Market Overview: March 10, 2026

Summary

• The biggest risk this time is the unpredictability of the shooting war itself. After U.S.-Israeli strikes on Iran, Iran harassed the strait, oil surged intraday to $115 on Monday, rates in the U.S., Europe, and the U.K. swung sharply in sync, and front-end trades betting on a BOE cut “completely collapsed.” Unlike the rate volatility triggered by the BOJ in January, that episode stayed at the economic level and remained controllable—the speaker judged at the time that 10-year U.S. Treasury yields would top out at 4.2%. This time, the setup is stagflation pricing, and the market is trading a bear flattening: short-end rates are rising faster than long-end rates.
• But the speaker believes oil prices will not pass through to core PCE in the short term. The Fed is focused on core PCE now; oil prices “could rise to 300 or 500” and, for a few weeks, only make gasoline more expensive—services, rent, and labor costs will not rise. The closest comparison is tariffs on China, whose pass-through to PCE was only marginal, at 0.2 to 0.5 percentage points. The one-year fundamental backdrop remains benign growth, disinflation, and an unemployment picture that does not look good; the real variable is that “war itself costs money,” and the story becomes harder to call if the war drags on.
• Iran is highly likely to stay hardline to the end. The speaker cited the framework of Kohan (phonetic), the State Department’s head of global affairs and a former Hillary adviser: after Khamenei was killed and made a “martyr,” Iran is effectively a decentralized government controlled by IRGC commanders; the new leader, Khamenei’s son, is neither an ayatollah nor influential in the military. The war has converted domestic tensions—90M people, 5M taking to the streets, and a crackdown that killed tens of thousands—into an external conflict. The commanders’ overriding interest is to “put the entire national disaster on the Americans,” making regime change “very, very difficult.”
• Positioning risk is asymmetric, with a major landmine around 6500. The S&P hitting around 6750 would flip short-term CTAs short, and around 6650 would flip medium-term CTAs short; if CTA positioning is flat, it would have to sell $21B. At around 6500, short-term gamma turns severely negative, accelerating selling. Meanwhile, top-of-book orders are only one-fourth to one-fifth of January, hedge funds sold through ETFs throughout February as volume reached a record high, and institutions are leaning bearish—thin liquidity plus negative gamma means “once sentiment turns bad, it falls very hard.”
• The upside is no easier. There is a very strong gamma wall at 6900-7000, with many option sellers positioned there; breaking through requires volatility to fall and CTAs to re-lever and buy. “The market still lacks sufficient capital.” On Monday, Trump appeared to hint before the close that the war was nearing an end, and the S&P reversed from a loss of more than 2% overnight while the Nasdaq reversed from a 2.5% loss to finish higher. But a shooting war is not a tariff war: “a shooting war carries the risk of getting out of control,” with World War I an example of a war nobody wanted that ultimately did.
• If there is a pullback, it is a buying opportunity, not the start of a bear market. U.S. equity valuations are not expensive: Nvidia may trade at roughly 12-13x forward P/E next year, with growth around 80% this year and potentially around 50% next year; Microsoft is at 22x P/E, roughly the same as when the market bottomed last April. The trade is to buy the accessories behind AI’s big shovels: if optics and memory keep pulling back because of deleveraging—Samsung and SK Hynix’s adjustments came from unwinding the dollar-won carry trade, not fundamentals—that is a good opportunity. The least-priced-in piece is the CPU: AI calls require CPUs directly; Nvidia has Grace and may later have an in-house CPU that sounds like Vera; news may be announced on March 16, at which point an x86 chip might be needed, “which would be a major help to Intel.”

Deep dive

1. Interest rates are the key variable to watch right now—this time the market is pricing stagflation, not an economic war

• The sharpest move last week was in rates, and not just in the U.S.: euro and sterling rates moved in tandem. The chain reaction was Iran blockading and harassing the strait → oil surged on Monday, reaching $115 intraday in Asia before falling back overnight → front-end trades betting on a BOE cut “completely collapsed.”
• The speaker contrasted this with January: the BOJ’s dovishness plus the Japanese government’s plan for large-scale physical expenditure caused long-end duration risk to collapse, spilling over into U.S. and European rates. But that episode “stayed at the economic level and was fully controllable”; even the tariff war “was, after all, an economic war, and once Trump started tackling it, it was controllable,” so he judged at the time that 10-year U.S. Treasury yields would top out at 4.2%. This time the setup is stagflation, and the short-term market trade is bear flattening—front-end rates are repricing faster than the long end.
• The rise in rates is also triggering deleveraging: the dollar-to-EM carry trade into the Korean won and Latin America was unwinding on Monday. This was money “funneling liquidity into EM markets”; de-risking could set off a chain reaction, with oil feeding into rates and credit, then equities. “This is what we see as the biggest risk in the market right now—and the least measurable.”

2. Oil prices are unlikely to pass through to core PCE in the short term; the fundamental script is unchanged

• The speaker sees a Fed shift toward rate hikes driven by oil as unlikely. The Fed watched CPI in the 1970s; today it watches core PCE—regardless of whether oil hits $300 or $500, within a few weeks it can only make gasoline more expensive; services, rent, shelter, and labor costs do not rise. The closest comparison is tariffs on China, whose pass-through into core inflation was only marginal, at roughly 0.2 to 0.5 percentage points last year. Oil must stay high for an extended period before gradually passing through to the front end.
• The one-year fundamental backdrop is unchanged: relatively healthy growth, disinflationary inflation dynamics, and an unemployment rate that “doesn’t look good.” The only caveat: “If this war goes on for a very long time… it’s hard to say, because war itself costs money.”

3. The war is actually helping the IRGC—the military hardliners are more likely to fight to the end

• The speaker cited Kohan (phonetic), the State Department’s head of global affairs, a former Hillary adviser who helped formulate Middle East policy: Iran is a theocracy that had contingency plans for a “decapitation strike.” If the supreme leader went incommunicado, each IRGC commander would launch missiles according to plan. That is why, once war began, Iran struck every nearby country with a U.S. military base or close U.S. ties—even “brother countries” like Azerbaijan and Turkey. This suggests Iran is now a decentralized government run collectively by a group of military commanders.
• Khamenei’s newly selected son is neither an ayatollah—his theological standing is weak—nor influential in the military, so his power is heavily constrained by the IRGC. During Iran’s economic crisis at the start of the year, 5M of its 90M people took to the streets and the crackdown killed tens of thousands. Khamenei and the IRGC, which was deeply tied to him, had been headed for “the pillar of historical shame… total ruin, only a matter of time.” By killing Khamenei and turning him into a martyr, the United States instead allowed the external conflict to overwhelm domestic grievances.
• The commanders are therefore more likely to fight to the end than seek a deal, putting “the entire national disaster on the Americans.” Overthrowing the IRGC regime would be “very, very difficult.” The cost of Tehran remaining hardline is also limited: its missile launchers are destroyed by the U.S. five minutes after launch, and its ammunition projection is being halved each day, so it has little visible ability to launch further ballistic missile attacks on Israel proper. But continuing to harass the strait and drive up shipping insurance is enough; a U.S. escort guarantee is “not a long-term solution,” and the market may shift to pricing in a different possibility: “this war may never end.”

4. Positioning map: 6900-7000 is a gamma wall; 6500 is a major landmine below

• Hedge funds sold through ETFs throughout February, with ETF turnover reaching a record high, while institutional shorts became “more determined.” Top-of-book orders were only one-fourth to one-fifth of January—5 levels on each side versus 25—and market makers had cut displayed orders by more than half. The buyers have thinned out.
• The downside traps: after months of sideways trading, short-term CTAs turn short at 6700-6750, and medium-term CTAs at 6650-6700. If S&P CTA positioning is flat, closing it requires $21B of selling. Dealer gamma has turned negative—“when prices rise it buys, when they fall it sells… once sentiment turns bad, it falls hard.” The acceleration point is 6500, where short-term gamma becomes “deeply negative” and heavy selling floods in—a “very large landmine” roughly 5% lower.
• The upside is also difficult: many option sellers are concentrated at 6900-7000. A breakout requires capital to build, volatility to fall, and CTAs to re-lever and buy; “the market still does not have enough capital.” On Monday, Trump appeared to hint before the close that the war was nearly over; after the S&P fell more than 2% overnight and the Nasdaq 2.5%, both reversed to close higher—a classic TACO-style rally. But the speaker’s caveat is worth noting: “A shooting war carries the risk of getting out of control. In World War I, no country really wanted war; everyone wanted to contain it, and it ultimately spiraled out of control.”

5. If it really sells off, buy the accessories for AI’s big shovels—the least-priced-in piece is the CPU

• This is not a valuation-driven collapse or a bear market: Nvidia may trade at roughly 12-13x forward P/E next year, with growth around 80% this year and potentially around 50% next year. “Look at any domestic company with 12-13x forward P/E next year and that kind of growth.” Microsoft is at 22x P/E, with valuation around where it was at the bottom of last April’s selloff. If a decline comes, it will be positioning-driven and “give investors plenty of buying opportunities.”
• The framework: semiconductor demand has been rewritten in the AI era. Nvidia “isn’t making GPUs now—it is building mainframes”; once it delivers a full solution, demand for optics, copper, connectivity, and memory changes too. “Who knew before that we would need something like Nokia?” Optics and memory had “run a little too far,” but another pullback would be a good opportunity—especially since the declines in Samsung and SK Hynix reflect the unwinding of a dollar-won carry trade and Korea-wide deleveraging, not fundamentals.
• The most overlooked link is the CPU: AI calls require CPUs directly. Nvidia has Grace and may later have an in-house CPU whose name sounds like Vera, but large-scale training and inference may still require an x86 chip—news may be announced on March 16, which “would be a major help to Intel.” The speaker also acknowledged a rethink on Intel, having previously believed its results this year would be difficult to achieve. His conclusion: “If we are weighing the upside and downside, I think the risks are relatively asymmetric right now.”