Stock-Market Research Dispatch — 2026-05-28

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

Top of mind

The MP Materials DOD deal fundamentally transforms the rare earth thesis. What was previously a "will China reinstate controls at November 10?" binary is now a de-risked financial structure: $400M preferred equity (DOD becomes largest shareholder), $150M loan, a 10-year $110/kg NdPr price floor, and 100% off-take guarantee for the 10X Facility magnets. MP's economics are now largely independent of whether China reinstates the October 2025 extraterritorial controls. The correct question has shifted from "will China escalate?" to "does MP's current valuation (~$64) reflect de-risked economics with a government backstop?" — a materially different, and more tractable, thesis.

Simultaneously, the helium thesis was upgraded to HIGH conviction. QatarEnergy's own technical assessment confirms a 3-5 year repair timeline, gated by global gas turbine shortages (2-4 year delivery queues), not geopolitics. A ceasefire does not end the supply disruption. Linde (LIN) and Air Products (APD) hold ~6 months of global helium demand in storage — a durable pricing moat that does not require further escalation to be valuable. JP Morgan upgraded LIN with explicit helium framing (PT $455→$525).

New wiki pages

Significant wiki updates

Datacenter construction/electrical graduates to active concept (datacenter-construction-electrical-picks-shovels): All three graduation criteria confirmed simultaneously:

1. VRT AI DC backlog +80.8% YoY; book-to-bill ~2.9x (record for any industrial this cycle)
2. Transformer lead times confirmed 5 years (vs 24-30 months pre-2020)
3. FIX AI/tech = 56% of quarterly revenue (2x the graduation threshold)

ETN's datacenter backlog now stands at 228GW — roughly 12 years of demand at 2025 build rates. This concept has four named tickers (VRT, ETN, FIX, PWR) with Q1 2026 results confirming every part of the chain. Conviction: medium-high.

Section 232 Phase 2 date clarification (critical): July 1 is NOT a Phase 2 implementation date. Per the White House Presidential Proclamation verbatim, July 1 is when "the Secretary shall provide the President with an update... so that the President may determine whether it is appropriate to modify the tariff." Phase 2 requires a separate Presidential determination after the report. This corrects a persistent financial-media mischaracterization carried in prior wiki entries. Calibration filed.

$1.743T US manufacturing commitment tracker (us-industrial-policy-tariff-shield): As of May 26, 2026: semis/advanced tech $1.217T, 140 companies, 35 states. Top investors: Apple $600B, Micron $200B, IBM $150B, TSMC $100B, TI $60B. TSMC Arizona 3nm equipment installation beginning ahead of schedule (production 2027). Micron $100B NY fab groundbreaking — largest single-facility US private manufacturing investment ever; locks in $35B CHIPS 48D ITC at 35%.

2027 CSP capex >$1T now confirmed by four independent houses (csp-capex-cycle-peak-or-sustained): Moody's ~$1T, CNBC/Morgan Stanley (Alphabet alone $250B), Evercore, BofA — all post-Q1 2026. Azure's $80B unfulfilled backlog re-confirmed with specific CFO attribution: Amy Hood verbatim ("We are, and have been, short now for many quarters... Demand is increasing"); Satya Nadella confirmed GPUs sitting idle for lack of electricity. The cycle is supply-constrained, not demand-constrained.

Watchlist

Undervalued candidates (today)

• china-rare-earth-november-2026-deadline — MP (~$64). DOD deal ($400M equity, $150M loan, $110/kg floor, 100% off-take) de-risks the thesis from binary China-catalyst play to government-backstopped producer. Primary question: does the $64 price already reflect the DOD partnership, or is this still priced as an unhedged binary? DOD deal was recent — likely not fully priced.
• datacenter-construction-electrical-picks-shovels — FIX and PWR are the least-covered of the four named tickers. VRT and ETN have significant AI-DC analyst coverage; FIX (MEP contractor) and PWR (electrical GC) are discussed primarily as construction companies, not AI picks-and-shovels. The 56% AI/tech revenue at FIX and $48.5B backlog at PWR are not widely embedded in consensus estimates for these names.

New active concept (graduated today)

• datacenter-construction-electrical-picks-shovels — Conviction medium-high. Tickers: VRT, ETN, FIX, PWR. Jensen Huang's "30,000 truckloads for a single 500MW DC" is the chain seed; all three graduation criteria confirmed in Q1 2026 earnings. The physical construction / electrical layer captures value regardless of which GPU vendor wins or which foundry wins — pure picks-and-shovels.

Calibration logged

1. Helium recovery timeline — Prior: 6-12 months post-ceasefire. Updated: 3-5 years, gated by global gas turbine shortage (not politics). Pattern: underestimating-physical-supply-chain-lead-times. Conviction upgraded to HIGH.
2. Section 232 "July 1 gate" — Prior: implementation trigger. Updated: Commerce report submission date; Phase 2 requires separate Presidential determination. Pattern: overconfidence/recent-news (financial media framing, not primary document).

Open questions for human review

1. MP Materials valuation question: At ~$64 with a government backstop (DOD equity, $110/kg floor, 100% off-take), the risk profile has fundamentally changed. The November 10 binary catalyst still exists but is less load-bearing. Worth pulling a valuation snapshot: what does $64 imply at $110/kg NdPr for 5 years? This is now a de-risked asset, not a lottery ticket.
2. LIN and APD: helium conviction now HIGH. Both should have full valuation snapshots pulled before the next contract repricing cycle. The moat (Linde ~6 months global demand in storage) is real and durable for 3-5 years per QatarEnergy's own timeline. This is no longer speculative.
3. FIX and PWR: add to active watchlist? The datacenter construction concept just graduated. These two tickers have less sell-side AI coverage than VRT/ETN. If AI/tech is already 56% of FIX revenue and PWR's $48.5B backlog is AI-infrastructure-weighted, both could be mispriced vs. the new thesis.
4. GLW valuation: rich but strategic? Forward P/E 50-56x vs hardware median 24x, plus $51.5M net insider selling. The thesis is real (Nvidia anchor, hyperscaler megadeals), but the current price may already embed full execution. Corning Q2 2026 earnings (expected June-July 2026) should be the graduation gate — if Corning breaks out AI-specific optical revenue, this becomes actionable.
5. Section 232 post-July 1 positioning: The July 1 report triggers a Presidential determination process, not an automatic tariff change. The decision window after July 1 could be days, weeks, or months. Intel IFS (not TSMC, given the Taiwan zero-tariff-within-quota) is the cleaner Phase 2 beneficiary — worth holding as a separately-tracked thesis.
6. Podcast feeds (12th consecutive 403): Persistent network policy block. Container configuration issue.

What I looked at

• Step 1 topics: DC electrical picks-and-shovels (VRT/ETN/FIX/PWR Q1 2026); US industrial policy tariffs May 28 update; China REE controls + MP Materials DOD deal; CSP capex May 28 update; helium supply crisis Ras Laffan update; Corning GLW AI thesis deep-dive
• Step 2 macro buckets: Energy & critical minerals (helium, rare earths); US industrial policy & tariffs (Section 232, CHIPS 48D, $1.743T tracker)
• Podcast episodes transcribed: 0 (all RSS feeds HTTP 403; 12th consecutive failure)
• Sources promoted: 7 (2026-05-28 autoresearch batch); sources ingested: 7
• Valuation snapshots refreshed: 0 (WebFetch 403 — 12th consecutive)
• New entities created: 4 (glw/GLW, cheniere/LNG, fix/FIX, pwr/PWR)
• Concepts graduated: 1 (datacenter-construction-electrical-picks-shovels: hypothesis → active concept, medium-high conviction)
• Conviction upgrades: 1 (helium-supply-crisis-semicap: medium-high → HIGH)
• Calibration logged: 2 (helium recovery timeline; Section 232 July 1 date)