CRDO's margin-scare damage from the September 1 print kept compounding — its 30-day return fell to -39.23% from -24.07% last cycle, now -7.4% below its 200-day — while SMCI's one-year deficit widened back out to -8.76% after nearly closing to -2.73% last cycle, with no fresh news receipt to explain either move this session. BE and OKTA both extended strongly (BE 1Y +309.79%, vs 200-day +41.6%; OKTA 1Y +81.06%, vs 200-day +60.6%). CEG turned genuinely weak — negative on the year (-10.46%) and below its 200-day (-3.6%) for the first time in this tracked window, worth a second look given the name's AI-power relevance.

Data as-of the 2026-09-11 close. Source: tech-insider-buys — price/technical fields only; this scan does not carry independent insider dollar-flow figures (see companion insider-scan for overlapping names' Form-4 data). Prior scan: 2026-09-04-tech-insider-buys.

CRDO's September 1 margin-scare damage kept compounding rather than stabilizing. ($162.95, RSI 33.2, 30D -39.23%, was -24.07% last cycle; -47.21% off its 52-week high; vs 200-day -7.4%, was -2.7%) — 2026-09-01-crdo-earnings still confirms the underlying beat (EPS $1.16 vs $1.03 consensus); the drop remains sourced to Yahoo Finance's own coverage of margin and guidance concerns flagged last cycle. This desk's current news pull (fetched 2026-09-08) carries no fresh update beyond aggregator quote pages, so the continued slide this week is reported without a new sourced cause — the original margin-scare driver is the best available explanation, but it does not by itself account for a further -15-point deterioration in the 30-day window.

SMCI's near-closed one-year deficit reversed back open, with no sourced driver for the move. ($40.10, RSI 60.5, 30D +6.62%, 1Y -8.76%, was -2.73% last cycle — widened again; vs 200-day +27.3%, was +26.2%) — this desk checked latest (fetched 2026-08-28, ok, not stale but predates this week's move) — the most recent item is an August earnings-results release, no fresh catalyst. Reported as an unexplained short-window swing, not glossed over.

CEG turned genuinely weak this cycle — negative on the year and below its 200-day. ($284.75, RSI 53.7, 30D +2.18%, 1Y -10.46%, vs 200-day -3.6%) — this is the first cycle in this tracked window where CEG shows both a negative one-year return and a sub-200-day price; prior cycles had it mixed but not both-negative. No stock-event or news check was run for CEG this session; flagged for follow-up given the name's relevance to the desk's AI-power/nuclear-fuel-cycle coverage.

BE and OKTA both extended their multi-cycle runs. BE ($275.75, RSI 65, 30D +16.27%, 1Y +309.79%, vs 200-day +41.6%) continues the large reversal tracked in the companion insider-scan — insider selling stays flat there, price-driven, not corroborated by fresh buying. OKTA ($166.50, RSI 58, 30D +13.03%, 1Y +81.06%, vs 200-day +60.6%) holds the widest 200-day cushion in this universe alongside TEAM.

MDB and PANW both cooled further off their September 1 confirmed beats, the same beat-then-sell-off shape flagged last cycle. MDB ($362.21, RSI 42.4, 30D -17.27%, 1Y +9.07%, vs 200-day +5.6%, was +7.7%) compressed its 200-day cushion further. PANW ($330.65, RSI 43, 30D -14.56%, 1Y +66.72%, vs 200-day +40.1%) stayed comfortably positive on the year despite the continued pullback — the mildest of the three beat-then-sell-off names tracked since September 1.

IONQ deepened its deficit further rather than repairing it. ($36.75, RSI 41.1, 30D -18.69%, 1Y -21.89%, was -6.15% last cycle; vs 200-day -16.6%, thesis state: intact-but-priced) — the widening gap here runs opposite to the narrowing flagged last cycle.

HUBS stays the deepest structural laggard in this universe. ($225.33, RSI 45.1, 1Y -54.64%, vs 200-day -12.7%, thesis state: intact-but-priced) — 30D +7.24% is a bounce, not a repair, against a one-year deficit that remains the widest tracked here besides IONQ's short-window collapse.

Table labels are technical states, not trade recommendations. thesis state: intact-but-priced (ARM, HUBS, IONQ) flags a name whose fundamental story the desk still credits but whose valuation or chart carries elevated risk — it is not a buy signal on its own. CRDO and SMCI are this cycle's clearest reminder that a sourced cause from a prior cycle does not automatically explain a continuation: CRDO's original margin-scare driver stands, but this week's further -15-point slide has no fresh receipt behind it, and SMCI's reversal has none at all. CEG's flip to both negative-1Y and sub-200-day is a genuine structural change worth a dedicated look given the desk's standing AI-power coverage, not just a line in a wide table. Per the desk's long-horizon discipline, BE and OKTA's unrated/no-flag status against +309.79%/+81.06% one-year returns is a reminder that thesis state coverage is incomplete in this universe — the absence of a flag is not itself a signal.

Capitulation (RSI < 25): None.

Oversold (RSI 25-40): CRDO (33.2). The single most damaged name in this universe by RSI.

Neutral / Strong (RSI 40+): All 41 remaining names, spanning AVGO (42.2, just crossed below its 200-day) to QCOM (66.3, near overbought).

• CRDO: does the slide stabilize, or does a fresh news pull surface a new driver for this week's further -15-point move?
• SMCI: the one-year deficit reopened with no sourced cause — check next pull for an explanation.
• CEG: first cycle both negative-1Y and sub-200-day — worth a dedicated news/event check given the desk's AI-power coverage.
• IONQ: deficit widened sharply (-6.15%→-21.89%) — confirm whether this continues or was a single-cycle swing.

• tech-insider-buys (2026-09-11)
• 2026-09-01-crdo-earnings
• latest (fetched 2026-08-28)
• 2026-09-04-tech-insider-buys (prior scan, comparison basis)