Iran Oil Blockade|Feds next move in doubt?

Strait of Fire

Oil traded above $105 a barrel on Friday — roughly 50 percent higher than before Feb. 28. The U.S. military struck two Iran-flagged tankers in the Gulf of Oman the same morning Secretary of State Marco Rubio said he expected Iran to reply to a peace proposal "later in the day." Those two facts should not coexist, and the market's inability to price them cleanly is the real story of this week.

The Strait of Hormuz has been effectively closed since the war began, cutting off the passage through which one-fifth of global oil normally flows. That closure has already produced a measurable economic chain reaction. Maersk, which handles 14 percent of the world's containerized goods, disclosed an extra $500 million in monthly fuel costs — and told investors it cannot absorb all of it. Customers will carry a share. Goldman Sachs analysts wrote that if supply disruptions continue, elevated oil prices could persist through 2027. That window matters because it is long enough to change how the Federal Reserve has to behave.

The Fed connection is the mechanism most investors are underpricing. St. Louis Fed President Alberto Musalem said this week that risks have been "shifting toward the inflation side." Average U.S. gasoline prices hit $4.48 a gallon as of May 5, up from $2.99 the day the war started — a 50 percent spike in roughly two months. That acts as a regressive tax on lower-income households, compressing discretionary spending and threatening the consumer base that companies like McDonald's and Domino's rely on for same-store sales recovery. If those two forces — inflation above target and softening consumer demand — land simultaneously, the Fed cannot cut rates to defend growth without fanning prices further.

Iran's own economic collapse adds a layer that markets typically treat as exogenous noise but shouldn't. First vice president Aref acknowledged that some prices jumped more than 100 percent in less than a week. The rial hit an all-time low of 1.8 million to the dollar. The regime is paying wages and food subsidies from a collapsing fiscal base. That pressure is the most credible source of a genuine ceasefire — not diplomatic goodwill. The closer Iran's government gets to an inability to make payroll, the higher the probability of a deal. And a deal has an immediate, quantifiable market consequence: oil back toward $70, jet fuel costs down, airline stocks recovering, and suddenly the Fed's inflation calculus shifts the other direction. The question is not whether a peace deal is possible. It is whether the timeline fits inside the window where inflation expectations can still be re-anchored.

Memory Supercycle

The same week Maersk warned of demand destruction from high fuel costs, three memory chip companies reported results that seemed to belong to a different economic era entirely. Micron surged 11 percent Friday on Cloud Memory Business Unit revenue of $5.28 billion at a 66 percent gross margin. SanDisk jumped 11 percent on quarterly revenue up 251 percent year over year. Western Digital rose 3 percent after its non-GAAP gross margin crossed 50 percent for the first time.

That divergence — collapsing shipping profits, exploding memory margins — is not a contradiction. It is a structural split in where AI capital is flowing, and it explains why the VanEck Semiconductor ETF has gained 50 percent year to date even as the broader economic backdrop carries oil-shock uncertainty.

The causal driver is high-bandwidth memory, or HBM. Every large language model, every AI inference workload, and every agentic pipeline requires HBM sitting adjacent to the GPU. Supply is structurally short. Micron's order books reportedly stretch into 2027. Analysts running 39 Buy ratings against 5 Hold calls on Micron stock have a forward price-to-earnings ratio of 6x — implying an earnings ramp so steep that even after a 151 percent year-to-date gain, the stock looks cheap on a two-year earnings horizon.

The counter-signal worth watching is hyperscaler capex. Micron, SanDisk, and Western Digital all derive the bulk of their revenue growth from a small number of cloud infrastructure buyers: Amazon, Microsoft, Google, and Meta. If any of those buyers signals a pause or reallocation of AI spending — which Microsoft faces pressure on, with hedge fund TCI reportedly cutting its stake over AI disruption concerns — memory demand could soften faster than production capacity can be wound down. The historical pattern for memory is sharp upcycles followed by sharper corrections. What is different this time is that capacity discipline among producers appears more measured than in prior cycles, and the AI infrastructure buildout has a structural demand floor that commodity memory never had. Whether that floor holds at current price levels, or whether it is already reflected in a 528 percent year-to-date move in SanDisk, is the threshold that determines whether this trade still has room or is already pricing 2027 perfection.

Apple's Quiet Pivot

Apple reached a preliminary chip-making agreement with Intel on Friday, and Intel's stock jumped 19 percent to an all-time high. That single headline resolves a question that has quietly been building since Apple moved its Mac processors off Intel's x86 architecture in 2020: whether Intel's costly pivot to foundry manufacturing could ever win back the kind of customer relationship that once generated billions in annual revenue.

The deal's significance runs through Apple's supply chain logic, not Intel's balance sheet. Apple's consumer chip production is almost entirely concentrated at TSMC. As demand for Apple Intelligence features drives higher memory content and faster silicon refresh cycles — management flagged significantly higher memory costs ahead — that TSMC dependency becomes a supply constraint, not just a vendor relationship. Analyst Daniel Ives at Wedbush raised Apple's price target to $400 from $350, estimating that roughly 20 percent of the world's population will access AI through an Apple device. That distribution thesis only holds if Apple can manufacture at scale. Intel's 18A node is now in mass production, and the successor 14A process — expected to begin production in 2027 — could place Intel at the leading edge of semiconductor manufacturing for the first time in years.

The supporting case for Intel's rerating is that this is not a charity arrangement. Microsoft has also placed orders for Intel's 18A node. TSMC's foundry dominance has created a single point of concentration risk in the AI hardware supply chain, and both U.S. policy and corporate procurement strategies are pushing to distribute that risk. Musk's Terafab commitment to Intel adds another institutional anchor to the foundry thesis, though execution risk remains substantial — Intel has spent years and billions getting to a competitive process node after a period of significant delays and financial hardship.

The forward condition that separates a durable Intel rerating from a one-day headline pop is whether Apple's agreement extends to the 14A node or is limited to near-term supply diversification. If Apple commits to 14A volume — which would not ship until mid-2028 — Intel becomes a structural beneficiary of the AI device cycle for years. If the agreement stays shallow, Intel gets a credibility boost without the revenue foundation that would justify trading at all-time highs. Apple's share price at $293 on May 8, up 48 percent over the past year, reflects the AI distribution thesis Wedbush outlined. Whether the Intel partnership deepens that moat or simply adds a supply footnote is the condition that Wedbush's $400 target is implicitly betting on — and the answer will not come from Friday's headline.