Daily Dispatch — 2026-05-15
Today's market snapshot (as of May 15, 2026)
| Ticker | Price | Notes |
|---|---|---|
| INTC | ~$115.93 | ATH $129.44 (May 11); ~225% YTD; BofA PT $96, Mizuho PT $124 |
| ASML | $1,584.51 | Sole High-NA supplier; Intel/Samsung/SK Hynix driving 2026 allocation |
| AMAT | ~$446.76 | Q2 $7.91B record (May 14 close); Citi PT $520; +4.5% post-earnings |
| TSM | $363.35 | +3.13% today; 4-year pricing power; Arizona 2nd fab 2H27 |
| KLAC | $1,869.19 | +6.01% today; 52% process-control share |
| LRCX | $299.15 | Etch + deposition; TSMC/Samsung/Micron/Intel all customers |
| MU | $776.01 | HBM supercycle; 52-week range $90–$819; only US-domiciled HBM |
| CCJ | $115.50 | Uranium; nuclear baseload thesis; -1.22% today |
| FCX | ~$66.93 | Copper deficit thesis; primary copper miner |
| SCCO | $188.50 | Copper miner; -1.74% today |
Key developments today
AMAT Q2 FY2026 — beat-and-raise confirms picks-and-shovels durability
Applied Materials reported Q2 results May 14: $7.91B revenue (+11% YoY), non-GAAP EPS $2.86 (beats $2.67 estimate), Q3 guided at $8.95B ±$500M (~13% sequential step-up). Gross margin 50.0%, highest in 25+ years. CEO Dickerson: AI demand "durable and multi-year," customer conversations extend into 2027 and 2028. TSMC joined EPIC Center as founding partner May 11 — AMAT now co-innovates with all major fabs simultaneously. Citi raised PT to $520 (from $420). AMAT DRAM exposure = 31% of CY2026 revenue, highest in semicap peer group. WFE trajectory: AMAT Q3 guide of $8.95B implies ~$35B annualized — already above EE Times' full-market $156B estimate, supporting Morgan Stanley's $185B+ model for 2027. Picks-and-shovels conviction raised to High for AMAT.
Apple-Intel deal — still preliminary, 7 days in
No public statement from Apple or Intel as of May 15 (7 days post-WSJ reporting). INTC pulled back ~10% from May 11 ATH ($129.44) to ~$110–116. BofA models the deal adding $10B/year Intel Foundry revenue by 2030 (PT $96, stock well above). Market is pricing in more than BofA's base case. Samsung is still exploratory — not confirmed as fallback. The deal remains the make-or-break signal for the INTC re-rate.
CSP capex confirmed durable — all 4 hyperscalers raised 2026 guidance
Comprehensive review of Q1 2026 hyperscaler earnings confirms the $725–830B capex cycle is supply-constrained, not demand-constrained. Sundar Pichai: "Cloud revenue would have been higher if we could meet demand." Google Cloud +63% YoY (suppressed). Azure AI doubled YoY for the 4th consecutive quarter. Alphabet enterprise backlog $462B (nearly doubled QoQ). 2027 consensus: Evercore + BofA both modeling >$1T combined hyperscaler capex; Goldman $1.15T cumulative 2025–2027. This confirms the upstream demand signal for all five picks-and-shovels chains. Bear case documented: 46% capex-revenue gap (Damodaran), 95% enterprise pilot fail rate (IBM survey) — real risk but doesn't negate supply-constrained framing.
SMIC $5.9B acquisition — China runs the same reshoring logic in reverse
SMIC acquiring 100% of SMIC North (Beijing 12-inch fab) for $5.9B — China's largest foundry M&A. SMIC North at 5nm-class ~30–40% yield (vs TSMC 80%+), accelerating domestic equipment adoption. This is the mirror image of the allied-reshoring thesis: China building semiconductor self-sufficiency to reduce dependency on Taiwan and allied nations. Two implications: (1) geopolitical bifurcation of supply chains accelerates; (2) removes a potential China-needs-Taiwan moderating force on escalation risk. Increases urgency of allied-nation reshoring ahead of China achieving self-sufficiency.
New hypothesis chains added
Three new thesis chains added to the wiki as hypothesis questions (not yet active — require revenue evidence):
- datacenter-construction-electrical-picks-shovels: Huang "30K truckloads per 500MW DC" → physical construction is a second picks-and-shovels layer. Tickers: VRT, PWR, FIX, HUBB. Watch: DC segment revenue >30% YoY.
- lattice-fpga-ai-cluster-management-layer: Lattice+AMI $1.65B → FPGA+BMC control plane for every AI rack, scales with cluster unit count. LSCC. Watch: post-AMI close, DC >50% with >30% growth.
- softbank-roze-robotics-datacenter-construction: SoftBank Roze + ABB → autonomous DC construction robotics; $100B IPO target H2 2026, pre-revenue. Low priority; revisit at S-1 filing.
Tesla Intel-14A Austin R&D fab — separate from Terafab
Tesla Q1 2026 earnings (April 22) disclosed a $3B standalone semiconductor R&D fab in Austin using Intel 14A, producing "a few thousand wafers per month" at pilot scale. Distinct from Terafab ($25B+ consortium). Two separate Musk-ecosystem Intel-14A commitments now documented: the Austin R&D fab (near-term, operational scale) and Terafab (production-scale, multi-year). Both reinforce 14A as the chosen node.
Thesis status update
| Thesis | Status | Conviction | Key risk |
|---|---|---|---|
| INTC re-rate via anchor customers | Active | Medium-high | Apple deal still preliminary; Samsung as third alternative; yield gap vs TSMC |
| Picks-and-shovels: ASML/AMAT/LRCX/KLAC | Active | High (AMAT raised today) | ASML already richly valued; China export drag on AMAT/LRCX |
| CSP capex durability ($830B+) | Active | High (confirmed today) | 46% capex-revenue gap; enterprise conversion risk |
| HBM supply bottleneck / MU | Active | High | SK Hynix dominant; MU execution on HBM3E ramp |
| Nuclear baseload / CCJ/CEG/BWXT | Active | Medium-high | CEG valuation rich; nuclear permitting slow |
| Copper supercycle / FCX/SCCO | Active | Medium | Mine output surprise could soften deficit |
| DC construction picks-and-shovels | Hypothesis | Low-medium | Pre-revenue signal; VRT already priced for AI |
| Lattice FPGA / AMI management layer | Hypothesis | Low-medium | AMI acquisition pending close (Q3 2026) |
Watch list for next session
- Intel Q2 2026 earnings (~July 2026) — key: IFS revenue guide, any Cook/Apple reference, 18A yield update
- Apple Q3 FY2026 earnings (~late July 2026) — Cook reaffirmation or pivot on Intel
- LSCC + AMI close (Q3 2026 expected) — trigger for lattice-fpga hypothesis graduation
- VRT / PWR data center segment revenue (next earnings) — trigger for DC construction hypothesis graduation
- Roze S-1 filing — low priority; revisit only when filed
- SMIC North CSRC approval — watch for China domestic fab progress updates
- Morgan Stanley / Street 2027 WFE estimate convergence — does $185B+ become consensus?