Daily Brief - March 30, 2026

Daily Brief Podcast · Mar 30, 2026

Daily Brief - March 30, 2026

Iran War Is THE Market Variable Right Now

Sources: The Real Eisman Playbook (Steven Cook, CFR), The Compound and Friends (DataTrek), Kara Swisher (Sen. Tillis)

The Iran conflict is the single variable moving markets. Every headline produces a swing. Here is the state of play:

Oil is at $102 and making new highs. Gas above $4/gallon in parts of the US. Trump announced a 5-day pause on bombing Iranian energy infrastructure claiming talks are progressing -- Iran denied any talks are happening.
Base case is a "messy middle" in 3 months: regime survives but weakened, Strait of Hormuz partially contested, heavier US military presence than planned. No clean resolution for markets.
Russia is providing Iran real-time satellite intelligence for targeting US and Israeli facilities. Trump has not confronted Russia, caught between national security and his desire for low gas prices. Russian oil sanctions were lifted to increase global supply.
Gulf states (UAE, Saudi, Qatar) are angry but aligned after Iran bombed Dubai's airport, Saudi oil facilities, and Qatari targets. Expect massive investment in hardened defenses, bomb shelters, and anti-drone technology -- bullish for defense contractors.
Watch the 60-day War Powers deadline. Sen. Tillis (R-NC, retiring) will not support continued engagement without an AUMF and clear strategic objectives. Congressional support could fracture at that inflection point.
Freedom of navigation through the Strait of Hormuz is non-negotiable. If Iran institutionalizes selective passage control, it is a permanent risk premium on global energy prices.

Actionable: Hold energy stocks. DataTrek says never sell energy making new highs. The 1990 Gulf War analog says markets bottom when oil peaks, not before. Defense and anti-drone stocks are positioned for increased Gulf demand.

Market Levels and Oversold Signals

Source: The Compound and Friends (DataTrek -- Nick Colas, Jessica Rabe)

S&P 500 key buy level: below 6,050 triggers a 2-sigma oversold signal with 92% historical win rate for positive 50-day forward returns (avg +9.6%). Currently at -8.1% on 50-day rolling basis.
Nasdaq key buy level: below 20,650 triggers 2-sigma oversold with 81% win rate. Currently at -10.8%.
VIX between 27-43 historically produces good forward returns. Above 43, outcomes depend on a policy response materializing -- and right now there is no obvious policy lever (unlike the April 2025 trade reversal).
Three headwinds converging: recession risk from high oil, war uncertainty, and potential Fed rate hikes from combined tariff + oil inflation. S&P already down 7% YTD. Years with all three factors average -19% declines.
Exxon (XOM) now trades at 25x earnings vs. Nvidia (NVDA) at 20x. Portfolio managers are gravitating toward energy dividend certainty over speculative AI reinvestment.

Stocks mentioned: XOM (record highs ~$175), NVDA (20x P/E), XLE (preferred over small-cap energy PSCE), IGV (software ETF, oversold -- contains PLTR, MSFT, CRM, CRWD, PANW).

Bill Ackman: "Stocks Are Stupidly Cheap"

Source: Motley Fool Money

Specific names flagged as irrationally undervalued:

Lululemon (LULU): ~11x trailing earnings, a valuation usually reserved for slow-growth retailers. Dominant brand, industry-leading margins, large untapped international runway (especially China).
Microsoft (MSFT): Low 20s P/E. Positioned as "the essential operating system for AI."
Alphabet (GOOGL): Low 20s P/E. Growing TPU business, AI integration across its ad machine.
Howard Hughes Holdings (HHH): Trading in the low $60s; Ackman bought at $100 last year. Business performing as expected operationally.
Fannie Mae (FNMA) / Freddie Mac (FMCC): Ackman calls these potential 10x plays. His cost basis on FNMA is ~$2.29/share (now ~$6), estimates worth $34+ if conservatorship ends. Trump administration discussing reprivatization and a potential $30B IPO. Risk: removing conservatorship could destabilize the mortgage market while rates are high.
SpaceX IPO: Expected to be the largest IPO in history -- $75B raise at $1.75T valuation. 30% retail allocation (vs typical 5-10%). Valuation driven by Starlink profitability, merged xAI/X platform, and orbital AI data centers. Caution: much rests on unproven future tech.
AI beneficiary aggregators: Jefferies flagged Expedia (EXPE), Instacart/Maple Bear, and Uber (UBER) as AI beneficiaries, not victims. Uber singled out as best positioned.

Private Credit: Liquidity Crunch Building

Source: Capital Allocators -- Kieran Goodwin (Sabah Capital)

Most time-sensitive credit market item. The private credit market ($350B in non-traded BDCs + $100B in interval funds, all since 2018) is showing stress:

Dividend cuts of ~10% at Blackstone B-CRED, Goldman, and Oak Tree have triggered redemptions doubling from 2.1% to 4.3%.
Bear case is reflexive: redemptions force selling, bids vanish, bank lines get pulled, repeat. Tail risk extends to insurance/annuity companies loaded with BBB-rated private credit.
SaaS/software lending is the weak spot. AI is disrupting 80% gross margins in software. Funds with 50%+ software concentration are particularly exposed. CLO equity tranches already absorbing disproportionate damage.
Opportunity emerging: Private credit secondaries -- quality loans could clear in the low 90s at ~11% unlevered yields. Sabah is tendering for Blue Owl OBDC2 at a discount to NAV.

Actionable: Reduce exposure to non-traded BDCs and interval funds heavy on software lending. Watch for more dividend cuts as leading indicators.

Goldman Sachs AI: Past Experimentation, Into Production

Source: Odd Lots -- Marco Argenti, Goldman CIO

47,000 employees use GS Assist daily (1M+ prompts/month). Goldman has already terminated third-party software contracts in favor of AI-built internal replacements.
The buy-vs-build equation has flipped for simple apps. SaaS tied to processes AI is transforming (developer tools, simple UX apps, survey/expense tools) faces disruption. SaaS tied to stable, regulated processes (accounting, ERP, systems of record) is more defensible.
Framework: Will the underlying business process change in 5 years? If yes, the incumbent software vendor is at risk.
Token costs per unit decline, but total enterprise AI spend rises significantly as reasoning and agentic workflows consume far more tokens. Bullish for GPU makers and cloud infrastructure.
Forward-deployed engineers from model providers (Anthropic, etc.) are the preferred integration model, disadvantaging traditional IT consulting/integration firms.
Goldman is NOT reducing developer headcount -- productivity gains go to building more from the existing backlog.

AI Skills Are Now a Cross-Platform Standard

Source: AI News & Strategy Daily

Anthropic, OpenAI, and Microsoft have agreed on a shared "skills" file format -- simple markdown files with metadata that agents (not humans) now primarily call. Key takeaways:

Skills compound; prompts do not. Every refinement persists. Shift practice from prompt iteration to skill iteration.
Design skills as agent-callable contracts with explicit output formats and composability for multi-step workflows.
Team deployment in three tiers: Tier 1 (org-wide standards), Tier 2 (methodology skills capturing senior practitioner craft knowledge -- highest alpha), Tier 3 (personal workflow).
Career move: Open-sourcing high-quality domain-specific skills works like an open-source portfolio for hiring.

Action item: If you use AI tools regularly, start converting your repeated workflows into versioned skill files now. This is becoming infrastructure, not a nice-to-have.

ARK Invest: Autonomous Vehicles at $34T by 2030

Source: ARK Invest Big Ideas 2026 -- Tasha Keene

Robotaxis can price profitably at $0.25/mile vs. ~$3 for ride-hail and ~$1 for personal car ownership.
Tesla (TSLA) has the largest cumulative autonomous data set and a 35-50% cost advantage per mile over Waymo's retrofitted vehicle approach.
Uber (UBER) and Lyft (LYFT) face direct disruption. Only ~140,000 robotaxis needed to replace all current US ride-hail volume -- roughly one month of Tesla production.
Baidu (BIDU) leads commercial driverless miles in China via Apollo Go. Chinese manufacturing scale means competitors could appear rapidly.
Value accrues to the autonomous technology platform layer, not vehicle manufacturers or fleet operators.

How to Actually Use AI Well

Source: Art of Charm -- Dr. Vivienne Ming

Only 5-10% of people become true "cyborgs" who integrate AI with their own reasoning for superhuman performance. Most blindly follow AI or ignore it.
When AI was redesigned to only ask questions (never give answers), twice as many people hit superhuman performance -- but they hated using it. Comfort and effectiveness are at odds.
Anthropic's own research: Developers using Claude Code were faster but learned dramatically less. AI is the new GPS -- better with it, worse when you stop.
Foundational skills (curiosity, resilience, meta-uncertainty) predict career outcomes better than technical credentials across 122 million people studied.

Practical actions: Use AI as a Socratic partner (prompt it to challenge you, not just answer). Keep a failure diary. Develop meta-uncertainty before consulting AI. Invest in foundational learning skills over certifications.

Mistral: Open-Weight Voice AI and Enterprise Customization

Source: Latent Space -- Pavan Kumar Reddy, Guillaume Lample

Voxtral TTS: 3B parameter open-weights text-to-speech model using novel autoregressive flow matching. Dramatically lower latency than competitors, designed for real-time voice agents at a fraction of ElevenLabs' cost.
Forge platform gives enterprises the same fine-tuning tools Mistral's internal team uses. Companies using off-the-shelf closed models are leaving value on the table by not leveraging their proprietary data.
Leanstral: Formal math proving in Lean, showing transfer learning from formal proofs to general reasoning. Formal verification industry expected to grow as AI lowers the barrier.
No dominant architecture exists for audio AI (unlike text transformers), meaning many low-hanging research opportunities remain.

Political Risk Monitor

Source: On with Kara Swisher -- Sen. Tom Tillis (R-NC)

Independent voter erosion is the red flag for 2026 midterms. 125,000 more Democrats voted in the NC primary than Republicans. 85% of independents have disengaged from or turned against the GOP. Suburban districts at risk.
Fed independence: Tillis is blocking Kevin Warsh's Fed chair confirmation until DOJ drops its investigation of Jerome Powell. No path to discharge from committee.
Filibuster is safe for now but remains a recurring pressure point the administration will keep pushing.
DHS funding impasse (tied to Save Act/voter ID legislation) needs resolution to avoid cascading impacts on TSA, Coast Guard, and other agencies.

Huberman Lab: Hormones and Sexual Orientation

Source: Huberman Lab -- Dr. Marc Breedlove

Science episode with no direct financial implications. Key finding: sexual orientation is significantly shaped by prenatal testosterone exposure and maternal immune responses. The fraternal birth order effect (each older biological brother increases probability of homosexuality by ~33%) is driven by maternal antibodies. Health note: CAH (congenital adrenal hyperplasia) carrier rate is ~1 in 12, far more common than most realize.