a verified conclusion.