Brent Crude
$107.85
WTI Crude
$102.45
Brent-WTI Spread
$5.40
Dubai Physical
$126.00
⚡ KEY SIGNAL: Dubai physical crude ($126) is trading $18+ above Brent futures ($108). Trump's jawboning suppresses paper prices, but physical supply constraints can't be talked away. Watch the physical-paper gap — when it collapses, futures reprice violently upward.
Risk Indicators
Hormuz Status
Closed
Yuan toll for CN/RU only
Brent War Premium
~$15-18/bbl
Goldman estimate
Paper vs Physical Gap
$18+
Brent $108 vs Dubai $126
Gulf Drone Defense
Improving
UKR interceptors deploying
Iran Internal Unrest
Elevated
Isfahan/Ahvaz strikes
U.S. Energy Insulation
Strong
13.6M bbl/day domestic
SPR Available
Yes
Strategic reserve buffer
Red Sea / Houthis
Active Threat
Salalah hit, Maersk halted
APRIL 6 — TRUMP ULTIMATUM
Iran must reopen Strait of Hormuz or face attacks on energy infrastructure & Kharg Island
30-40% Probability
Hormuz Reopens by Mid-April
- Brent Target
- $80-90 by Q3
- S&P 500 Impact
- Recovery rally, +5-8%
- Sector Rotation
- Energy down, consumer up
- Trigger
- Skinny deal / regime collapse / Kharg seizure
35-40% Probability
Partial Reopening / Toll Regime
- Brent Target
- $95-110 sustained
- S&P 500 Impact
- Sideways, ±3%
- Sector Rotation
- Energy flat, defensives lead
- Trigger
- Yuan toll system formalized, limited tanker flow
20-30% Probability
Escalation / Infrastructure Hits
- Brent Target
- $130-150+
- S&P 500 Impact
- -10-15% correction
- Sector Rotation
- Energy spikes, broad selloff
- Trigger
- Kharg strike / Saudi facility hit / full blockade
Feb 28
U.S.-Israel strikes begin
Mar 2
Strait of Hormuz effectively closed
Mar 4
Brent surges past $82
Mar 9
Brent hits $115; Saudi/Bahrain infrastructure hit
Mar 11
Iranian strikes hit AWS facilities in UAE/Bahrain — banking outages
Mar 15
Fujairah port attacked
Mar 18
Ukraine deploys 201 anti-drone experts to Gulf
Mar 20
Isfahan steel worker strikes begin
Mar 24
Houthi drone hits Salalah (Oman)
Mar 27
Ukraine-Saudi defense deal signed
Mar 28
Ukraine-UAE & Qatar defense deals
Mar 30
WTI settles above $100 for first time since 2022
Apr 6
TRUMP ULTIMATUM DEADLINE
Apr 19
BCA supply-doubling inflection point
Conflict / Macro Watchlist
| Ticker | Name | Thesis |
|---|---|---|
| XLE | Energy Select SPDR | Direct oil price exposure |
| XOP | Oil & Gas E&P | Leveraged to WTI upside |
| ITA | iShares Defense ETF | Defense spending tailwind |
| DBA | Invesco Agriculture | Fertilizer/food chain disruption |
| UNG | U.S. Natural Gas Fund | LNG rerouting premium |
| EWG | iShares Germany | Short — EU energy vulnerability |
| INDA | iShares India | Short — Hormuz-dependent crude imports |
| TLT | 20+ Year Treasury | Rates risk if inflation spikes |
⚠ This is a monitoring framework, not investment advice. Always verify current prices and consult a financial advisor before making trading decisions.
⚡ THESIS: Hyperscalers are on track to spend $600-700B on AI infrastructure in 2026. The companies that supply the physical layer — power, cooling, connectivity, construction, compute — are the picks-and-shovels of this cycle. Ranked below by base-case 3-year total return potential.
Key Assumptions
Hyperscaler Capex
$600B+/yr
Through 2028
Grid Bottleneck
Persists
Behind-the-meter favored
Liquid Cooling
Standard by 2027
100K+ GPU clusters
Construction Labor
Scarce
= pricing power
Tier 1 — 120%+ Total Return Potential
| # | Ticker | Company | Category | Thesis |
|---|---|---|---|---|
| 1 | BE | Bloom Energy | Power Gen | Behind-the-meter fuel cells. GE turbines sold out through 2028. Only near-term on-site DC power alternative. |
| 2 | MRVL | Marvell Tech | Custom Silicon | Custom AI accelerators, DPUs, electro-optics. Design wins with every major hyperscaler. Revenue inflecting. |
| 3 | MOD | Modine Mfg | Cooling | ~$10B cap, pure-play liquid-to-chip cooling. Margin expansion as DC mix dominates. |
| 4 | AMKR | Amkor Tech | Packaging | CoWoS advanced packaging is THE GPU production bottleneck. Capacity sold out. |
| 5 | MU | Micron | HBM Memory | HBM is the second bottleneck after packaging. Super-cycle pricing, demand visibility through 2028. |
Tier 2 — 80-120% Total Return Potential
| # | Ticker | Company | Category | Thesis |
|---|---|---|---|---|
| 6 | CLS | Celestica | Contract Mfg | Design-to-build lock-in with hyperscalers. Smaller cap, backlog surging. |
| 7 | NVT | nVent Electric | Liquid Cooling | Pivoting entire company toward DC infrastructure. Smaller base = bigger moves. |
| 8 | VRT | Vertiv | DC Power/Cooling | Record backlog, duopoly with Schneider. Already run up but earnings sustain it. |
| 9 | FIX | Comfort Systems | DC Construction | Mechanical contractor. Skilled labor scarcity = pricing power. Peak construction 2026-2028. |
| 10 | EME | EMCOR Group | DC Construction | Electrical/mechanical DC build. Record bookings converting to revenue now. |
Tier 3 — 50-80% Total Return Potential
| # | Ticker | Company | Category | Thesis |
|---|---|---|---|---|
| 11 | PWR | Quanta Services | Grid + DC Build | $60B+ backlog, decades of grid work ahead. Size caps upside slightly. |
| 12 | GEV | GE Vernova | Turbines/Grid | Gas turbines sold out through 2028. Extreme visibility but P/E >50 limits return. |
| 13 | NVDA | Nvidia | GPUs | Best moat in tech. ~$2.5T cap means law of large numbers constrains % returns. |
| 14 | AVGO | Broadcom | Custom AI/Net | Google TPU supplier, networking ASICs. Fortress moat. ~$900B cap. |
| 15 | CIEN | Ciena | Optical Net | DC interconnect demand rising. Cheaper valuation = room to re-rate. |
| 16 | ANET | Arista Networks | DC Switching | Best-in-class 800G. Priced for perfection. Steady compounder. |
Tier 4 — 30-50% Total Return Potential
| # | Ticker | Company | Category | Thesis |
|---|---|---|---|---|
| 17 | TSM | TSMC | Chip Fab | Irreplaceable. ~$800B cap, Taiwan geopolitical risk is permanent discount. |
| 18 | GLW | Corning | Fiber Optic | Monopoly position, optical segment repricing higher. Moderate growth rate. |
| 19 | ETN | Eaton | Power Distro | Fortress moat, $90B+ cap. Reliable compounder, not a double. |
| 20 | APH | Amphenol | Connectors | In every rack. Well-owned, $90B+ cap. Limited upside from here. |
| 21 | HXSCL | SK Hynix | HBM Memory | HBM co-leader with Micron. OTC liquidity poor; Korean listing preferred. |
| 22 | HUBB | Hubbell | Grid Electrical | Solid grid play, less DC-specific torque than peers. |
Tier 5 — 15-30% Total Return Potential
| # | Ticker | Company | Category | Thesis |
|---|---|---|---|---|
| 23 | SBGSY | Schneider Electric | Power/UPS | Excellent moat, $130B+ cap, diversified. Steady, not explosive. |
| 24 | TEL | TE Connectivity | Connectors | Diversified exposure dilutes the DC upside. |
| 25 | TT | Trane Technologies | HVAC | DC is growing but still minority of revenue. |
| 26 | ABBNY | ABB | Transformers | Most diversified = least DC leverage. |
| 27 | CAT | Caterpillar | Generators | Every DC has backup gensets. Rounding error in CAT's revenue. |
Risk By Tier
Tier 1 Downside
-40 to -60%
If capex pullback
Tier 2-3 Downside
-20 to -35%
Valuation compression
Tier 4-5 Downside
-10 to -20%
Defensive, bond proxies
Kill Switch
Capex Pullback
Hyperscalers flinch
⚠ Rankings reflect Gabriel's base-case estimates as of March 31, 2026. Not investment advice. REITs excluded per request.
■ FIRST IN HISTORY: Data centers are being targeted as military objectives. Iranian retaliatory strikes hit AWS facilities in UAE and Bahrain, causing banking and enterprise outages. Submarine cable chokepoints in the Strait of Hormuz are now a live vulnerability.
$300B
Gulf AI Spend at Risk
165
Missiles Intercepted (UAE, 1 weekend)
541
Drones Intercepted (same weekend)
5GW
Stargate UAE Planned Capacity
Major Deals in the Crosshairs
| Deal | Partners | Value | Status |
|---|---|---|---|
| Stargate UAE (Abu Dhabi) | OpenAI, G42, Oracle, Nvidia, SoftBank | 5GW campus | 200MW cluster for 2026 — uncertain |
| AWS Saudi Region | Amazon Web Services + Humain | $5.3B | Mandatory security reassessment |
| Nvidia → Riyadh | Nvidia + Humain (Saudi PIF) | 18,000 Blackwell GPUs | Chip export policy in flux |
| AMD → Humain | AMD + Humain | $10B partnership | Under review |
| UAE Multi-Billion DC Campus | Undisclosed hyperscaler | Multi-$B | "Far from finalized" — Reuters |
Base Case Assessment
- Existing builds continue. Sunk cost + sovereign prestige. UAE and Saudi won't abandon what's in the ground.
- New greenfield commitments slow 20-40%. Insurance costs spike, hyperscalers quietly delay FIDs on unstarted projects.
- Net global DC capex impact: modest. Gulf is ~5-8% of total pipeline. Some spend redirects to Texas, Nordics, Southeast Asia.
- If hostilities end within 6 months, investment resumes at ~90% of prior pace. Strategic logic (cheap energy, sovereign capital) unchanged — just higher risk premium.
Spillover: Where Gulf Capex Redirects
U.S. (Texas, Virginia)
Biggest Winner
Already 60%+ of pipeline
Nordics (Sweden, Finland)
Growing
Cheap hydro, cold air
Southeast Asia
Emerging
Singapore, Malaysia, Indonesia
Impact on DC Infra Stocks
Neutral to Positive
Rerouting favors US contractors
▲ NET ASSESSMENT FOR STOCK LIST: Gulf disruption is mostly a non-event for infrastructure suppliers. They sell globally. Rerouting builds to US/European soil is neutral-to-positive for Quanta (PWR), EMCOR (EME), and Comfort Systems (FIX). The real risk remains whether hyperscalers collectively flinch on total capex — and so far, they haven't.
■ METHODOLOGY: Simulated portfolio inception on Feb 27, 2026 (last pre-war close). Tracks all 23 DC infrastructure + compute stocks vs S&P 500 (SPY) benchmark through March 31. This is a hindsight exercise — see caveats below.
-3.8%
Equal-Weight Portfolio
-3.7%
Tier-Weighted Portfolio
-5.0%
S&P 500 (SPY)
+1.3%
Alpha vs Benchmark
Tier Performance (Feb 27 → Mar 31)
Tier 2 (80-120%)
-0.8%
Best tier — beat SPY by 4.2%
Tier 3 (50-80%)
-1.0%
Beat SPY by 4.0%
Tier 1 (120%+)
-4.3%
Beat SPY by 0.7%
Tier 5 (15-30%)
-7.2%
Trailed SPY by 2.2%
Tier 4 (30-50%)
-8.5%
Trailed SPY by 3.5%
S&P 500
-5.0%
Benchmark
Individual Stock Returns (Ranked)
| Tier | Ticker | Company | Category | Feb 27 | Mar 31 | Return | vs SPY |
|---|---|---|---|---|---|---|---|
| T1 | MRVL | Marvell Tech | Custom Silicon | $81.69 | $98.71 | +20.8% | +25.8% |
| T3 | CIEN | Ciena | Optical Net | $348.70 | $382.92 | +9.8% | +14.8% |
| T2 | EME | EMCOR Group | DC Construction | $724.62 | $737.30 | +1.7% | +6.7% |
| T2 | CLS | Celestica | Contract Mfg | $277.63 | $279.36 | +0.6% | +5.6% |
| T2 | NVT | nVent Electric | Liquid Cooling | $118.36 | $117.89 | -0.4% | +4.6% |
| T3 | GEV | GE Vernova | Turbines/Grid | $873.07 | $869.00 | -0.5% | +4.5% |
| T3 | NVDA | Nvidia | GPUs | $177.18 | $174.01 | -1.8% | +3.2% |
| T2 | VRT | Vertiv | DC Power/Cooling | $254.83 | $249.55 | -2.1% | +2.9% |
| T3 | PWR | Quanta Services | Grid + DC Build | $563.08 | $548.65 | -2.6% | +2.4% |
| T3 | AVGO | Broadcom | Custom AI/Net | $318.88 | $309.39 | -3.0% | +2.0% |
| T4 | HUBB | Hubbell | Grid Electrical | $511.63 | $491.43 | -3.9% | +1.1% |
| T2 | FIX | Comfort Systems | DC Construction | $1428.63 | $1371.22 | -4.0% | +1.0% |
| T1 | MOD | Modine Mfg | Cooling | $227.25 | $216.63 | -4.7% | +0.3% |
| T4 | ETN | Eaton | Power Distro | $374.75 | $357.20 | -4.7% | +0.3% |
| T5 | CAT | Caterpillar | Generators | $742.83 | $707.76 | -4.7% | +0.3% |
| SPY | S&P 500 | Benchmark | $684.12 | $649.67 | -5.0% | — | |
| T1 | AMKR | Amkor Tech | Packaging | $47.73 | $45.13 | -5.4% | -0.4% |
| T3 | ANET | Arista Networks | DC Switching | $133.50 | $122.77 | -8.0% | -3.0% |
| T5 | TT | Trane Tech | HVAC | $461.21 | $417.11 | -9.6% | -4.6% |
| T4 | TSM | TSMC | Chip Fab | $373.53 | $336.08 | -10.0% | -5.0% |
| T4 | GLW | Corning | Fiber Optic | $150.38 | $135.18 | -10.1% | -5.1% |
| T1 | BE | Bloom Energy | Power Gen | $155.67 | $134.67 | -13.5% | -8.5% |
| T4 | APH | Amphenol | Connectors | $145.77 | $125.85 | -13.7% | -8.7% |
| T1 | MU | Micron | HBM Memory | $412.20 | $335.15 | -18.7% | -13.7% |
Key Takeaways
- Portfolio beat SPY by +1.3%. Both equal-weight (-3.8%) and tier-weighted (-3.7%) outperformed the S&P 500 (-5.0%) during the conflict period.
- Tier 2 was the sweet spot. Mid-cap infrastructure plays (EME, CLS, NVT, VRT, FIX) collectively lost only -0.8% — best tier by far. Labor moat thesis held up.
- Tier 3 also strong. Large-cap compute/grid names (GEV, NVDA, PWR, AVGO) showed defensive resilience at -1.0%.
- Tier 1 was mixed. MRVL (+20.8%) was a monster, but BE (-13.5%) and MU (-18.7%) dragged the tier down. High-conviction small caps = high dispersion.
- Tier 4 underperformed. "Safe" large caps (TSM, GLW, APH) fell harder than the market. Diversification didn't help when everything sold off.
- 15 of 23 stocks beat SPY. The basket provided broad outperformance, not just from a few winners.
⚠ HINDSIGHT CAVEAT: This is NOT a true backtest. The ranking model (Gabriel) was built on March 31 with full knowledge of the war's progression. Look-ahead bias contaminates these results. Treat this as feature analysis, not a prediction track record. Real out-of-sample tracking begins April 1, 2026.
Jim's AI Confirmation Bias — Lessons Learned
Jim correctly identified that Gemini was tilting bad news his way. The "daily morning briefing" format is particularly dangerous — it creates compounding urgency that drives emotional trading. Here are specific patterns to watch:
1
Gemini omitted Ukrainian anti-drone deployment until prompted — always cross-check for omissions
2
AI "daily briefings" create urgency bias — compare weekly trendlines, not daily noise
3
Worst-case scenarios ($185+ Brent) presented without probability weighting
4
Physical vs futures price divergence is more informative than either price alone
5
Trump "jawboning" artificially suppresses futures — watch Dubai physical for ground truth
6
Cross-reference AI summaries against EIA, IEA, and Goldman primary reports