US Pre-Market Brief | 2026-10-01(ET) Thursday

Coverage window: 2026-09-30 (Wed) 16:00 ET regular-session close → 2026-10-01 (Thu) ~09:25 ET pre-market snapshot (drafted ~10:00 ET; post-open behavior not included); after-hours/pre-market quotes are timestamped. All times Eastern (ET). Today is the first trading day of Q4; last night's Micron after-hours earnings are the absolute core of this window.

0. Today's One-Line Summary

1. The strongest catalyst is "Micron's beat-and-raise landing": MU FQ4'26 revenue $54.23B (+379%, first time above $50B) far above consensus ~$51.2B, non-GAAP EPS $33.42 (vs $31.45), gross margin 87.0%; FQ1'27 guidance revenue $61.5B±1.5B (about 8% above consensus ~$56.8B), EPS $38.15±1.0; management said memory supply stays tight through 2028, and 2027 HBM is sold out with rising prices — but the stock isn't buying it: after-hours it spiked then faded to $1,056.75 (-0.8%), and pre-market at 9:25am it traded at $1,054.60 (-0.99%). A textbook "good-news exhaustion" setup; the only pre-market point of dispute is the "slight squishiness" in gross-margin guidance cited by Deutsche Bank (86.25%, a small sequential decline).
2. The AI chain is broadly higher pre-market, spreading from memory to the whole chain: NVDA +1.0% ($230.74, MU demand validation + $150B buyback floor), GOOGL +1.32.3% (overnight launch of the flagship Gemini 4 Argon model, beating OpenAI/Anthropic on most benchmarks, still rising pre-market), LRCX/AMAT each +1.8% pre-market (per Reuters 4:48am), SNPS +0.32.5% (FY27 EPS guidance $19.04–19.12, ~10% above consensus $17.32), ORCL +2% (report says Tencent is leasing 100K chips from its Southeast Asia data centers, single source, to be verified); in Asia, Nikkei +3.30% and KOSPI closed up nearly 2% (Samsung +2.79%, SK Hynix +3.21%) set the overnight tone.
3. Macro is the other pole today, and it's bearish: the 10Y Treasury yield early this morning rose further above yesterday's close (5.293%) to 5.32–5.33%, and the 30Y to 5.66% — both the highest since 2002; Q3 was the worst quarter for Treasuries since 1994; Dow futures at one point hit a three-month low (around -0.5%), Russell 2000 futures -0.58%, VIX 16.84 (a two-week high). The soft PCE knocked the October rate-hike probability from 71% a week ago to ~37%, but the divergence of "improved rate-cut expectations while the long end keeps selling off" continues — fiscal/issuance supply is the main narrative, with 8 Fed speakers today (Williams said Wednesday "can wait until December to hike"; see this series' 9/30 recap for the quote).
4. Single-stock events: CEG +2.5% pre-market (20-year nuclear PPA with Amazon: Calvert Cliffs totaling 690MW, including 190MW of new uprate, >$3B investment); VICR +12~15% pre-market (third raise of Q3 sequential growth guidance to >30%, on AI power VPD licensing revenue); NKE reports tonight after hours (consensus EPS $0.44, revenue $11.33B), and pre-market already touched a 52-week low of $35.16; geopolitics: Iran says it received the US reply on the Gulf ceasefire proposal, and the US military completed its withdrawal from Iraq, but WTI is still +2.2% at $92.39.
5. Pre-market snapshot (4:48–5:13am): Dow futures -0.52%, S&P -0.02%, Nasdaq 100 +0.29/+0.35%; 10Y 5.32%, 2Y 4.91%; DXY 101.75 (-0.29%); gold $4,158; BTC $83.6K; today's data: 8:30 jobless claims (consensus 209K, prior 197K, official actual not obtained at press time), 10:00 ISM manufacturing PMI (expected 55.0, prior 54.6) + August construction spending; tomorrow night (10/2) the September jobs report is the ultimate event of the week.

1. News Overview

⚠️ Data-fetch failure note (ops): this machine's yfinance has been rate-limited by Yahoo across the board for days (YFRateLimitError, fast_info fails on the first call); 0 fields in this report came from yfinance; all stock data went through stockanalysis.com (direct FetchURL) + MarketWatch + multi-source cross-checking; 52-week ranges verified against CNBC/Yahoo/Morningstar/IR. Unresolved disputes in this report: ① conflicting VICR 9/30 closing prices from two sources (Morningstar $276.06 vs stockanalysis $288.94); we adopt the latter, which is internally consistent with the pre-market quotes; ② VICR/CEG/NKE YTD gains are estimates; YTD has inter-source discrepancies (flagged item by item); ③ SNPS's stockanalysis quote page lags; replaced with four sources incl. CNBC/Robinhood/Morningstar; ④ this morning's 8:30 initial claims official actual not obtained (body marks "no reliable data"), 10:00 ISM not released at press time; ⑤ NKE consensus (EPS $0.44 / revenue $11.33B) not verified by a second source; ⑥ ORCL-Tencent lease news is a Stocktwits single source, unverified.

2. Strongest Themes, Descending

3. Overall Stock Strength Board

Bullish/watch zone (sorted by total score descending; pre-market prices are 8:20–9:25am ET snapshots):

Note: fundamentals/valuation/consensus are stockanalysis.com aggregate single-source basis unless otherwise noted (not verified by a second source); 52-week ranges cross-checked across multiple sources; ties broken by source authority and liquidity.

Bearish/avoid zone:

4. Stock Scoring Model (Total 100, Itemized Scores Required)

• MU (81): source quality 15 (company earnings/IR, highest tier) + catalyst directness 20 (own results + guidance) + earnings elasticity 15 (revenue +379%, GM 87%, FQ1 accelerating again sequentially) + moat 14 (HBM triopoly, 26 long-term agreements covering >35%, supply discipline through 2028) + expectation gap 6 (forward 6× cheap, but after +273% YTD with no pre-market gain = expectations full) + sustainability 10 (through 2028) + trading profile 10 (top liquidity/options depth) − risk 9 (good-news exhaustion, GM guidance dispute, IV crush, after-hours spike-and-fade).
• NVDA (76): 13+14+14+15+6+9+10−5 (10Y pressures valuation, ecosystem substitution narrative, only -2.5% from prior high so limited room).
• GOOGL (70): 13+15+12+14+7+9+10−10 (capex $132B squeezes FCF, limited initial access, internal skeptics, macro rate pressure).
• VICR (70): 15+20+15+8+7+8+7−10 (92× P/E, 28× P/S, small float, already +15% pre-market, guidance includes one-off licensing component to be verified).
• CEG (65): 15+17+10+12+6+10+9−14 (net debt $24B, rate-sensitive, 9/30 -3.99% already demonstrated, uprate 2030+).
• SNPS (63): 13+16+12+12+7+8+8−13 (net debt $7.23 billion, interest coverage 1.66×, +21% share count dilution, already front-ran +4.78% on 9/30).
• META (60): 8+10+10+14+5+8+10−5 (no new hard catalyst, at highs).
• ORCL (58): 8 (single source) +15+10+12+7+8+9−11 (unverified, rich valuation).
• RKLB (55): 12+14+8+7+6+6+7−5 (losses, speculative profile).
• LLY (54): 14+12+8+13+4+7+8−12 (data not a surprise, healthcare policy, priced in).
• NKE (36, bearish basis): 12+12+3+8+3+5+9−16 (52-week low, earnings eve, FCF covering dividends 111%, weak China, tariffs).

5. Top Stock Deep-Dives

5.1 MU 美光 — The Core of the Whole Session, "Blow-and-Raise but the Stock Won't Rally" (S | 81)

• Related news: FQ4'26 after hours on 9/30 (Micron IR¹): revenue $54.23B (+379%, first time above $50B) vs expected ~$51.2B; non-GAAP EPS $33.42 (GAAP $32.87) vs $31.45; GM 87.0%. FQ1'27 guidance: revenue $61.5B±1.5B (vs consensus ~$56.8B, about 8% higher), EPS $38.15±1.0, GM ~86.25%. Operating cash flow $43.97B, quarterly dividend $0.15; customer supply agreement commitments raised to $32B; 26 strategic customer agreements covering over 35%; 2027 HBM bit supply agreements fully signed with rising prices; management says memory/flash supply stays tight through 2028.
• Catalyst logic: FQ1 guidance accelerating +13% or more sequentially + long-term agreement lock-in = cyclicality weakened by "contractualization" — the core argument for the valuation logic migrating from "cyclical" to "quasi-growth." The answer to the pre-market -1% lies in two places: ① consensus too full (+273% YTD; any beat is treated as confirmation of linear extrapolation, not a surprise); ② the "squishiness" of the 86.25% GM guidance cited by DB — the gross-margin peak dispute is the bears' only anchor.
• Theme and stage: mid-stage of the AI memory supercycle, getting crowded; last night was the "midterm exam," grade A+, but the stock reaction shows pricing has already moved to "final exam."
• Fundamentals check (stockanalysis.com, 10/1 pre-market basis): TTM revenue $133.19B, net profit $84.97B, FCF $58.96B, net cash $68.27B (huge single-quarter jump, consistent with the massive FCF); TTM P/E 14.3×, forward P/E 6.0×; 49 covering analysts at Strong Buy, mean target $1,521 (+42.8%).
• Pre-market and technicals: 9/30 closed $1,065.11; after-hours high $1,083 faded to $1,056.75; pre-market 9:25am $1,054.60 (-0.99%), intraday range $1,062–1,084. 52-week range $179.61–1,255; -16% from the 52-week high. With implied volatility realized, IV crush lands and option-side profit difficulty rises.
• Final call: Priority deep-dive, but let the market vote first — verification points: the direction chosen between the $1,056 support and $1,100 resistance after the open; a volume-backed reclaim of $1,083 (the after-hours high) falsifies "good-news exhaustion."

5.2 NVDA 英伟达 (A+ | 76)

• Related news: Indirect beneficiary of MU earnings (memory is a coincident indicator for AI servers); the $150B buyback (total authorization $235B) keeps providing a floor; 9:25am pre-market $230.74 (+1.03%).
• Catalyst logic: MU's "tight supply through 2028" = extended shipment visibility for NVDA; the buyback changes the supply-demand structure on pullbacks. No new catalyst of its own — a "flagbearer" type of sympathy rally.
• Theme and stage: AI compute leader, consolidating at highs since 9/22, only -2.5% from the 52-week high of $236.54, at the decisive battleground of the prior high.
• Fundamentals check: TTM revenue $302.97B, net profit $192.88B, FCF $127.01B, net cash $23.61B; TTM P/E 28.9× / forward 19.0×; 61 Strong Buy, target $327.70 (+43.5%).
• Final call: Priority deep-dive; verification point: whether the $233–236 prior-high band can break out on volume — a breakout opens new-high space, failure keeps the $222–236 box.

5.3 GOOGL Alphabet (A | 70)

• Related news: Launched the flagship Gemini 4 Argon model after hours on 9/30 (The New Stack⁹): beats OpenAI/Anthropic on most benchmarks (strongest in knowledge work, mixed on coding); API pricing $2/$10 per million tokens (95% discount on cached input); 1M output cap; initial limited access only for trusted testers / cybersecurity defense agencies. JPMorgan keeps Overweight; internal skeptics also exist (Bloomberg report).
• Catalyst logic: model lead → Cloud/API share → stronger ads/subscription monetization; aggressive API pricing is "volume for price" to win the developer ecosystem — pressure on gross margin short term, ecosystem lock-in long term. This is the landmark event of "the AI narrative coming back to GOOGL from its rivals."
• Theme and stage: in the three-way AI model race GOOGL completes a key positioning move; the stock is up only +10% YTD, a clear laggard among the mega-cap tech.
• Fundamentals check: TTM revenue $445.87B, net profit $244.12B (single-source aggregate basis, not verified by a second source; TTM net profit includes large non-operating gains, so TTM P/E 17.3× is therefore distorted — forward P/E 25.6× is the real trading multiple); capex $132.4B leaves FCF at only $53.27B (P/FCF 79×) — capex is the core valuation controversy; net cash $121.68B; 61 Strong Buy, target $429.46 (+24.8%).
• Pre-market and technicals: 9/30 closed $344.08 (+0.93%), intraday $341–360; pre-market $348.50 (+1.28%, 8:20am) / $351.57 (Robinhood); 52-week range $235.84–408.61, -15% from the high.
• Final call: Watch closely; verification point: a volume breakout above the 9/30 intraday high of $360; if the rollout pace (speed of access expansion) beats expectations, the re-rating slope steepens.

5.4 VICR Vicor (A− | 70)

• Related news: 10/1 pre-market company announcement (Vicor IR¹⁴): Q3 sequential growth guidance raised from >20% to >30% on increased VPD licensing revenue — the third raise this year (May: raised Q2; 9/21: raised Q3 to >20%; today: raised again).
• Catalyst logic: business-model re-rating from "selling power modules" to "modules + IP licensing": Q2 already beat (EPS $1.04 vs $0.652), and Q3 accelerates further on top of the beat; Vertical Power Delivery patents + the US ITC import ban form a legal moat against competitors; backlog $380M.
• Theme and stage: the sharpest name in the AI data center power-delivery architecture upgrade theme; after a ~25% pullback from the 6/30 high of $382.65, this is the second-leg launch signal.
• Fundamentals check: TTM revenue $474 million (+22%), net profit $145 million, GM 56.63%, net cash $446 million; but TTM P/E 92.8×, P/S 28.1×, forward P/E 58× — pure growth pricing; only 4 covering analysts (Buy, target $386.25); FY26 consensus EPS $3.12, FY27 $6.35 (implying a doubling of earnings); next earnings 10/20.
• Pre-market and technicals: pre-market $333.00 (+12~15%, 9:21am snapshot; an earlier read had +12.2%; the two sources disagree on the 9/30 close ($276.06 vs $288.94), and the gain range holds under both bases); extremely volatile intraday; float of only ~25.69 million shares (insiders hold 44.2%) — the liquidity discount is a double-edged sword. 52-week $48.31–382.65.
• Final call: Watch closely but never chase the +15% gap-up; only consider if the pullback toward the $300 area (former platform) holds; the 10/20 earnings will verify the "quality" of the new guidance (product revenue vs licensing mix).

5.5 CEG Constellation Energy (A− | 65)

• Related news: 9/30 after-hours announcement (BusinessWire¹²): signed a 20-year PPA with Amazon — Calvert Cliffs totaling 690MW (including 190MW of new uprate, online 2030–2032), supporting >$3B of investment; plus a PJM 13-state retail power supply agreement; revenue certainty supports 20-year plant relicensing.
• Catalyst logic: the hyperscaler nuclear long-term contract queue gains another member (after Microsoft/Meta/Walmart); the "power is the hard bottleneck of AI" narrative keeps getting contract-validated; the 20-year contract transfers power-price risk to the buyer, and CEG gets certain cash flow.
• Theme and stage: leader of the AI power theme, but the stock is ~-27% YTD (rates killing the valuation); 22 covering analysts with zero Sell, target $347 vs current $260 — a textbook case of "fundamental narrative mispriced by the market."
• Fundamentals check: TTM revenue $31.27B, net profit $3.47B, OCF $4.21B; net debt $24.0B (D/E 0.76, leverage up after the Calpine deal, interest coverage 6.0×); forward P/E 20.5×.
• Final call: Watch closely; short term it's a tug-of-war between "rates vs contracts," and a 10Y pullback is the bottle-opener for the stock; talk trend reversal only above $270 (the September platform).

5.6 SNPS Synopsys (B+ | 63)

• Related news: long-term financial model released at the 9/28 investor day, still brewing on 10/1: FY27 (ending 2027/10) revenue guidance $11.1–11.2B, non-GAAP EPS $19.04–19.12, ~10% above consensus $17.32; near 15% CAGR through FY30; driven by custom silicon demand and the chip design collaboration with OpenAI.
• Catalyst logic: the AI custom chip wave → design complexity explosion → EDA usage and pricing both rise; the "OpenAI collaboration" turns AI labs directly into a customer channel.
• Fundamentals check: GM 82.9% (software profile unchanged); but net debt of $7.23 billion after the Ansys deal, interest coverage only 1.66×, share count +21% YoY (dilution); GAAP EPS depressed by deal amortization (TTM P/E 74× distorted, forward 24.7× is the reference); 25 Strong Buy, target $545.93 (+31.5%).
• Pre-market and technicals: 9/30 already +4.78% to close $434.94 (four sources agree); pre-market $436.01 (flat) – $445.6 (+2.5%), 52-week $362.55–539.48 (-19% from the high).
• Final call: Watch closely; the "guidance beat + deep YTD drawdown" combo has re-pricing room, but leverage and dilution cap the slope; watch the pullback into $425–430 (the breakout platform) for support.

5.7 META (B+ | 60)

• Related news: Sensor Tower: the Muse AI agent passed 5 million downloads in three weeks, early pace faster than competitors (via Stocktwits).
• Catalyst logic: validation on the AI application side — Meta converts open-source model advantage into a consumer entry point; for the core ads business it's a far-dated option, not a current-EPS driver.
• Fundamentals: previously validated repeatedly (revenue +28%, gross margin ~81% territory); no reliable pre-market data for now. 9/30 closed $725.18 (-1.84%), holding the $710 verification level.
• Final call: Watch only — no hard-catalyst day; wait for the quality of support on a pullback toward the $710 area.

5.8 ORCL (B | 58)

• Related news: Stocktwits single source: report says Tencent is leasing 100K chips from Oracle's Southeast Asia data centers for AI infrastructure; +2% pre-market.
• Catalyst logic: if true, it's a large order of "Chinese demand landing in US cloud via a third location," a real positive for OCI utilization; but not verified by a second source, and this "leased compute" model carries high geopolitical sensitivity (risk of regulatory reversal).
• Final call: Watch only — upgrade only after a Reuters/Bloomberg-grade second source lands.

5.9 NKE Nike — Earnings Tonight, Pre-Market Touches 52-Week Low (Avoid | 36)

• Related news: FQ1'27 after hours tonight (confirmed date 10/1 AMC, stockanalysis calendar); consensus EPS $0.44, revenue $11.33B (not verified by a second source).
• Catalyst logic: the Win Now turnaround enters the "verify management credibility" phase: China demand (last fiscal Q4 China revenue -20%), weak North America, tariffs pressuring gross margin (GM 43.25%, still below the historical 45%+); what the market wants is not a beat, but evidence that "the downcycle has a floor."
• Fundamentals check: TTM revenue $46.4B, net profit $3.11B; net debt of $2.02 billion (NKE swung from net cash to net debt — a qualitative change in state); TTM P/E 16.9× / forward P/E 21.3× (forward EPS ~$1.65, earnings still being revised down); dividend yield 4.6%, payout ratio 78%, FCF payout ratio 111% — dividend sustainability in question; 42 Hold, target $45.63 (+28.9%, huge dispersion); shorts 5.78% and up 24% this month; ~-43% YTD, 52-week $35.16–76.97, pre-market $35.16 already touches a 52-week low.
• Final call: Avoid — two-way high volatility on earnings eve, with trapped longs and oversold conditions coexisting; wait for post-close guidance: gross margin stabilizing + China decline narrowing = can upgrade to watch; another guidance cut = opens space below $30.

5.10 RKLB Rocket Lab (B | 55)

• Related news: signed a 20-launch Electron commercial contract with Synspective (largest in company history), +3% pre-market.
• Catalyst logic: commercialization validation of launch cadence; more certainty in Electron small-rocket cash flow; Neutron remains an option.
• Final call: Watch only — the order is positive but persistent losses make this a theme-sentiment position; Q3 delivery data is the next verification point.

6. Bearish/Avoid List

7. Intra-Theme Rankings

Theme 1: AI Memory Supercycle (Strength ★★★★★)

Theme 2: AI Models/Software (Strength ★★★★)

Theme 3: AI Power Bottleneck (Strength ★★★★)

8. Opening Verification Signals

• Pre-market: ① the direction of MU's cash-market re-pricing ($1,056 support / $1,083 after-hours high / $1,100 resistance) — the single most important market signal; ② the quality of the AI-chain gap: whether NVDA/GOOGL gap-and-hold (holding over 50% of the pre-market gain) or gap-fill; ③ VICR's volume in the first 15 minutes after the +15% gap-up — small-float names that open high rarely stay high.
• Intraday: ① can the Dow hold in the first 30 minutes after the open (futures already at a three-month low; if the cash index gaps down and keeps falling, the AI-only-rally structure comes under pressure); ② SMH vs XLU relative strength — "tech huddle resisting the drop" was the September pattern; if SMH follows MU up while the broad market doesn't, it confirms a structural market; ③ the immediate reaction to 10Y 5.32%: yields push higher → valuation pressure across the board; a pullback → long-duration rebound.
• Options sentiment: MU post-earnings IV crush lands; open-interest changes in the $1,050–1,100 calls are worth watching; VICR has no options depth to speak of (liquidity discount); NKE's IV is rising into tonight's earnings — do the IV math before buying options pre-earnings.
• Risks: ① mixing pre-market and post-open data (this report's data is a pre-9:25am snapshot); ② single-plant risk: if only chips rally while soft tech / the broad market doesn't, the "AI faith vs bond rout" tug-of-war tilts toward the bond rout; ③ the macro triple: ISM at 10:00 + 8 Fed speakers — any "December hike pulled forward" talk hits both valuations and sentiment; ④ wait-and-see money ahead of tomorrow night's jobs report may make today's second half thin.

9. Final Conclusions

① Today's 5 Most-Watched Stocks

② Today's 3 Strongest Themes: ⑴ AI memory supercycle (validated by MU earnings, sustainable through 2028, representatives: MU/NVDA); ⑵ AI power bottleneck (CEG-AMZN PPA + VICR's third guidance raise, representatives: CEG/VICR); ⑶ AI model race (Gemini 4 + SNPS guidance, representatives: GOOGL/SNPS).

③ Today's Avoid Directions: NKE (earnings eve, 52-week low); small caps / pure duration (10Y 5.32%); chasing VICR's pre-market +15% gap-up; single-source ORCL news not chased.

④ Final One-Line Call: Today is a head-on tug-of-war between "the hardest earnings in AI history vs the highest rates since 2002" — MU proved the industry trend, the 10Y decides the valuation ceiling; at the open watch two prices, $1,056 and $233 — whoever holds sets today's tone, and before tomorrow night's jobs report, position size matters more than direction.

⚠️ Risk warning: this list is only a pre-market information roundup for observation and does not constitute investment advice. US stocks carry high volatility and pre-market gap risk; post-earnings IV crush and guidance reversals exist; auto-generated content may have information timeliness gaps or factual errors — defer to company disclosures/SEC filings, and do not use this directly as a trading basis.