← Research · knowledge/reports/2026-10-30-1100-semis-rout-pre-earnings-gap-risk.md
2026-10-30 11:00 ET: The chip rout at 11:05 (cluster 1 breaking lower while INTC stabilizes), why it reached NVDA/ANET/AVGO/ORCL, and gap risk into next week's semis prints (ON, AMD, ANET, SMCI, ARM, QCOM, MCHP)
Slot: Fri 2026-10-30 11:00 ET hourly research. Status: FINAL, ~11:00–11:20 ET; research partner (whiggity) input in §8. Screening/research, not investment advice.
TRAI data used: /api/latest = same 8:53 AM premarket run as the 09:00/10:00 reports (2026-10-29_premarket_2026-10-30T12-53-19-775Z; regime transitional, VIX 18.59 then). /api/events (generated 12:17 UTC today). /api/correlations (60d to 2026-10-22, stale, S-068).
Intraday quotes: Yahoo chart API 5-min bars, last bar 15:05 UTC (11:05 ET), vs the 10-29 close; OR30 = 9:30–10:00 range; VWAP computed from 5-min typical price × volume. Unofficial, may be delayed (Yahoo chart API).
Previous slots today: 09:00 covered the INTC print and premarket read-through; 10:00 covered candidate OR30 checks, telecom, LLY and the software split. This hour does not repeat the INTC print details. It covers how the semis complex is trading after the open, why it spread past PC-exposed names, and the gap-risk set-up for next week's chip prints.
1. Focus and why
Semis are the largest mover in the universe and are extending, not stabilizing: SMH −6.13% and SOXX −7.64% are both below their OR30 lows, while INTC itself (−27.99%) is back inside its range and +1.08% above VWAP. The selling has moved from the news name to the group. Seven cluster-1 names report in the next four sessions, so this is the set-up for next week's gap risk.
2. Cluster 1 at 11:05 ET (vs 10-29 close)
| Name | Last | OR30 range | vs OR30 | vs VWAP | Next print (TRAI /api/events) |
|---|---|---|---|---|---|
| INTC | −27.99% | −30.36…−25.32 | inside | +1.08% | reported 10-29 AMC |
| QCOM | −10.29% | −8.17…−4.98 | below | −2.18% | 11-04 AMC |
| ARM | −10.52% | −11.24…−8.87 | inside | −0.65% | 11-04 AMC |
| AMD | −9.43% | −10.42…−8.13 | inside | −0.07% | 11-03 AMC |
| DELL | −9.41% | −13.15…−7.24 | inside | −0.23% | 11-24 AMC |
| HPQ | −8.09% | −15.32…−8.42 | above | +1.82% | — |
| ANET | −7.01% | −7.75…−5.01 | inside | −0.18% | 11-03 AMC |
| SMCI | −5.69% | −7.00…−4.21 | inside | −0.22% | 11-03 AMC |
| AVGO | −5.36% | −5.29…−3.77 | below | −0.72% | — |
| ON | −4.81% | −4.56…−3.28 | below | −0.59% | 11-02 AMC |
| MU | −4.55% | −5.86…−3.08 | inside | +0.15% | — |
| MCHP | −4.27% | −3.57…−1.29 | below | −1.10% | 11-05 AMC |
| NVDA | −3.47% | −3.72…−2.18 | inside | −0.21% | 11-17 AMC |
| TXN | −3.02% | −3.06…−1.10 | inside | −0.81% | — |
Indices: SPY −1.34% (below OR30, −0.08% vs VWAP), QQQ −2.78% (inside), SMH −6.13% (below), SOXX −7.64% (below). Earlier 11:03 pass: AMAT −6.65%, LRCX −5.82%, ASML −5.15%, TSM −4.60%, MRVL −7.93%, ORCL −5.31%, VRT −4.66% (Yahoo).
Read:
- The two direct PC names that gapped hardest (DELL, HPQ) are bouncing off their lows (HPQ above its OR30 high, DELL ~3.7pp off its −13.15% low). The names extending lower are analog/MCU (ON, MCHP), handset (QCOM) and AVGO, i.e. the selling is spreading by "chip exposure", not by "PC exposure".
- QCOM is the weakest name vs its own range (−2.18% vs VWAP). MarketWatch ties it to handset and enterprise-hardware demand fears (MarketWatch, 2026-10-30). This fits the Counterpoint/IDC Q3 smartphone shipment data (−20%/−22% y/y) already in the 10:00 report §6. Tickeron says QCOM's move is sector-driven, not company-specific (Tickeron, 2026-10-30); I found no QCOM-specific news.
3. Why it spread past PC names
- Intel's guide (recap only): Q4 revenue $15.0–16.0B, non-GAAP GM 38% (from 44.1%), CFO sees the PC slowdown "continuing at least through the first half of next year"; data center revenue +78% to $7.3B but supply-limited by substrates; Israel fab accident costs ~$300M of Q4 revenue (Yahoo Finance/TIKR, 2026-10-30 09:39). Reuters (via Investing.com) calls it Intel's sharpest one-day decline on record, with >$137B of market value lost (Reuters via Investing.com, 2026-10-30).
- Second driver: AMZN. MarketWatch and Investing.com both pair Intel with Amazon's results as "fresh concerns around the AI trade" and doubts about the payoff from AI capex, after a week of GOOGL/AMZN/META prints (MarketWatch; Investing.com on AMD, 2026-10-30). That's the channel to NVDA, ANET, AVGO, ORCL and VRT, which have little PC exposure.
- Third: memory pricing. The PC/phone slowdown is blamed on high DRAM/NAND prices as suppliers shift capacity to HBM (Reuters). That's a cost for OEMs and a revenue tailwind for MU (record FQ4 revenue $54.2B, per the Yahoo/TIKR piece). MU is the cluster-1 name holding above VWAP (+0.15%), consistent with that split, n = 1.
- AMD-specific: it reports in two sessions, so holders are de-risking into the print (Investing.com). Analysts are leaning the other way: Jefferies reiterated Buy, $650 PT; Susquehanna raised to $650 from $620 (same source). JPMorgan estimates Intel's server-CPU revenue share fell from 77% to 53% since late 2022, so Xeon supply shortfalls are share AMD can take (Yahoo/TIKR piece). Counter-case is real but didn't help today.
4. Next week's cluster-1 prints: dates and TRAI reaction history
| Name | Date / timing | Confirmed by | TRAI last-4 avg abs gap | in ATR | fill rate |
|---|---|---|---|---|---|
| ON | Mon 11-02 AMC | TRAI /api/events (not separately checked) | 2.73% | 0.52 | 100% |
| AMD | Tue 11-03 AMC | Investing.com | 8.94% | 1.59 | 25% |
| ANET | Tue 11-03 AMC, call 4:30 ET | Arista IR, 2026-10-06 | 8.87% | 2.00 | 0% |
| SMCI | Tue 11-03 AMC | TRAI /api/events | 9.97% | 1.61 | 0% |
| ARM | Wed 11-04 AMC, call 5:00 ET | Arm newsroom, 2026-10-05 | 7.62% | 1.01 | 50% |
| QCOM | Wed 11-04 AMC, call 4:45 ET | Qualcomm IR, 2026-10-22 | 6.99% | 2.08 | 0% |
| MCHP | Thu 11-05 AMC | TRAI /api/events | 5.70% | 1.26 | 25% |
NVDA is 11-17 AMC, DELL 11-24 AMC. All are cluster 1 (avg ρ 0.588), so each print is also a sympathy event for the others. AMD's history: last four reaction gaps −2.72%, −11.20%, +15.26%, −6.57% (TRAI /api/events, AMD note).
5. Links to today's candidates
None of today's ten bucket names is in cluster 1 (CVNA, GE, CMG, DASH are singleton clusters in TRAI's 60d cut). At 11:05: GE fade-short −2.35%, below OR30 and VWAP (still working); HD/LOW still below OR30 lows; CVNA, MO, DASH, BA, CMG, GILD, CRM inside their OR30. The semis rout acts on candidates only through index beta (CVNA β 2.17 per the 09:00 report). Nothing in this hour changes gap or event risk on a candidate for the rest of today.
6. Analysis
- Breadth of the sympathy move scales with "chip-ness", not "PC-ness", by late morning. The direct PC OEMs are retracing while analog/MCU/handset/AVGO make new lows. If that persists into the close, a "second-order sympathy" pattern exists: names with an upcoming own print (ON, QCOM, MCHP, AMD) keep falling while the reported name stabilizes. n = 1; needs the event study in S-086.
- Pre-earnings drawdowns change the gap set-up. AMD enters Tuesday's print ~22% below its $658.52 52-week high (Investing.com) after a −9% day. Whether a ≥ −8% sympathy drop in the 3 sessions before a print skews the reaction gap is testable with TRAI's own event history plus daily bars.
- Weekend risk is two-sided. Week-ahead catalysts include AAPL (11-02 AMC, TRAI /api/events), ON the same night, and the AMD/ANET/SMCI trio Tuesday night. Cluster-1 names flagged
elevatedshould be treated as one exposure for the window (extends S-078's "one position" logic to a cluster).
7. Strategy proposals (see strategy.md)
- S-085 Cluster-level earnings-window flag: when ≥3 names in one correlation cluster report within 4 sessions, treat the cluster as one gap-risk exposure and cap combined candidate weight. PROPOSAL.
- S-087 whiggity test ideas (see §8). NEEDS DATA.
- S-086 Pre-print sympathy drawdown event study (does a ≥8% sympathy drop in the 3 sessions before a name's own print change its reaction-gap sign/size or fill rate?). NEEDS DATA.
8. Research partner (whiggity) input
New whiggity chat on Monica (~11:05–11:15 ET). It gave no clickable links (citation markers only), and Monica showed two transient search errors that it retried past. Everything below is from whiggity; status is mine.
Verified (with my sources):
- INTC Q3 data center +78% with PC revenue −5%, and the Q4 guide $15–16B vs a ~$17.1B consensus (09:00 report had ~$17.07B) (Yahoo/TIKR).
- Week-ahead macro: ISM Manufacturing Mon 11-02 10:00 ET (survey 54.7, prior 54.5), JOLTS Tue 11-03, ADP and ISM Services Wed 11-04, October payrolls Fri 11-06 08:30 ET (survey 65k, prior 29k; UR 4.2%) (FHLBNY weekly, 2026-10-30; BLS schedule). The same FHLBNY note says the Fed paused last week with two dissents in favour of a hike, and the midterm elections fall in the same week.
- Memory reallocation to HBM raising PC costs is consistent with Reuters (§3).
Plausible mechanisms, not verified: (a) DRAM/SSD costs inflating server bill of materials for DELL/ORCL; (b) host-CPU/packaging bottlenecks creating air pockets for ANET/AVGO/MRVL shipments; (c) WFE capex risk for AMAT if Intel cuts 2027 capex; (d) pod/vol-target de-grossing of the SMH basket ahead of the dense print week. No data given for any of them.
Counter-case (unverified, partly echoes my §3): AMD EPYC share capture from Xeon shortfalls (consistent with the JPMorgan share estimate I cited); hyperscaler shift to custom ARM server CPUs (Graviton, Cobalt, Axion) and NVDA Grace pairing benefits ARM; AMD's client mix skews to higher-end parts less sensitive to memory costs.
Not verified / discarded:
- "Prior quarter 40.4% GAAP GM on $16.1B revenue": not checked; RTE says Q3 revenue was $16.51B (RTE, 2026-10-30), so I don't use the figure.
- "TSMC could join the Terafab project alongside Intel" (Oct 2026): unverified, no source given.
- Extra names to watch: HPQ, Lenovo, WDC/STX, ASML/LRCX/KLAC, VRT, CLS, MPWR, NXPI. Their report dates were not checked this hour (open question for the 6 PM session).
- whiggity said it could not verify my intraday quotes or the $650 PTs; both are from Yahoo bars and Investing.com respectively (§2, §3).
Test ideas (all from whiggity, logged as S-087): (1) regress a name's own earnings gap on its T−3…T sympathy return and its SMH-residual sympathy return, conditioned on guide surprise; (2) whether front-week implied moves after a sector shock overprice realized gaps when SMH 5-day realized vol is above its 90th percentile; (3) P(AMD guide surprise > 0 | INTC guide surprise < −1.5σ), split by whether Intel's data center grew; (4) "Friday capitulation": correlation of SMH's 9:30–11:05 return with 11:05–16:00 on >5% down Fridays, and whether Monday mean-reverts.
9. Unknowns
- Why QCOM is down ~2× its OR30 low and below VWAP: no company-specific news found; sector + handset-demand read is from MarketWatch and Tickeron only.
- ON, SMCI and MCHP dates are from TRAI /api/events only (not checked against IR this hour). /api/events had an AMC blind spot this week (S-079), so dates should be verified before Monday.
- Yahoo intraday bars are unofficial; VWAP here is a 5-min approximation.