Daily Brief — March 31, 2026

Tuesday, March 31, 2026

---

🌍 Geopolitics

Iran Opens Strait of Hormuz to Some Ships — But the Damage Is Done

Iran made a partial concession Tuesday, allowing limited commercial shipping through the Strait of Hormuz after weeks of near-total blockade. But "limited" is doing a lot of heavy lifting. Daily oil transit through the strait has collapsed from 20 million barrels to a trickle, and the 53% surge in crude prices over March tells the real story.

The backdrop: 7 Iranian missile salvos struck Israel on Monday. Two more UNIFIL peacekeepers were killed in southern Lebanon. Israel announced plans to further expand its "buffer zone." And in the U.S., the White House floated the idea of passing war costs to regional states — a trial balloon that landed with a thud in Gulf capitals.

This matters because: Brent crude just posted a 55% monthly gain — the largest since the contract's inception in 1988. WTI settled above $100 for the first time since 2022. The Hormuz blockade directly threatens 20% of global seaborne oil trade, and while Asian nations with massive strategic reserves (China: 1.4B barrels, Japan: 350M) can weather 150-180 days, India and South Korea have just 33-60 days of import cover. Europe's EuroStoxx 50 is down 12%. The U.S., as a net energy exporter, is structurally insulated — but not immune to the second-order chaos.

Trump Signals Willingness to End Iran Campaign — Markets React Instantly

The Wall Street Journal reported that President Trump has signaled willingness to end the U.S. military campaign against Iran. Stocks immediately rallied on the headline. Oil swung wildly between gains and losses. But the mixed messaging — simultaneously floating "take the oil" scenarios — is generating a credibility gap that's hurting GOP positioning heading into midterms.

This matters because: Markets have "priced in" a 4-6 week conflict. If this extends past mid-April, analysts warn of a fundamental repricing of risk that could push equities into true correction territory. The divergence between headline-driven rallies and underlying deterioration makes this a trader's market, not an investor's market.

---

📈 Financial Markets

Markets Whipsaw: Dow +0.1%, S&P -0.4%, Nasdaq -0.8%

A volatile Monday saw sharp divergence despite dovish Powell remarks. Treasury yields retreated sharply, but the relief rally in equities couldn't hold. The Nasdaq and Dow both entered correction territory last week. Key culprit beyond geopolitics: Google's "TurboQuant" AI memory compression breakthrough hammered semiconductor stocks.

The numbers: Micron (MU) has fallen 30% over the last 8 trading sessions. Samsung and SK Hynix dropped 6%+ in Seoul. The fear: Google's algorithm could dramatically reduce AI memory chip demand. The reality: Micron's HBM capacity is sold out through all of 2026, so near-term demand isn't at risk. This is a classic case of long-term thesis anxiety colliding with short-term fundamentals.

Oil: Brent at $112.90/bbl (+0.16%). WTI at $102.30/bbl (-0.5%). Record monthly surge for both benchmarks.

Powell: "We Can Wait and See" — No Rate Hike Despite Oil Shock

Fed Chair Powell spoke at Harvard Monday, delivering the message markets wanted: the Fed views the oil price spike as a transitory supply shock and sees no need to raise rates from the current 3.50%-3.75% range. He called the current rate "a good place" to observe unfolding events including the Iran war and tariff impacts. The Fed will "look through" the oil shock rather than react to it.

This matters because: This is the Powell Put in action. By explicitly ruling out rate hikes in response to the oil shock, Powell is telling markets the Fed won't compound geopolitical pain with monetary tightening. Treasury yields fell sharply in response. For insurance companies and other fixed-income-heavy portfolios, this means the yield environment stays favorable but won't improve further near-term.

---

🤖 AI & Tech

Anthropic's "Mythos" — The Most Powerful AI Model Ever Built, Leaked by Accident

In one of the most embarrassing security lapses in AI history, Anthropic accidentally left draft blog posts about its next-generation AI model in a publicly accessible data store. Fortune and independent cybersecurity researchers discovered nearly 3,000 unpublished assets, including details of "Claude Mythos" — which Anthropic acknowledges is "a step change" and "the most capable we've built to date."

Key details from the leak: Mythos introduces a new model tier called "Capybara" — larger and more powerful than Opus, the current top tier. The draft blog claims "dramatically higher scores on tests of software coding, academic reasoning, and cybersecurity." The model is already being tested with early access customers. Anthropic also had leaked details of an invite-only CEO summit in Europe targeting enterprise sales.

This matters because: The irony of an AI safety company having a major data leak is not lost on anyone. But the substance is what matters: if Mythos/Capybara delivers on the leaked benchmarks, it represents a genuine capability leap. For enterprises evaluating AI vendors, this changes the calculus. The cybersecurity risk angle — Anthropic itself flagged "unprecedented cybersecurity risks" — adds fuel to the regulation debate.

Google's TurboQuant Shakes the Semiconductor World

Google DeepMind published research on a new compression algorithm that could dramatically reduce the memory required to run AI models. The implications sent shockwaves through memory chip stocks: Samsung, Micron, SK Hynix, Western Digital, SanDisk, and Seagate all sold off hard.

This matters because: The AI infrastructure buildout has been predicated on ever-increasing demand for memory chips, especially HBM (high-bandwidth memory). If TurboQuant or similar techniques can reduce that demand significantly, the entire capex thesis for memory chipmakers needs revision. Counterpoint: these efficiency gains historically increase total compute usage (Jevons paradox), which could ultimately increase memory demand. The market hasn't figured out which narrative wins yet.

---

🏢 Insurance Industry

U.S. Treasury Convening Insurance Regulators Over Private Credit Risks

Treasury Secretary Scott Bessent is expected to announce as early as Wednesday the first in a series of meetings with all 50 U.S. state insurance regulators focused on the $2 trillion private credit market. The Treasury has no direct regulatory authority over insurance, but Bessent is positioning the department as a "convening authority, resource and forum."

Key concerns on the agenda: fund-level leverage, consistency of private credit ratings, use of offshore reinsurance, and liquidity of private credit investments. Bessent warned that working Americans' retirement accounts should not become "a dumping ground" for "rotten" assets.

This matters for Johal Insurance: This is the most significant federal-level engagement with insurance regulation in years. The private credit nexus — where PE-backed insurers have been loading up on illiquid assets — has been a ticking time bomb that state regulators haven't had the resources to monitor adequately. If Treasury succeeds in improving transparency, it could reshape how insurers allocate their investment portfolios. Watch for downstream effects on commercial insurance pricing and carrier stability. Any carriers heavily exposed to private credit could face ratings pressure.

---

🚀 Startup Spotlight

Gimlet Labs — $80M Series A for Multi-Silicon AI Inference

What they do: Gimlet Labs has built a "multi-silicon inference cloud" — software that optimizes AI model inference across different chip architectures (NVIDIA, AMD, custom silicon). Think of it as an abstraction layer that prevents vendor lock-in at the inference stage.

The raise: $80M Series A led by Menlo Ventures — an unusually large Series A, signaling high conviction.

Why it's interesting: As the AI industry shifts from training (dominated by NVIDIA) to inference (where the long-term revenue lives), the ability to run models efficiently across multiple chip types becomes a massive competitive advantage. Gimlet Labs is betting that the future of AI compute isn't NVIDIA-only — and given Google's TurboQuant news this week, that bet looks increasingly smart. Every dollar saved on inference is a dollar that goes to the bottom line for AI-deploying enterprises.

---

💎 Potential 10-Bagger: Palantir Technologies (PLTR)

The thesis in one line: Palantir is the only pure-play AI-for-government company at scale, and we're entering the most defense-intensive period since the Cold War.

Current setup: Palantir just won a major Pentagon contract (announced March 27). The Department of War (rebranded from Department of Defense) is accelerating procurement of AI-driven battlefield intelligence, and Palantir's Gotham and Foundry platforms are the incumbents. Revenue growth has been accelerating — commercial adoption via AIP (Artificial Intelligence Platform) is the kicker Wall Street underestimates.

The 10-bagger case:

• Defense spending supercycle: The Iran war, coupled with rising great-power competition, is driving a structural increase in defense budgets globally. Palantir sits at the intersection of AI + defense + intelligence — the three fastest-growing government spend categories.
• Commercial flywheel: AIP boot camps are converting enterprise customers at unprecedented rates. The TAM expansion from government-only to government-plus-commercial doubles the addressable market.
• Earnings leverage: Palantir's margins are expanding as its platform business scales. The next earnings report (May 11) could be a catalyst if commercial growth accelerates.

The risk: Valuation is rich. PLTR has always traded at a premium to fundamentals. If the Iran conflict resolves quickly, the defense spending narrative may deflate. And competition from Microsoft/Azure Government and AWS GovCloud is real.

Verdict: In a world of permanent geopolitical instability, AI-powered defense intelligence isn't a luxury — it's infrastructure. Palantir is building the operating system for 21st-century warfare. Position sizing matters: this is a high-conviction, high-volatility name. Not financial advice — do your own research.

---

Johal Capital Daily Brief | March 31, 2026 | 🦊 Zanu

---

Disclaimer: Informational only; not investment advice.