Long-end yields, not the policy rate, are doing the Fed’s tightening. AI and memory equities bounced after a forced seller cleared, but credible voices dispute whether that marked a bottom. Contracted memory, power and optical demand survived; financing quality now separates exposure.
Jim Bianco argues that the Fed’s hold left bond traders to tighten conditions themselves. MacroVoices remains bearish on TLT, but gives a clean invalidation signal: bonds that stop falling on a hot inflation print or weak auction would warn that crowded duration shorts are exhausted.
Jim Bianco sees the Situational Awareness block sale as the visible, exhaustible seller that often accompanies a low. CryptoCondom says the fund was a symptom of leverage plus AI capex and revenue anxiety, not a market-clearing event. Korean evidence leans fundamental: one interview cited roughly 70% five-year Samsung LTA coverage, while unaudited Reddit excerpts reported firm long-term commitments and memory tightness through 2027. A second-hand FT report still expects a slow recovery because volatility deters institutions.
Morgan Stanley attributes much of the AI-infrastructure selloff to technical pressure. The harder evidence is lower in the stack: Entergy described Meta’s five-gigawatt Louisiana site and a contract that makes data centers bear incremental and fixed costs; Quanta showed record backlog and cash generation; AXT reported record indium-phosphide revenue; and Lam’s 2027 setup carried unusual multi-year visibility. An unaudited Reddit post sharpened the cash-conversion split between Microsoft and Meta.
Mark Minervini initiated a small NET long for relative strength while retaining SPY as a portfolio hedge, a cleaner signal than broad dip-buying. Separately, QTR kept APP and PLTR on an explicit avoid list, and Fidenza Macro disclosed a June exit from AI semiconductors and infrastructure into cash. The common thread is position specificity rather than blanket risk-on.
Mark Minervini explicitly initiated a small NET position on relative strength and constructive price action; this was a disclosed trade, not an inferred purchase.
The forced seller may have cleared, but high-quality voices disagree on whether it marked a durable low; a second-hand FT report expects Korean institutional re-entry to remain slow.
Credit-aware commentary favors assets near credible contract value; Quanta backlog and AXT’s InP milestone pass a harder evidence test than sector beta, but neither post disclosed a purchase.
Minervini retained an SPY hedge while buying selected individual setups, keeping index caution distinct from stock selection.
QTR seconded repeated warnings to stay away from APP and PLTR; this was an explicit avoid stance, not a disclosed short.
Fidenza Macro disclosed exiting the portfolio in June and moving to cash before declaring the equity bull market topped; the author claim is unaudited.