In the movie Margin Call, Jeremy Irons plays an investment bank CEO on the eve of the 2008 financial crisis who insists his only job is to figure out where the market is headed next. When he resolves to dump the firm’s mortgage-backed securities, he bangs the table and yells, “This is it! I’m telling you, THIS IS IT!”

I started my career as a geopolitical analyst in 2010 at George Friedman’s Stratfor, where my first assignment as a lowly intern was a red-team analysis of what would happen if Iran closed the Strait of Hormuz in response to a US/Israeli attack. The closure would have been it. Yet we never got to the point where George called us into the office at 3 am and banged on a table, precisely because geopolitical analysis demonstrated that Iran’s asymmetric leverage in blocking passage of oil through the Strait was sufficient to ward off any incursion.

If you could go back and tell that intern the Strait of Hormuz had now been closed for three months and counting, I’d have pictured a world out of Mad Max: Fury Road—not one with Brent crude near $90 and US markets at record highs. Even if Hormuz reopens on July 1, the world has lost 2 billion barrels of production. And markets couldn’t seem to care less.

The Situation 

There are two narratives competing right now for what is happening in the global economy. The first is a bullish growth story driven by AI, the most significant technological revolution since the internet.

The top-performing market in the world YTD is South Korea, up 112%, and that growth is underpinned by AI-related semiconductors. Many other sectors in South Korea are struggling, especially with their exposure to higher energy prices due to the closure of the Strait of Hormuz. To wit: South Korea’s “market” growth is driven primarily by Samsung and SK Hynix.

TSMC, the Taiwanese company with a monopoly on production of cutting-edge AI chips, has doubled over the last 12 months.

China is mired in a real estate downturn and has failed to increase domestic consumption; meanwhile, consumer prices are starting to rise there as well. But China has seen a significant boost in hardware exports tied to AI with integrated circuit exports doubling year-on-year and exports of automated data processing machines and parts up almost 50% over the same period. Moreover, while Chinese consumption continues to disappoint, demand for AI compute in China has exploded, increasing by 1,000X (not a typo) since the start of 2024. Toto—yes, the bidet manufacturer—saw its stock price surge after a report that suggested the ceramics used in its semiconductor components could make it “the most undervalued and overlooked AI memory beneficiary.” Even toilets are part of the global AI growth story.

At the national level, AI is also driving the US economy. Corporate investment in AI drove roughly one-third of US GDP growth in 2025. Google, Amazon, Microsoft, and Meta will spend roughly $725 billion on capex in 2026, up 77% year-on-year; S&P 500 blended earnings growth rose to 27.1% in the first quarter, with Alphabet, Amazon, and Meta as the largest contributors to that figure.

Per Craig Fuller at FreightWaves, AI data center investment is now running at roughly $20 billion every two weeks, against $20 billion per year (inflation-adjusted) during the construction of the Interstate Highway System. (A single 500-megawatt data center requires about 30,000 truckloads of concrete, structural steel, and copper.) Freight volumes are up roughly 11% year-on-year in 2026, ISM Manufacturing PMI was back in expansion at 52.7 in March, and capital goods now account for a record 41% of all US imports. On the power side, cumulative electric and gas utility capex by the 47 largest US investor-owned utilities is projected to surpass $1 trillion over 2025–2029, with 2025 capex alone up 22% year-on-year to $212 billion. And AI-driven data center power demand is forecast to grow more than thirtyfold by 2035, from 4 GW in 2024 to 123 GW.

The second narrative is a bearish stagflation story. Even before the war started, the economy was K-shaped: Nearly half of consumer spending comes from the top 10% of US households. The combination of tariffs and the closure of the Strait of Hormuz is conspiring to increase prices. Inflation is accelerating. April 2026 CPI came in at 3.8% year-on-year, the highest reading since May 2023, with energy up 17.9% and food up 3.2%. The personal savings rate has collapsed to 2.6%, the lowest since July 2007 and half what it was when President Trump took office.

Real wages are not keeping pace with inflation: From April 2025 to April 2026, wages grew 0.24 percentage points slower than inflation. Nominal wages, the literal dollars earned regardless of cost of living, increased by 3.6% while inflation stood at 3.8%. Also, US real wages have resumed a sharp downward trajectory since March.

The so-called AI-driven growth story? Last week, the Commerce Department lowered its Q1 GDP growth rate from 2% to 1.6%. Chevron CEO Mike Wirth went on Bloomberg TV last Friday and said shortages in the US could be “weeks” away. Inventories are tight, demand is strong, prices are elevated, and there’s risk they go higher. At the global level, the closure of Hormuz means structurally higher energy and food prices through at least the end of this year—and likely into the next.

Geopolitics Must Be in Your Toolbox

The US war in Iran has demonstrated that the Strait of Hormuz is not a geopolitical Sword of Damocles. It did not bring the global economy to a halt; it didn’t even send Brent crude above $100 a barrel three months into its closure. Major economies are much less dependent on oil than they were in the 1970s: Oil intensity of GDP has declined 70% since the 1973 Middle East energy crisis.

The rise of renewables and other forms of energy, and the increased efficiency of power-consuming machinery, support this ongoing trend. Saudi Arabia and the UAE have already announced plans to build new pipelines to get around the Strait of Hormuz, which means the importance of the chokepoint will continue to decline in the coming years. In about three years, we will see Iran, the UAE, Saudi Arabia, Oman, and other Middle Eastern countries compete for the easiest and most reliable trade route rather than the open hostility that has defined this year. Despite what the CEO of Chevron said, there are reasons to believe that oil prices may continue to be range-bound even if the US and Iran play “Deal, No Deal” for another two-plus months.

The deeper problem is that the expectation of disaster has obscured the corrosive effect of inflation, which takes place over a long period. Prices were already creeping higher before the war began: The war accelerated and steepened their rise, adding energy as a significant upward force on the overall price consumers are facing. What happens when the AI growth story begins to falter? Will the Magnificent Seven increase their capex another 77% in 2027? Even if they spend at similar 2026 levels, the rate of change will have gone down, and the steroid-like boost to overall growth figures will start to diminish. Meanwhile, inflation will keep eating into the appetite of the consumer, who has essentially been YOLO-ing since the realization that governments overreacted to the COVID-19 pandemic.

The AI growth story has clearly defeated the geopolitical risk story thus far in 2026. There is a cautionary tale here: While it’s self-serving to argue that understanding geopolitics is critical to making decisions about markets now and in the future, it is no less true. But geopolitical analysis is just one tool in the toolbox. If you try to build a house with nothing but a hammer, you will fail. In this case, only using the hammer suggested that one should expect $200 crude within six weeks of Hormuz closing. It even told you that the US wouldn’t attack Iran because of the potentially deleterious economic impact. (Full disclosure, that was my position in January. It was wrong.)

But don’t throw the hammer out with the bathwater. Geopolitics must be in the toolbox, and its import is especially pronounced over longer time horizons. Other factors have been more important than geopolitics in 2026; based on how things have gone, that looks poised to continue, irrespective of the war and whenever it ends. But eventually, the push and pull story between AI-driven growth and real economic constraints will shift. Inflation will continue to eat away at the purchasing power of the consumer. Confidence when lost collapses quickly.

In the meantime, as we say in New Orleans, laissez les bons temps rouler.

Map/Chart of the Week

Blind Spot

In 1965, Intel co-founder Gordon Moore wrote a paper entitled “Cramming More Components onto Integrated Circuits.” Moore predicted a doubling of transistors every year for the next 10 years. You don’t need to be a tech maven to understand the importance of transistors; they are one of the basic building blocks of modern electronics, controlling the flow of electrical signals within microchips. I.e., they are very important. So important that Moore’s observation became known as “Moore’s Law” and has signaled success in the semiconductor industry ever since. Intel remained the dominant chip manufacturer for decades, but around 2018 it ceded process leadership to TSMC on manufacturing execution. While Intel’s 10nm node slipped repeatedly into 2019, TSMC was shipping its leading-edge node in volume with high yields.

Last week, an employee at Chinese company Huawei delivered a speech, “New Semiconductor Path in Practice,” in which she presented the “Tau Scaling Law,” a “new principle for guiding the future development of the semiconductor industry.” She proposed that the physical size of transistors is not as important as how fast computing tasks get done—and that Huawei can use design, architecture, and other new techniques to improve performance without shrinking the size of transistors.

Here’s what’s important to understand: Huawei’s strategy is born of necessity. During his first administration, President Trump determined Huawei’s lead in 5G infrastructure (not AI) posed a national security threat to the US. He attempted to kneecap Huawei and other Chinese companies with US tech export restrictions that blocked Huawei’s access to ultraviolet EUV lithography equipment, the machinery without which building chips with smaller and smaller transistors is impossible. (Dutch manufacturer ASML has a monopoly on that equipment, and while it has chafed against US restrictions, it appears to largely have followed through.) Ultraviolet EUV lithography equipment is incredibly hard to develop. It’s one of the few technologies out there where the question isn’t “how long will it take China to duplicate” but rather “can China duplicate this technology in an amount of time that would make it relevant.”

For years, experts have been saying Moore’s Law is effectively dead (Nvidia CEO Jensen Huang made the assertion in 2022). There have also been previous reports of technologies that might break ASML’s monopoly over EUV lithography equipment. Perhaps so… but in the meantime, the most advanced chips made today—the ones critical to the global AI infrastructure buildout—still operate within the old paradigm.

When asked about Huawei’s announcement, Huang described the development as a breakthrough for Huawei but hardly a threat to TSMC, which he claimed had already been pioneering related technologies and capabilities for years. I can’t offer a definitive answer. What I can note is that the US’s attempt to destroy Huawei not only failed but made Huawei a more innovative company, one that is challenging what we know and understand about some of the most important hardware underpinning the entire AI story. 

Reader Question

Finally…

What I’m watching: NBA Playoffs

What I’m reading: Annals of the Former World, John McPhee

What I’m listening to: This One’s for You, Luke Combs