0) THEME OF THE WEEK
Three interlocking threads:
Iran War — Strait of Hormuz approaching 4+ weeks closed. "Taco trade" extended 5 more days (Trump delayed his own deadline). Iran says no talks exist. Classic Column 1 → Column 2 migration risk.
AI/SaaS Disruption — Claude Co-work keeps getting named as the killer of per-seat enterprise SaaS. Usage-based infrastructure = fine. Per-seat = endangered.
Technical Exhaustion of Sellers — NASDAQ RSI hit 29. VIX at 30+. Three specific signals still missing before the real bottom is confirmed (VIX 40-60 spike, breadth thrust, big green volume node).
1) WHAT ACTUALLY MATTERS THIS WEEK
Theme | Status | Key Number |
|---|---|---|
Iran / Hormuz closure | 🔴 Live, unresolved | Brent crude ~$100-105/bbl. $120+ = stagflation risk |
Claude Co-work disruption | 🔴 Live, accelerating | Anthropic ARR: $1B → $9B → $14B → $19B in 3 months |
Google TurboQuant | 🟡 Digesting | 6x memory compression. Deepseek-style overreaction on Micron |
Meta legal verdicts | 🟡 Ongoing | $4.2M LA + $375M NM. 2,000+ pending suits. Section 230 at risk |
Robinhood $1.5B buyback | 🟡 Fresh | Barely covers SBC dilution. Starts end of Q1 |
Elliott / Synopsys | 🟢 Actionable NOW | Multi-billion stake disclosed Sunday. SNPS at $416, target $540 |
2) CONSENSUS (What multiple sources agree on)
✅ Iran ends eventually — question is weeks vs. months vs. year+ ✅ Hyperscaler AI capex is real and unstoppable ($600B+ in 2026 from 4 companies alone) ✅ Per-seat SaaS is in structural trouble (Salesforce, ServiceNow, Adobe) ✅ Amazon is historically cheap — 9.7x projected 2027 FCF ✅ Nvidia is not expensive — ~21-24x forward PE with 60% operating margins ✅ Meta is a buy — Forward PE sub-20, core business unchanged, $500 is the deeper buy zone ✅ Gold long-term thesis intact despite short-term forced selling ✅ DCA in tranches — nobody can call the bottom
3) KEY DISAGREEMENTS
Salesforce: Buy or Avoid?
Bull (Everything Money): $195 = 12.5x FCF, $290 DCF intrinsic value. 150K enterprise customers aren't leaving overnight.
Bear (Brian Stofl, sold the position): New users will never pay $175/seat when Claude costs $20. Terminal value collapses. Verdict: Both correct on different time horizons. Not a 2026 trade.
How bad is Iran really?
Optimists: "Ends in weeks, buy aggressively."
Skeptics (Brian Stofl, Jordan/Dumb Money): Iran may have 6-12 months of drone/missile inventory. No Pentagon transparency. No clean exit path. Verdict: Skeptics more right on uncertainty; optimists may be right on eventual outcome.
Robinhood: Buy or pass?
Bull: Down 36%, buyback authorized, app downloads at all-time high.
Bear (Travis/Asymmetric): 14.8x price/sales historically elevated. Users flat. Better to use cash for growth. Verdict:High-beta market bet, not asymmetric opportunity.
4) TRADE IDEAS WITH TIME LABELS
⚡ ACT ASAP
SNPS — Synopsys | ~$416 Elliott Management (76B AUM, best activist track record ever) disclosed multi-billion stake. Controls ~70% of global EDA market — every AI chip being designed uses their software. In penalty box post-Ansys acquisition. Elliott playbook: buybacks + margin expansion = re-rating.
Trigger: Market opens Monday; analyst upgrades likely
Target: $540 consensus | $650 high
Invalidation: Broader market -10%+; Ansys integration fails
Conviction: Medium-High | 🟢 Fresh
VRT — Vertiv | ~$267 Joined S&P 500 on March 23. $10T+ in passive assets must now buy it. Record $15B backlog (+109% YoY). Orders grew 252% YoY. Co-designs liquid cooling with Nvidia directly — baked into every Vera Rubin data center blueprint. Norwegian sovereign wealth fund bought $538M position.
Trigger: Passive index fund mandate (mechanical, ongoing now)
Target: $320 bull analyst target
Invalidation: Market overwhelms inclusion bid
Conviction: High | 🟢 Fresh
📅 NEXT WEEK
NVDA — Nvidia | ~$175-180 6 months sideways while the company got significantly better. Grock acquisition = training + inference platform combined. $1T AI chip demand by 2027 (Jensen, conservative). DCF probability-weighted upside ~25%. Heaviest institutional call option buying among all Mag7.
Trigger: Hold 200-day MA at $179; Iran deescalation signal
Invalidation: Broader flush; hyperscaler capex pause rumors
Conviction: High | 🟢
META — Meta Platforms | ~$540-560 Down 30% from highs. Legal fines are rounding errors vs $200B revenue. Core business untouched by Iran. Forward PE approaching 16-17x (trough). Multiple DCF estimates: $645-$900. Watch $500 for heavier buy.
Trigger: No new shock verdicts; Iran deescalation
Invalidation: Congress acts on Section 230; more landmark verdicts
Conviction: High | 🟢
AMZN — Amazon | ~$185-195 9.7x projected 2027 FCF — cheapest in modern history. AWS at 24% growth, fastest in 13 quarters. Every Amazon capex cycle has eventually paid off. Sentiment finally shifting.
Trigger: Accumulate; no specific catalyst needed
Invalidation: AWS slows below 20%; recession materializes
Conviction: High | 🟢
AMD | ~$95-105 #1 named beaten-down semiconductor by multiple sources. Reclaimed 50-day MA on March 25 (+7.5%). OpenAI and Meta partnership developments. High beta to SMH recovery.
Trigger: Hold above 50-day MA; semiconductor sector follow-through
Invalidation: Drops back below 50-day MA; broader flush
Conviction: Medium | 🟢
PANW — Palo Alto | ~$160-170 Inverse head & shoulders. Neckline ~$170. Target ~$199. Fresh technical setup, confirmed working per active trader.
Trigger: Break and hold above $170
Invalidation: Drops below left shoulder; Claude Mythos fears hit cyber harder
Conviction: Medium (technical trade) | 🟢
🗓️ NEXT 30 DAYS
CRDO — Credo Technology 200%+ revenue growth. 68% gross margins. 88% market share in AI cluster high-speed connectivity. 51% off highs. 884 institutional holders. Analyst consensus ~$200 (90%+ upside). The cables connecting GPUs inside every major hyperscaler.
Conviction: High | 🟢
ARM — ARM Holdings | ~$135 March 24: AGIC CPU for agentic AI announced. Lead customer: Meta ($135B 2026 capex). TSMC 3nm manufacturing. Other customers: OpenAI, Cloudflare, SAP. Revenue $4.7B trailing — adding "billions." Analyst first-ever buy rating. Ships H2 2026.
Conviction: Medium-High | 🟡
SOFI — SoFi | ~$15-16 $3.6B new loan platform business (fresh). Muddy Waters report rejected by two major analysts who maintained/raised targets ($25 Goldman, $38 Isizuo). 57th largest US bank. SoFi Plus launching with 4.5% APY. App downloads at all-time company high.
Conviction: Medium-High | 🟢
ORCL — Oracle | ~$143 AI database as the enterprise cross-sell. 18% forward revenue growth. Deep Fortune 500 relationships. Down from $344 highs. DCA target: under $150.
Conviction: Medium-High | 🟢
MGM Resorts | ~$22-25 Japan resort (2030) could be worth more than entire current $9.5B market cap. Marina Bay Sands comparable generates $2.9B adj. EBITDA. Buying back 10-15% of shares annually. Bet MGM crossing to cash generation. Completely ignored by AI/Iran narrative.
Conviction: Medium-High | 🟡
Gold / GLD Down $1,000+ from January all-time high. Turkey forced-sold 58 tons in 2 weeks. Gulf states selling to maintain dollar pegs. Russia April export ban adds more pressure near-term. But: long-term debt/dollar debasement thesis unchanged. Smart money indicators turning positive.
Strategy: DCA in tranches, avoid leverage, expect more pressure before April
Conviction: Medium-High (long-term) / Low (timing) | 🟢
Defensive: PG, KMB Both near cheapest valuations since 2023 on forward metrics. P&G DCF: $198 intrinsic vs ~$144 current. KMB: 13x forward P/E. Low beta. Consumer staples outperforming in current environment.
⚠️ Sources are 5 days old — verify current prices before acting
Conviction: Medium | 🔴 Verify first
5) RISK MAP
Risk | Timeline | Severity |
|---|---|---|
Iran escalates (infrastructure damage, not just transit) | 0-2 weeks | Catastrophic — market -15%+ |
Fed signals rate hike | 0-4 weeks | High — multiple compression |
Q1 earnings miss on energy/supply chain | 1-2 months | Medium-High |
AI efficiency narratives keep hitting (TurboQuant, Mythos, etc.) | Ongoing | Medium |
China economic slowdown from Hormuz | 1-3 months | Underappreciated |
Baby boomer retirement fund outflows | 2026-2028 | Long-term structural headwind |
6) IF WE'RE WRONG, WE'RE WRONG BECAUSE...
Iran is actually Column 2 — and we won't know for 4-6 more weeks while every DCA buyer accumulates losses
Claude Co-work disruption is faster than the "slow moat erosion" bulls think — Anthropic doubling ARR off $9B in 3 months is not slow
The real bottom isn't in yet — VIX needs to spike to 40-60, breadth thrust needs to fire, large green volume node needed. None of these have happened. Every "buying now" call is technically premature.
Gold selling cascade is larger than estimated — Russia April ban + Gulf state selling could push gold to $3,500 before reversing
Nvidia timeline extends — Grock ships Q3 2026, hyperscalers historically pause capex cycles; if either slips, the re-rating waits
Not financial advice. Probabilistic scenario planning only.