TLT short interest at 150.5M shares (01/2026, ShitMyChartSays chart via commonsenseplay): a baseline of ~20M from 2002–2013, spikes to ~35M (2013 taper tantrum), ~38M (2016 election), ~60M (2022 inflation peak), then a straight climb from the 02/2023 “soft landing bottom” through 100M (01/2025) to the record. The thesis built on it: “the biggest short squeeze of 2026 will be US bonds — Bessent is manufacturing it right now,” with Treasury’s long-end buybacks (Treasury buybacks are not debt reduction) as the squeeze trigger. The poster is 40% long TLT and says so, which is more than most.
Three readings before believing the squeeze. TLT’s shares outstanding grew several-fold over the same period, so a chunk of the “record” is Records in a growing series — short interest as a share of float is the honest series and isn’t shown. A large part of ETF short interest is not a directional bet: it is duration hedging, basis trades and create-to-lend, which do not cover in a rally. And the mechanism that would squeeze — a state buyer meeting a crowded short — is real but small ($4B a shot vs the stock, see Chanos in Treasury buybacks are not debt reduction); the Stock-bond correlation flips with the shock type regime says the flight-to-quality bid that used to do this job is unreliable. What the chart does establish: long-duration Treasuries are the consensus short of the cycle, and consensus shorts are where Most-shorted baskets outperform does its work. A crowded position marks risk, not timing — same rule as Call-volume extremes on the other side.