July 25, 2026 · Decision research
Crowding Positioning
Finding
The crowded-long semis trade is unwinding in real time; the SOFR-short and dollar-long crowded trades are NOT, and the implicit "both will unwind on a dovish surprise" coupling is the weakest assumption in the sweep.
What changed
The PHLX Semiconductor Index is down ~18% in July 2026 alone, with the Intel-led leg driven by AI-capex skepticism, yet BofA's July FMS still shows 82% naming "long global semis" the most crowded trade (a record, third straight month) and 0% reporting a short. Leveraged-fund SOFR shorts rose to a fresh record 2.97M contracts (~US$700B notional) through June 23, while FedWatch still implies only ~22% odds of any cut through year-end and J.P. Morgan expects zero cuts in 2026 plus a hike in September 2027.
Decision
Keep "consider," but narrow it to the trade that the data actually supports, defensive on long semis (trim, hedge, or pair), and DO NOT pre-emptively fade the SOFR short or the dollar long until a Fed dovish catalyst surfaces. Implied probability of the catalyst is too low to justify the size of the position.
Watch next
(i) July 29–30 FOMC statement + press conference and any SEP revision; (ii) FedWatch implied path, a decisive move above ~70% for a December cut is the trigger for fading SOFR-short crowdedness; (iii) SOX basing action: a close back above the July highs on rising volume with hedge-fund net long crowding extending would invalidate the bearish positioning thesis; (iv) weekly CFTC TFF report (next print August 1) for changes in leveraged-fund SOFR net; (v) BofA Bull & Bear Indicator, any rise above ~8.0 would reinforce that positioning risk is still elevated but capitulation is not yet visible.