brain/
← all contexts
project

stock-market

Stock-Market Research Dispatch — 2026-06-02

Auto-generated by daily research routine. Not a source; not ingest-pending material. Summary of what changed in the wiki today.

Top of mind

SpaceX roadshow opens June 4 — the RSP/Mag7 trade arms on June 11 pricing. Roadshow runs June 4-10, pricing June 11, listing June 12 as SPCX on Nasdaq. The spacex-ipo-passive-shortfall-to-equal-weight-rerate armed position triggers at pricing. Conservative passive buying consensus is $15-30B (not the $60B Goldman figure, which applies only to QQQ under aggressive float-multiplier assumptions). The hedge-fund warehousing mechanic (~$16.5B coverage gap on day one) forces Mag7 selling into thin top-of-book liquidity. Long RSP / short Mag7 sized for a 9-12 month seasoning window. S&P 500 rule change (S&P Index Committee) still unconfirmed as of June 2.

ARM Holdings thesis graduated today — and ARM +15.73% to $408.85 on June 1. Q4 FY2026 earnings (May 6) confirmed: Neoverse datacenter royalties "more than doubled year-on-year." CEO Rene Haas is guiding another doubling in FY2027. The arm-holdings-vera-cpu-datacenter-royalty graduation gate was set for Q2 FY2027 (August 2026) — it was met 3+ months early, at Q4 FY2026. ARM surged +15.73% on June 1 to $408.85, likely triggered by Jensen Huang's "Vera CPU = $200B CPU market" call at GTC Taipei. ARM is now above all analyst consensus targets ($241 average) but Mizuho raised their target to $425. Thesis has been active-graduated; the near-term entry is no longer obvious at 188× fwd P/E.

Jensen Huang's Q1 FY2027 primary source contradicts the CUDA moat erosion thesis. Huang on the May 20 earnings call: "I expect that LPX and other SRAM-based decode-focused... accelerators will always be, will be a niche product for some time." cuda-moat-erosion-to-nvda-rerate conviction downgraded medium → low-medium. The nvda-cuda-moat-erosion short position is flagged for exit review — NVDA +6.26% on June 1 on the Q1 beat ($82B, +85% YoY) and GTC Taipei keynote. Source credibility is asymmetric: Feldman (Cerebras CEO) made the moat-erosion claims; Huang (Nvidia CEO) on an earnings call directly called the competing architecture niche. NVDA fwd P/E 22.58× on $5.43T market cap; analyst avg $296.81 (+32% upside from $224).

November 10 dual cliff confirmed: tariff truce AND REE suspension expire together. The October 30, 2025 Busan summit set both the US-China 90-day tariff truce AND the Wave 2 rare earth suspension with the same clock — both expire November 10, 2026. This is not two separate events; they're jointly conditioned. Yttrium collapsed from 333 MT/month to 17 MT/month (-95%), recovering to only ~20 MT (-70% below Jan 2025 baseline). MOFCOM's latest statement omits rare earths entirely — China is treating REEs as a separate bilateral negotiation. DOD January 1, 2027 non-China sourcing mandate is on the books but MP Materials and USA Rare Earth both track 2028 scale timelines; feasibility of full compliance by 2027 is in question.

Strongest-conviction buys

From step 4c ranking. Research output, not advice.

#TickerMechanismConvictionValuation vs baseFundamentals (one line)Next catalyst
1MUhelium-cliff-to-hbm-supply-crunch + S232 + hbm-cowos-as-binding-bottleneckmedium-highYes — fwd P/E 10.99× near 52w high; earnings June 24 = re-pricing eventAir Liquide 30-yr/$250M Idaho partnership (structural helium advantage); HBM4 50% pricing premium; Q3 consensus $34.27B/$19.55 EPS; Mizuho $1,150 PTQ3 FY2026 earnings June 24 — the thesis re-pricing event
2CEGpjm-capacity-prices-to-nuclear-premium + nuclear-baseload-for-ai-data-centersmedium-highYes — $265.70 vs analyst avg $365.73 (+37% upside); -9.5% from entryQ1 primary source: FCF $8.4B/yr → $11.5-13B/yr (2028-29, +45%); hyperscaler capex +75% YoY; PJM June FERC submissionCalpine lockup June 30 (25M shares) → near-term dip → entry window per ceg-calpine-lockup-entry-window
3CCJai-capex-to-power-and-materials-cascade step 5ahighYes — $112.59 vs analyst avg $130.25 (+15.7% upside)Q1 $845M/$509M EBITDA; India $2.6B LT deal; uranium spot $85.20/lb; LT $90/lb (14-yr high)Meta 7.8 GW nuclear PPA volume timeline; uranium spot vs $90 threshold

Watchlist

Active theses with current valuation context — stockanalysis.com, June 1 2026 closes.

TickerThesisConvictionPriceUndervalued vs base?Next catalyst
MUhbm-supply-bottleneck + helium + S232medium-high$1,035.50Yes at 10.99× fwd P/EJune 24 earnings
CEGnuclear-baseload-for-ai-data-centersmedium-high$265.70Yes — analyst avg $365.73Calpine lockup June 30
CCJnuclear-baseload-for-ai-data-centershigh$112.59Yes (+15.7% to avg)Uranium LT pricing
TSMtsmc-capacity-shortfall-and-pricing-powerhigh$435.63No (near 52w high $449)Q2 2026 earnings (July)
ASMLpicks-and-shovels-leading-edge-fab-buildouthigh$1,628.57No (near 52w high $1,654)48D H2 fab groundbreaking demand
NVDAai-capex-to-power-and-materials-cascadehigh$224.36Marginal (+32% to analyst avg)Q2 earnings (August)
SCCOcopper-supercycle-ai-data-centersmedium-high$194.62MarginalFCX Q2 earnings; LME trend
INTCus-fab-capacity-bottleneckhigh$109.33No (flagged for exit; -9.4% from entry, -17.7% sector-excess vs SMH)VLSI June 14–18; 18A-P peer-reviewed data
ARMarm-holdings-vera-cpu-datacenter-royaltymedium-high$408.85No (+69% above analyst avg $241; Mizuho $425 PT ≈ current)FY2027 H1 royalty data; Vera shipments

Undervalued candidates (today)

  • MU (hbm-supply-bottleneck + helium-cliff-to-hbm-supply-crunch + S232): At $1,035.50, fwd P/E 10.99×. The multiple is remarkably cheap for a company with dual forcing functions converging in Q3 and June 24 earnings as the re-pricing gate. Air Liquide 30-yr partnership gives structural helium advantage vs all Korean peers. Analyst consensus $717 is stale vs recent price surge; Mizuho $1,150 is the active bull target. Strongest buy.

  • CEG (nuclear-baseload-for-ai-data-centers): $265.70 vs analyst avg $365.73 (+37.7%). Q1 primary source confirms FCF trajectory ($8.4B → $11.5-13B, +45%). The -7.66% June 1 drop and Calpine lockup June 30 create a near-term technical overhang (hypothesis ceg-calpine-lockup-entry-window targets entry at $270-285 dip around lockup). PJM June FERC submission is the pending positive catalyst. Not a buy at $265 today — wait for the post-lockup dip.

New chains to investigate (hypothesis-stage)

Updated theses (existing active)

  • cuda-moat-erosion-to-nvda-rerate (conviction medium → low-medium): Jensen Huang's "niche" primary source. Short position nvda-cuda-moat-erosion flagged for exit review — NVDA is delivering on earnings while the thesis source (Feldman) is a competitor.

  • arm-holdings-vera-cpu-datacenter-royalty (status: hypothesis → active): Graduation gate met Q4 FY2026. ARM is now a full entity in the wiki. New entry is not obvious at 188× fwd P/E / 69% above analyst consensus.

  • spacex-ipo-passive-shortfall-to-equal-weight-rerate (armed → active on June 11 pricing): RSP/Mag7 trade triggers at SPCX pricing. S&P rule change still unconfirmed. Conservative passive $15-30B remains the planning figure.

  • tsmc-saturation-to-intel-anchor-stack (step 4 updated): Google 3-5yr LTA now primary-source confirmed via CEO Lip-Bu Tan Q1 2026 earnings call. Google joins AWS, MSFT, Apple-preliminary, Terafab as named anchors.

  • helium-supply-crisis-semicap (HIGH conviction, maintained): Samsung chairman "until 2030" disruption estimate; Micron Air Liquide 30-yr partnership; TSMC CEO CC Wei primary-source corroboration on Middle East gas/chemical risk.

  • china-rare-earth-november-2026-deadline (dual cliff now confirmed): November 10 = tariff truce + REE suspension, jointly conditioned from Busan summit. Urgency elevated.

Primary-source upgrades today

Nine earnings transcripts and autoresearch syntheses ingested. Key primary-source additions:

  • CC Wei (TSMC CEO) — Q1 2026: N2 "good yield," N3 crossover H2 2026, tightness through 2027, Middle East gas/chemical risk. Three quotes wired to tsmc, helium-supply-crisis-semicap, tsmc-saturation-to-intel-anchor-stack.
  • Lip-Bu Tan (Intel CEO) — Q1 2026: Google 3-5yr LTA (step 4 of tsmc-saturation-to-intel-anchor-stack primary-sourced), packaging "billions of dollars per year."
  • Jensen Huang (Nvidia CEO) — Q1 FY2027: $82B +85% YoY, VeraRubin Q3 "full production," LPX "niche" (conviction downgrade trigger).
  • Joseph Dominguez (CEG CEO) — Q1 2026: hyperscaler capex +75% YoY, PJM June FERC submission.

Portfolio health (June 1 closes)

PositionEntryMarkPos%Sector-excess
MU (long)$766.58$1,035.50+35.1%+26.8% ← best in portfolio
NOW (long)$124.37$135.86+9.2%+3.3%
TSM (long)$397.28$435.63+9.7%+1.4%
LRCX (long)$289.24$317.12+9.6%+1.3%
SCCO (long)$189.85$194.62+2.5%+3.3%
CCJ (long)$116.93$112.59−3.7%+3.3% (beat URA)
INTC (long)$120.61$109.33−9.4%−17.7% ← flagged for exit
GEV (long)$1,071.98$950.54−11.3%−10.4% ← worst sector-excess
CEG (long)$293.60$265.70−9.5%−4.9% (watch Calpine lockup)
NVDA (short)$219.51$224.36−2.2% (short)NVDA vs SMH −4.8pp inception, but conviction lowered ⚠

Unpriced: Samsung 005930.KS, SK Hynix 000660.KS, SMNEY, CBRS (entry ambiguous). These four positions have flagged ‹fill› entry_refs; not scored.

Exit discipline notes:

  • INTC: Now -9.4% position, -17.7% sector-excess. Every prior exit-discipline flag has been confirmed. VLSI June 14–18 is the remaining catalyst; if 18A-P peer review disappoints, no remaining thesis to stay in.
  • NVDA short: Mechanism conviction lowered. Jensen Huang's "niche" characterization is a primary-source contradiction of the thesis. Thesis source (Feldman) is a direct competitor. The short P&L is negative (-2.2%). Review for exit.
  • GEV: High-conviction thesis, worst sector-excess. FERC ruling end-of-June is the catalyst. Hold through FERC; exit if ruling disappoints.