Phone alerts still come now and then. On August 18 mine buzzed because missiles were “directed towards the country.” A few hours later the Ministry of Defense confirmed two ballistic missiles from Iran, one falling inside territorial waters, one into the sea beyond. By the next morning the UAE had suspended all trade and financial dealings with Tehran, and I was back in at home in Sharjah talking to companies about geopolitical risk and nonmarket strategy.

And yet the day-to-day here is boringly, almost defiantly normal (which I say as a good thing considering how long I’ve lived here). The malls are full. The conferences are mostly back up and running. The beaches on the Palm refilled within weeks of the first barrage (and then cleared out again as the summer heat ratcheted up).

If you want to understand what this war has actually done to the UAE and what it might still do (as opposed to sensationalist headlines), let’s start with how things got back to feeling “normal” so fast and then try see where things are likely to go.

Subscribe now

The oil pivot

For two decades the UAE has said it would wean itself off of oil dependent. Most people outside the UAE filed that under aspirational Gulf branding, somewhere between an indoor ski slope and islands built to look like the world. Then the war happened and we saw how far the UAE had actually come.

Non-oil activity hit 79.4 percent of GDP in the first quarter of 2026, up from 78 percent a year earlier. Non-oil exports jumped almost 24 percent in the first half of the year, to $123 billion. The IMF still expects the government to run a surplus(!) in 2026 with very low government debt. Foreign investment barely flinched: a senior UAE minister put the share of unaffected flows at 98 percent, and the country kept adding companies through the worst of the Iranian attacks.

Then on May 1, the UAE walked out of OPEC after six decades so it could pump on its own terms and funnel the proceeds into everything that is not oil. You do not make a move like that from a position of weakness. The longterm plan of weaning off of oil was supposed to pay off sometime in the 2030s. The war shifted the timeline forward and also proved the position correct. When roughly four-fifths of your economy has nothing to do with the barrel, an oil-price spike and a hit to your energy infrastructure is a bad quarter, not an existential crisis.

Let me be fair about what this does and does not mean. The Washington Post ran a sharp piece arguing the war made the move-beyond-oil project harder, not easier, because billions now have to go to rebuilding damaged facilities and buying interceptors instead of AI and tourism. I think that’s a fair point. But “we have to spend on defense and repairs” is a temporary diversion because this transition has been happening for more than a decade. And it’s a very different problem from the one people were forecasting in March.

Subscribe now

Little Sparta earned the name (again)

James Mattis reportedly coined “Little Sparta” years ago for a small country that invests in its military and is willing to use it. As Spartan by blood who lives in the UAE, I love the nickname. And the most kinetic phase of this war showed why the label came into being and what all that work behind the scenes was for.

The UAE has one of the best missile defense systems on the planet. Not in theory, but in battle-tested practice. Barrage after barrage, the interception rates ran high: three of four one day, eight of nine another, twenty-six of thirty-five drones on a “bad” night. I don’t want to minimize the real impacts in lives and physical damage, but the Iranian regime bombarded the country and it absorbed the heaviest sustained aerial assault in the region while holding casualties to single and low double digits. That is what modern defense preparation looks like in practice.

Preparation is also about diplomacy paired with that strength. The UAE’s position is strong because it still had relations with Iran that it could use as leverage. A country can only weigh up asset freezes and vessel seizures if there are assets and vessels to seize. When the August missiles came, the answer was not a statement of concern. It was cutting Iran off completely. Tehran now knows that hitting the UAE is expensive, and that the UAE can make it more expensive still. That is what keeps the malls full. (It’s also what increases the risks for other, less prepared states, but that is a post for another day).

Subscribe now

Companies aren’t leaving because where would they go?

Every crisis in the UAE brings the “exodus” story from those outside the country and this one did too. In the first week Nvidia and Amazon shut their Dubai offices, some Google staff got stranded after a sales kickoff, and wealth managers in Singapore started fielding nervous calls. If you only read the March news, you would think the foreign companies were packing up and ready to head home.

They were not. The finance sector saw no meaningful departures. DIFC added 775 companies in the first quarter, up 62 percent on the year. Canva, Century 21, and Novo Nordisk all set up regional bases during the war. Chinese firms in drones, payments, and flying cars are using Dubai as their springboard into the region and beyond. Henley’s verdict on the year was blunt: diversification and optionality were the prevailing themes, not an exodus.

The reason is simple, and a little deflating for the doom-and-gloom crowd. Ask the executive who threatens to leave and you’ll get largely the same response: leave for where? Riyadh is competing hard and will win some of this business but faces most of the same risks with potentially greater exposure and most of the Gulf sits under the same threat umbrella. Singapore is a long way from the customers that UAE offices service. The entire point of being in the UAE is the access to markets you reach from it, a couple of billion people across the Gulf, South Asia, and East Africa within a short flight, wrapped in fast licensing, no personal income tax, and courts built to reassure foreign capital. You do not give up that access because of attacks that were largely repelled from a threat that was largely known pre-crisis. A hedge, maybe. A second passport, sure. But nowhere close to a full-scale relocation of company offices and investments.

Subscribe now

The obvious soft spot: real estate

So the elephant in the room that a lot of economics reporting has already pointed out: Dubai property is down since the start of the conflict. Prices slipped up to 7 percent per month in the February-to-April window. Apartments and villas each fell about 3 percent in the second quarter. Transaction values ran roughly 16 percent below last year’s record.

But look closer and it’s obvious that this has very little to do with the conflict. Fitch noted that they expected a 15 percent correction running through late 2026 back in the spring of 2025. UBS had flagged Dubai as one of the most bubble-prone property markets in the world in September 2025. Developers were on track to hand over around 110,000 units this year against a ten-year average of 27,000 or roughly triple the average annual rate. That is a textbook oversupply bubble, and the slowdown had already started before the first alert went off.

Prices per square foot are still higher than they were a year ago. The froth came off sure but the market did not “collapse”. At the worst of it, the Dubai real estate stock index declined 34 percent while actual home prices only moved 4 to 7%. Anarock called that the widest gap between sentiment and asset value in any Dubai crisis on record.

That the conflict coincided with what was already predicted to be the end of a real estate cycle does not say much at all about the conflict’s impact on the UAE. It says a lot more about the cyclical nature of real estate than anything else and how those cycles can coincide with other macro factors in good and bad ways. As I’ll explain below, it ended up being a rather fortunate coincidence.

Subscribe now

The real: grocery bills and the expat family

Here is the risk I would actually watch over the next 6-9 months for companies with offices in the UAE. It has less to do with missiles and everything to do with cost of living.

The Strait of Hormuz has been effectively closed to commercial shipping since March. For roughly five months the Gulf has been flying in and trucking around what it used to float in, and that cost lands in the checkout aisle. Add to that the dramatically increased cost of diesel and the fact that around 80 percent of what the region eats is imported and you get a recipe for dramatic, sustained price increases. Container rates from China to Jebel Ali jumped to several times their pre-war level, insurance and port congestion stacked on top of that, and diesel and petrol spiked as much as 70 percent in a single month. Dubai’s consumer price index ran 5.3 percent higher year on year in July alone, easing from a 5.7 percent peak in June but still the hottest reading in the GCC, with transport alone up almost 12 percent. Set that against the United States near 4 percent and most of Europe lower still, and the UAE is experiencing a bigger cost-of-living jump than much of the developed world.

So the real extant question is what do expat families do in response. The odd relief valve is rent: with all that oversupply, leases are actually getting cheaper, down a couple of percent this year, and people are moving to lower cost units. But the moving itself is also revealing. Anecdotally, I am hearing about families trading a Downtown lease for a cheaper unit out in Jumeirah Village Circle, or crossing the border into Sharjah where both the rent and the schools cost less. Parents shopping the list of schools under Dh50,000 a year. Cars sold, takeout purchases cut, the extra flight home swapped out for a staycation. None of that signifies panic or a mass exodus but, over the next six to nine months, it could lead to return decisions that wouldn’t have happened otherwise.

And here is why it matters more in the UAE than the weakness of the real estate sector. When four-fifths of your economy is non-oil, the economy is more dependent upon (primarily domestic) consumer spending. Retail, restaurants, schools, clinics, leisure, the entire consumer-facing machine that the diversification away from oil is built upon depends on the median household feeling flush enough with cash to spend. Pay is not keeping up (and again anecdotally it doesn’t seem to be for most employees) because the global economy is tight already and firms are already absorbing higher input costs on softer demand.

Squeeze the household from both ends, higher prices and flat wages, and the belt-tightening does not stay in one household. It surfaces as the empty table on a Thursday night, the membership nobody renews, the unit that sits unsold, and it ripples straight through the sectors the UAE has staked its future on. This all functions as a long, grinding tax on the cost of living and it could quietly cool the consumer engine underneath the whole non-oil model.

Subscribe now

So what is the takeaway?

Six months in, the picture is clear. This is not a country hanging on by its fingernails. It spent twenty years building for exactly this type of shock and it largely worked. The economy proved it can run without the Strait open. Little Sparta proved it can take a punch and throw one. The businesses proved they had nowhere better to be. The property wobble turned out to be a bubble deflating on schedule.

The real trouble over the next six to nine months is quieter and slower than any of the kinetic attacks. It is the household budget, and whether a prolonged squeeze on the cost of living drags down consumer spending that the entire non-oil model requires.

The missiles make the headlines, but the grocery bill is actually what matters. That is where the next chapter of this story gets written and what companies to look at for their current Gulf risk exposure.

On the bright side for companies this also means that the solution for their own risks sits on their balance sheet in the form of adjusting wages and benefits to meet the increased cost demands. But that, as always, depends on how prepared the company was in the first place with sufficient balance sheet protection to make such changes.

As always, easier said than done for most companies.

My answer frustrates both of them (it’s an annoying habit of mine): What do you mean by “work”?

The Bud Light boycott “worked” in a way. It did real damage. Anheuser-Busch InBev’s US revenue fell by roughly $1.4 billion in 2023, its market value fell by about $26 billion in the weeks that followed, and it cut roughly 400 North American corporate jobs. Bud Light lost a twenty-year hold on first place and, by early 2025, was still roughly 40 percent below its pre-boycott sales level, according to former AB InBev executive Anson Frericks. Conservative drinkers said they would leave, and a lot of them did. Many never came back.

And AB InBev absorbed it, though not painlessly. US revenue fell 9.5 percent in 2023, shipments fell 12.7 percent, and a year later the company was still clawing back market share in increments of 10 to 20 basis points every few weeks. In September 2025, Michelob Ultra became the best-selling beer in America.

Michelob Ultra is owned by….AB InBev.

Boycotts can do real financial damage to companies. What they much less reliably do is change the behavior the boycott was called over, which is usually the whole point of doing the financial damage. That gap between hurting a company and moving its decisions is where most of these campaigns live or die. And most of them die.

Subscribe now

Four conditions for a boycott to achieve a political objective

For a consumer boycott to change what a company does, four conditions usually have to be met.

First, there has to be a substitute the consumer genuinely prefers, or at least tolerates. Tomlin’s event study of 125 US boycotts between 1978 and 2017 found that boycott announcements had a statistically significant negative effect on the shareholder wealth of targeted firms, with larger effects where competition was heavier. If there is nowhere else to go, the threat is a lot less threatening.

Second, the company has to be legally and structurally able to do what you are asking. Sometimes it is. Sometimes it isn’t. Sometimes the legal obstacle is real, and sometimes the real obstacle is that the company is not actually the entity you think you are boycotting. I’ll return to that below.

Third, the demand has to be something the company can actually do. Demanding that a company extricate itself from the global economy is not something a multinational can do. Or demanding that an oil company stop using oil. A boycott with no obvious company-based solution is not much of a negotiating tactic. It is just a permanent reputation tax, and companies eventually learn to live with permanent taxes or all kinds.

Fourth, somebody other than the boycotters has to care about the outcome. King’s study of boycotted firms between 1990 and 2005 found that media attention and reputational pressure were important predictors of whether companies conceded to boycotters’ demands. Sales declines, by contrast, were not a statistically significant predictor of concession.

The most important thing may not be whether the company is losing money in other words, but whether everyone - investors, media, home governments, uninterested customers - can see that the company is “losing”.

Subscribe now

Political boycotts often meet the first condition and the fourth. They are more likely to run into trouble with the second and third. So which boycotts work given these conditions? The ones that meet the four criteria above tend to be narrow in scope with concrete demands that can be made by a single decision maker where the campaign can be undone swiftly.

Cracker Barrel for instance reversed a new logo in eight days last August, after its market value fell sharply amid a backlash. The episode did not end with the CEO being fired immediately, as some later retellings imply. Julie Masino is only now stepping down, nearly a year later.

What about when they don’t have their intended outcome? An interesting case is from the Middle East where I live and work.

Since October 2023, consumers across the Middle East have boycotted primarily American food and beverage brands over their perceived connections to Israel. For some brands it hit hard. Alshaya, the Kuwaiti group that operates Starbucks across the Middle East and North Africa, cut roughly 2,000 jobs in March 2024, more than 10 percent of its Starbucks workforce. Americana Restaurants, which operates KFC and Pizza Hut across the region, saw revenue fall 9 percent and profit fall 39 percent in 2024.

Alshaya is Kuwaiti. Americana is listed in Abu Dhabi and Riyadh. Neither of them do any business with Israeli firms. The businesses employ people locally and operate largely regional supply chains. The logo of one company is franchised by other local firms in other nations with no other real relationship between them - these are entirely separate companies operationally and in terms of ownership that happen to share branding on a slice of their products and services.

In Starbucks’ case, the Middle Eastern business is operated by a local partner rather than being a chain of company-owned American stores. Starbucks itself does not have stores in Israel but licenses its branding to a separate Israeli company.

So the company losing the sale in a Gulf shopping mall is not having any impact on an American shareholder or an Israeli franchisor, and the immediate economic hit is instead on a company and a workforce in the same region as the person doing the boycotting perhaps even with the same political commitments.

The logo above the door is the only connection between the boycott target and the boycott impact.

An additional complexity is legal, and it takes me back to the second condition above. All those franchise agreements are legal enforceable contracts that pre-date the boycott. They can’t be discarded without legal repercussions. And US companies are further subject to antiboycott rules that can make certain forms of cooperation with foreign boycotts unlawful. A company that participates in a prohibited boycott can face substantial civil penalties, criminal exposure and loss of certain tax benefits. It also has to report certain boycott requests to the Commerce Department, generally on a quarterly basis.

Sometimes then it’s not that companies “won’t” do what the boycott asks; it’s that they cannot legally do what they’re being asked.

Subscribe now

A connected problem: depreciation of the boycott “asset”

In the language of business, a boycott is a depreciating “asset”.

It has a peak value, a schedule, and a salvage value. It is worth most at the beginning, before a single customer has stopped buying, because that is the moment when neither side knows what the threat is actually valued at. The organizers do not know how many people will stop buying. The company does not know either. Nobody knows how serious people are on this issue, how deep the brand defections will run, how well the campaign will be marketed, whether it will stay regional or go global, or how long any of it will last.

That uncertainty matters.

An executive looking at a large, potentially unbounded revenue risk will want to do a great deal to make it disappear. It is one reason concessions, when they happen, often happen quickly.

But then the data starts to arrive about the impacts of the boycott. Scanner data. Foot traffic. The next earnings call. Each one converts a little uncertainty into hard numbers that are a record of what people actually did rather than what they said they were going to do.

By month three, the company knows what the boycott costs. Its analysts know. Its investors know. Management can start planning around the new number. Nobody had to stop the boycott. Some of the uncertainty simply disappeared and what the company sees at that point maybe something that it largely shrugs at because it sees that the impact is marginal at best and recoverable at worst.

To quote Rasheed Wallace: “ball don’t lie.”

Subscribe now

That is why there is a curious asymmetry here. The boycott is most useful as a threat before it’s proven exactly what its impact will be.

Delacote’s work shows that the consumers most willing to boycott tend to be the ones who consume the least, while the heavy users whose absence would show up most clearly in the numbers are less willing to give the product up. And there is a gap between what people say they will do and what they actually do. 53 percent of Americans say they would avoid a company whose stance they dislike; only 30 percent do it.

So here is the paradox.

The boycotters’ leverage is the threat to leave. But the only way to demonstrate that the threat is real is to leave. Once they do, they have converted an uncertain threat into a measurable loss which can then be managed, recovered, and built upon. Renewing it is then almost impossible because the threat will not be taken seriously.

A strike is renewable. Workers go back to work and can strike again. A boycott does not have the same reset potential. Once customers have made the switch and a company has learned the cost of the boycott, the threat is harder to recreate in exactly the same form.

Every customer who stayed has also now demonstrated that this issue was not enough to move them. Every customer who left but returned has demonstrated that there weren’t real substitutes that worked for them long-term.

The company comes out of the experience with a (maybe) smaller but (definitely) better-understood customer base.

King’s numbers make more sense once you look at the boycott in the paradox context. Reputational exposure can move a company because reputation is still uncertain and still in play. Realized sales losses are different: by the time everybody knows the number, the number is already part of the company’s model.

And the reversion data post-boycott is not exactly encouraging for boycotters.

Chavis and Leslie found that the 2003 boycott of French wine cut sales by 13 percent over the six-month period they studied and then had reverted within a year. Amazon’s transactions on the day of the February 2025 “economic blackout” came in 1 percent above the average of the previous eight Fridays. Americana went from a 39 percent profit decline in 2024 to FY2025 profit up 38 percent, on 9.7 percent like-for-like growth. Then came first-half 2026 profit up 59 percent. The company expects to open as many as 130 more restaurants this year.

That is not the same thing as proving the boycott was irrelevant. But it is a reminder that a large hit in one period is not the same thing as permanently changing the business and its decisions.

Subscribe now

The company decision: concede, ride the wave, or push back

If you are running a company under boycott, you have three choices.

Concede. A concession is most valuable while the threat is still unpriced, which often means early. Cracker Barrel reversed in eight days, contained the immediate controversy and got the issue off the company’s front page.

But concession has a price.

It tells future campaigners that the company is vulnerable. It tells employees that management will reverse itself under pressure. It can tell investors that the next political fight may be handled the same way. You can usually only pay that price once.

You should think very carefully before making it a habit.

Ride the wave. Makes sense when the loss is isolated and the boycott demand has no obvious “off” switch, which describes a lot of political boycotts. It stops making sense if the boycott occurs simultaneously with something that changes buying behavior for reasons that have nothing to do with the boycott.

Target is an example. After its DEI rollback in January 2025 it missed sales expectations by nearly half a billion dollars, and Brian Cornell said the backlash, tariffs and weak consumer confidence all played a role and that he could not reliably separate them. That inability to separate them is a warning to companies trying to ride the wave: you better hope nothing else goes wrong.

If you cannot tell which part of the loss belongs to the boycott, this is not a good strategy.

Push back and argue your case in public. This is what the Gulf operators largely did: explain local ownership, local employment, local supply chains and what the boycott actually hits. Americana’s subsequent recovery shows that this can be a viable strategy, although the recovery itself is about more than the public argument alone.

The catch is that a company genuinely aligned with the issue a boycott targets will find that the argument sounds like evasion. Don’t push back in that case. Otherwise you have to make a political argument you don’t actually believe in and your company will be caught out.

Subscribe now

The consumer question: a result or living your values?

Everything above is about the company response.

Turn the tables around now, because most of us have stood on the other side of it in a supermarket aisle. As consumers, we need to stop asking whether boycotts “work”. Instead we need to ask which of two entirely different things you are trying to do. Are you trying to get a desired outcome? Or are you trying to live your values?

In this case, they demand different behaviors, and most people never decide which one they are doing.

Living your values. My maternal grandmother’s family were Greeks driven out of Anatolia in the early twentieth century. The land they left grew a lot of things, apricots and nectarines above all.

I love dried apricots. Most dried apricots in a supermarket come from Türkiye or California. I check the label, and if it says Türkiye, I put it back.

I have no illusions about the impact that this decision has. My grocery habits will never trouble the Turkish dried-fruit industry. The farms are too large and diversified, and the families working that land today had nothing to do with 1922.

I do it because I do not want my money going toward something I find ethically problematic. That is a good enough reason on its own.

It is also one of the main reasons people boycott. Klein, Smith and John found that self-enhancement and the desire to make a difference predict boycott participation, and that even among consumers who regarded a company’s conduct as outrageous, only a minority actually stopped buying.

Boycotting then can be partly a statement about the values of the person making it. There is nothing wrong with that. It is also the only version the boycott that cannot really “fail”, because it does not depend on anyone else showing up or changing company behavior. It comes down to where you believe your money should go.

If that is what you are doing, do it and stop worrying about whether it “works.”

The answer is no. And the answer is irrelevant.

You do you.

Subscribe now

Looking for a result? “Ball don’t lie”. If you want a specific decision or political result, you are doing something that loses value from the day you start doing it. It only works if the demand is built for speed, narrow enough for someone to say yes, legal enough for them to be able to do it, and concrete enough that everyone knows when it has been met.

That is what the campaigns that win tend to have in common.

A boycott of every company that touches another company that touches a country in the global economy cannot do that because then, it’s just a demand to remove yourself from the global economy.

Neither can a boycott that is ill-defined or cannot realistically end.

So decide which one you are doing as a consumer. Then behave accordingly.

This particular driver had a theory about the current situation in Iran: the on-again, off-again cycle of strikes and ceasefires, he insisted, was an attempt by US government officials to take advantage of the resulting market swings.

He is not a market participant himself, other than as a buyer of petrol for his taxi. But he had noticed what anyone paying attention has noticed. Oil surged after the strikes on Iran’s nuclear sites and then dropped more than five percent in a day once markets decided the Strait of Hormuz would stay “open”. The pattern has repeated with each truce announcement since. If you knew the timing of the openings and closings in advance, you could indeed make a fortune. So I told him, in so many words, that his theory was more plausible than I’d like it to be.

But then, as often happens when you speak to strangers, things got…weird. I next heard about how the same “secret hand” behind this profit-taking had also orchestrated the September 11 attacks, pushed Greece’s 2024 same-sex marriage law through parliament, and controlled all media. That last one is among the oldest and most lethal tropes in existence, and it arrived in the conversation with the same casual confidence as the oil conspiracy.

His finger on the pulse of the body politic had become a finger on the pulse of crazy town.

That ride is what I want to write about today. I’ve written before about grand corruption and the risks it poses to societies and their politics, particularly in the US. Today I want to talk about conspiracy theories and risk analysis: why conspiratorial thinking is mostly a hindrance to understanding the world, and why, in one very specific context, analysts ignore it at their own peril.

Subscribe now

The conspiracy right in front of you

Here’s the irony at the heart of conspiracy theories. Very often there is a conspiracy, and the theorists refuse to see the most obvious ones. Worse, once the real conspiracy is uncovered and publicly disclosed, the disclosure itself becomes further evidence of a cover-up.

September 11 is a standard example. It was a conspiracy in the traditional sense: a plot by a terrorist organization to fund, train, and deploy 19 hijackers onto four planes and fly them into buildings with the intent of taking down US financial, government, and military systems. The plot was real, secret, coordinated, and later documented in exhaustive detail. And yet, more than two decades on, one in five Americans says it’s likely the US government was behind the attacks, to say nothing of non-Americans. The actual conspiracy wasn’t satisfying or plausible enough for these folks. Nineteen men with box cutters, directed from training camps in Afghanistan, felt too small and too improbable for an event that large. So people went searching for something more usually with the underlying assumption that it was in fact the US government itself that committed the act: from directed energy weapons (whatever those are) to controlled detonations, amateur sleuths assume that they have the “real” answer to what happened despite loads of evidence to the contrary and the conspirators themselves taking credit for it.

Why do we do this? The research actually pretty extensive on the question. Conspiracy theories serve three critical human needs: epistemic (a clear cause for a chaotic event), existential (a feeling of control and safety), and social (the belief that you and your group see what others can’t). And they meet all three at absolutely zero cost to the believers because they are almost always unfalsifiable. No evidence can prove them wrong, so belief carries no risk of ever being embarrassed.

My taxi driver’s “secret hand” explained everything to him from Iran to oil markets, from family law to the press. It made him, in his own eyes, one of the few who really understood what was going on in the chaotic world around him. It cost him nothing to hold but gave him the psychological satisfaction of understanding a chaotic world.

If only the world was that simple.

Subscribe now

Conspiracy theories as counter-narratives

Conspiracy theories don’t grow randomly though. They grow against whatever the established narrative is, and the established narrative is usually set by governments, elites and the media outlets align with them. This is why you almost never see a major outlet run a conspiracy theory against its own audience’s perceived common sense. Conspiracy theories must point against narratives established by others.

Take two US examples, both of which were falsifiable to varying degrees. We saw claims of a stolen 2020 election circulate through right-leaning media, where half of Republicans still believe them, and not through left-leaning media. More than sixty cases were filed; every one decided on the merits failed, many before Trump-appointed judges.

On the other side, the most breathless versions of the Trump-Russia collusion story ran through left-leaning outlets, a body of coverage the Columbia Journalism Review later dissected at book length. The evidence of Russian interference in the election was extensive and documented. But the Mueller investigation did not establish that the Trump campaign coordinated with the Russian government, and outlets ended up revisiting sourcing they had leaned on hard. That retreat then spawned a counter-conspiracy (a counter-counter-narrative?) from the right: that the whole thing was a “Russia hoax“ and no Russian interference happened at all. Conspiracy theorists are nothing if not creative.

Each of these theories, and my taxi driver’s, is a counter-narrative to whatever the theorist believes is the standard story. All of us like to believe the ones that indict our opponents’ values and beliefs, and to dismiss the ones that contradict our own. They are, in a way, one of the means people use to explain why the observable evidence refutes their own beliefs about the world.

But that counter-narrative structure also gives conspiracy theories a real use in identifying political risk in a very specific context. If they are a standing check on official explanations and the officials are broadly corrupt in a real and obvious sense, conspiracy theories are often the only competing explanation available to officials explanations that make any sense.

Subscribe now

Just because you’re paranoid doesn’t mean they aren’t after you

The counter-narrative check matters most in this one type of political setting: when a society is beset by grand corruption, the abuse of high-level power that benefits the few at the expense of the many.

The research backs this up, and it is unsettling when paired with my previous warnings about grand corruption in the US. One study puts it about as plainly as one can: people suspect foul play when foul play is objectively there. Conspiracy beliefs as a result run higher in more corrupt countries, and public sector corruption is fertile ground for them across Western and non-Western societies alike.

This makes sense when you think about what grand corruption means for official explanations. A government awards a port concession to an unqualified bidder, or blocks a cheap medicine, or fights a war longer than any strategic logic supports. The official narrative has to explain all of that behavior without mentioning the real reason, because the real reason is evidence of a crime. (As a small aside, I often hear that corruption is “okay” in a particular place because it is so widespread; in exactly none of those places is corruption legal or publicly socially accepted.) So the official story about a major event in places with grand corruption is almost guaranteed to be false. Conspiracy theories then compete as a narrative on merit because they can take the grand corruption into account in ways that the official narrative cannot.

This is the narrow sense in which conspiracy thinking helps an analyst or company executive in looking at what is really happening and what the risks of operations really are in a place with grand corruption. Behavior that looks irrational on its face often becomes rational the moment you factoring the grand corruption.

In place without grad corruption, I find that the irrational actions can be explained by the fact that, well, humans are irrational and incompetent despite themselves. These remain the right default hypothesis in places without grand corruption. But in an environment where grand corruption is documented and widespread, that default should flip: when a policy keeps failing and keeps being renewed, follow the money.

Subscribe now

Disciplining our minds for counter-narratives

Here’s the rub. “Follow the money” is a pathway, but conspiracy theorists often treat it as a verdict. Simply finding someone who benefits from an action does not actually explain why something is happening. The world is filled with opportunists; that doesn’t mean those same opportunists manipulated the system to have the opportunity available. That is the entire difference between analysis and paranoia.

If my driver’s oil theory is a hypothesis, it would need to be tested. What trades would we expect to see before each strike and each ceasefire? Did any market participants actually make those trades? Whose disclosures should show those trades? What would distinguish insiders trading on advance knowledge from ordinary speculators reacting fast? Those are (largely) checkable questions, and investigating them is how you turn a suspicion into something approximating a real answer.

The test I’d offer is the one my driver failed. Call it the taxi driver test: if your theory explains oil prices, Greek marriage law, and the global media with the same mechanism, it explains none of them. A real hypothesis is expensive because it needs testing. It is also expensive because it means you can be wrong. It commits you to specific actors, specific payments, specific evidence that could prove you wrong. That’s why I try to be up front about my hypotheses about the world: so that you and I have a record of how I thought about things then and we can compare them to the evidence of what happens in the world over time.

A conspiracy theory on the other is free, which is exactly why it buys nothing most of the time.

Where do you draw the line between healthy suspicion of official narratives and conspiracy thinking? I’d like to hear it, especially from readers living in places where the official narrative has earned the suspicion.

The research says the European side has this one right. The Helsinki Businessmen Study followed executives for 40 years and found that executives who took three weeks of vacation or less died earlier than those who took more, even when they exercised and ate well at the same rates. Rest turned out to be the critical factor. I write about resilience for a living. It would be strange to skip the cheapest form of it. Most Americans do: fewer than half use all their vacation days.

Enjoy the piece below and see you in two weeks.

Subscribe now

Someone at a corporate event recently asked me how a normal person, meaning not an analyst, not someone who does this for a living, is supposed to keep up with the news when all of it is biased. The question was sincere. People want a trusted source. They want to point at one outlet and say: that one!

There isn’t one. I have spent about twenty years assessing how the same event reads from Washington, Beijing, Athens, and the Gulf, first as a researcher in conflict zones and now advising companies on political risk. If an unbiased outlet existed, I haven’t found it (and I’ve looked as it would have saved me a lot of reading). The search for one is a fool’s errand.

Subscribe now

The framing problem

Start with the vocabulary. Misinformation, disinformation, fake news. They’re not using the word “lies” - and people do lie, really, they do - because most of this doesn’t actually fall into the category of a lie. Much of what gets these labels though is real: the events actually happened (see my note on AI at the end of the piece though). What the “flaggers” are flagging is the context: a true fact pulled out of one narrative and dropped into another. Whoever yells “misinformation!” usually disagrees with the framing rather than the facts. People treat the terms above as technical categories, something a platform can flag and delete, something you defeat by finding the right filter you trust.

No fact-checking layer fixes that, because there is nothing false to catch. The dispute is over the story the fact got placed inside, and stories don’t fail fact-checks.

So change the goal. You want a set of sources whose biases you can name, each clearing a single low bar: it reports things that actually happened. That bar is lower than people expect and it is enough. Once you know which way a source leans, you can read it, subtract the lean, and keep the facts. A bias you can see has already lost most of its power over you.

Subscribe now

Naming bias

In America this means reading across the domestic divide, yes. The more useful move is reading foreign outlets in English. A paper in Hong Kong or Athens or Dublin generally has no stake in the American red-blue fight. It carries its own domestic concerns and its own commercial interests instead. The South China Morning Post for instance has been owned by Alibaba since 2016, and its editorial line reflects that. I read it anyway, because I know that going in, and because Hong Kong’s establishment view of a story is a data point no American outlet will hand me. A different bias is actually what I want (and you should too). It shows you the parts of a story your home outlets skip because those parts fit neither side of a fight you already know.

This is usually the part people don’t like. Reading this way forces you to notice your own biases, and that is uncomfortable. It is easier to read three outlets that agree with you and call yourself informed. Deliberately reading an outlet that sees the world in a way you find annoying/wrong/morally debased/whatever is the whole exercise.

The research backs this up. Psychologists call it inoculation: expose people to a weakened dose of a persuasion technique, teach them to name it, and they get measurably better at telling reliable from unreliable news without sliding into reflexive cynicism about everything. A large study found the effect holds across age, education, and political ideology with two caveats: the effect is moderate, and it fades if you stop practicing it. Which is exactly why this needs to be a habit rather than a one-time media literacy seminar.

Even with all of this, you will sometimes get it wrong. I have. In November 2022 a missile killed two people in Przewodów, Poland, and the AP reported, citing a senior US intelligence official, that Russian missiles had crossed into NATO territory. Every outlet in my rotation carried it, because everyone downstream of the wire did. For a few hours the accepted fact was that Russia had struck a NATO member. Then Polish and NATO investigators concluded it was most likely a Ukrainian air-defense missile that fell while intercepting a Russian barrage, and AP retracted. I believed the first version, and so did my carefully assembled, bias-mapped reading list, every source of it. Agreement across your sources lowers the odds of being wrong. Nothing gets those odds to zero. The honest thing is that, when the correction comes, to update in public, since pretending you always knew is its own kind of bias.

Subscribe now

My Method (but you probably need your own for what you care about)

None of this requires more hours. It requires structure. Here is how I stagger my own intake for the things I care about. I will say that I have intentionally left out sports news because I doubt most of you care about the New York Rangers.

Daily, for what happened. This tier is about immediacy. I want to know what moved, from enough angles that no single narrative sets the terms. The FT and WSJ for markets and the business read, which is its own bias and a useful one for me since I care about geopolitical risk’s impact on companies. The New York Times for the American center-left institutional line. Drudge Report because it remains the fastest headline instinct on the internet, and because it is a live lesson in why you re-check your sources: it spent two decades as the house aggregator of the American right, turned on Trump around 2019, and media-bias trackers now rate it lean left. My reasons for reading it survived its flip. The characterization I had of it did not. South China Morning Post for the view from Hong Kong. Kathimerini for Greece and the eastern Mediterranean. BBC for the wire-level baseline. The Irish Times because a small, sharp country with no dog in the American fight often covers the world more plainly than the giants (plus, it’s in English). Semafor because its house format separates the facts from the reporter’s view and then tells you where the disagreement is, which does some of the subtraction for you. When I need the minute-by-minute, Twitter and Reddit, with every caveat those two deserve stapled on.

The point of the daily tier is watching the same event refract through nine or ten sets of interests and noticing where they diverge. The divergence is the information in addition to the base level fact of what happened.

Weekly, for what it means. Once I know what happened, I want people who have had a few days to think, and I want them arguing. The Economist for the globalist-liberal establishment read. The Atlantic for the American intellectual center-left. Der Spiegel’s international edition for the German and continental view. The New Statesman and the Spectator as a deliberate pair, left and right of British commentary, read against each other on purpose. The Diplomat for Asia-Pacific coverage that Western weeklies treat as a sidebar and it treats as the main event.

Monthly and quarterly, for whether any of it matters. These I read across the whole period rather than in one sitting because they’re too long. Foreign Affairs and Foreign Policy for the American foreign-policy establishment talking to itself, worth hearing precisely because it is the house view. Le Grand Continent for the European strategic conversation, which I translate from the French myself, because the slowness of translating forces me to actually absorb the argument. Asia Policy from the National Bureau of Asian Research for rigorous regional analysis. And a rotating stack of academic journals, which correct the distortion a year of headlines quietly builds. Daily reading without the slow stuff here can warp your sense of what is trending versus what merely got covered a lot.

The staggering is the method. Dailies tell you what happened. Weeklies tell you what to make of it. The long-cycle reads tell you whether the thing you read about all year actually mattered. No single source does all three, and no single source is unbiased. The closest thing to an unbiased source is a system, and you have to build it yourself.

The person at the corporate event wanted a name. The answer is a method. I understand the disappointment. A name would be easier. It would also be outsourcing your judgment and calling the result truth.

Subscribe now

An exception

One more note on the AI exception that is coming. Earlier I said there is nothing false to catch, that the fight is over framing rather than facts. AI-generated video is where that stops being true. It produces footage of events that never happened, and footage is the one format people still treat as proof. So far the problem is smaller than the panic suggests. Most fabricated video is still cheap, still detectable, and still gets caught within hours. That window is closing fast, and it will close faster than most newsrooms are ready for.

A reading system cannot fix this one. When the disputed thing is whether the event happened at all, subtracting the lean gets you nothing. This is the rare case where technology and regulation have to carry the load: provenance standards like C2PA that cryptographically mark where a file came from, which most platforms currently strip on upload, and disclosure rules like the EU AI Act’s Article 50, which begins enforcement on August 2 of this year and requires AI-generated and manipulated content to say so. Neither is close to sufficient yet. Both attack the right problem, because forgery, unlike framing, is actually solvable at the platform layer. Watch how those two develop over the next year. For everything else, keep reading widely.

That is where we are now. When I last updated my scenario weights around the Iran war, Managed Fragmentation sat at 75%. It was the base case, and a comfortable one for planning purposes: assume the grind continues, build buffers, watch the tails. As of this week, I have Managed Fragmentation at 50% and all-out war at 40%. A durable ceasefire with real peace holds the remaining 10%.

When one scenario dominates, you can plan for it and hedge the tail cheaply. When two scenarios of roughly equal weight point in different directions, hedging both is expensive and planning for either means being wrong about the other almost half the time. Shannon entropy, the mathematical measure of uncertainty I have written about before, peaks exactly here: when outcomes approach equal probability. This is, formally and practically, the moment of maximum uncertainty. Most planning frameworks quietly assume a dominant base case. Right now there isn’t one.

Read the numbers above again. The gap between the two most likely futures for the Iran war, for energy markets, and for the global economy are now roughly equal. A coin flip with a thumb on one side.

Here is what the past ten days did to my weights. The interim ceasefire signed in June collapsed in early July. The United States restored its blockade of Iranian ports and has now run more than a week of consecutive nightly strikes, expanding beyond the coast to targets in northern Iran. Two American service members are dead and one is missing, the first US combat fatalities from Iranian fire since March. Iran has widened its retaliation to Bahrain, Qatar, Kuwait, Oman, and Jordan. Mediators floated a 10-day ceasefire; Washington waved it away. Brent has climbed back toward $90, up roughly 30% from its July lows. The Strait of Hormuz, closed by Iran since early March, saw 14 transits on a recent Sunday against 37 the week before, and the ships that do move are moving under US military escort along a contested lane hugging the Omani coast that Iran keeps attacking.

Long story short: the thing I have been calling Managed Fragmentation is straining at its upper boundary. Whether it holds or breaks is the question that will define the rest of 2026.

Subscribe now

Why I’m doing this post differently

Readers of this newsletter know I have mostly avoided writing the thing I actually do for clients: assign probabilities to scenarios, then map the business consequences of each. That has been deliberate. Business impact analysis is only useful when it is specific to the business’ footprint, suppliers, and financial positioning. The world does not need another generic geopolitical risk newsletter, and I have no interest in writing one.

This moment is the exception. When two outcomes are near parity, the mapping itself becomes the scarce good, because almost no business plans for two futures at once. So this week I am going to walk through what each road looks like between now and September, and then what each means for companies and macro investors. Consider it a look at the kind of analysis that usually stays behind a client wall, published because the moment calls for it.

One disclosure before we start, because I hold myself to it: I advise multinationals on exactly these questions, and I keep a strict wall between that advisory work and any investment activity. Everything below is directional scenario analysis, not trade construction, and certainly not financial advice.

First, the 10%: why peace is so unlikely

Every scenario map should clear its least likely branch first, so let’s do that.

A durable ceasefire requires at least one leader who benefits from being seen to settle. Look around and try to find one. Netanyahu faces an election on October 27 with final party lists due September 9 and a Likud primary on August 4; his path to political survival runs through a war record, not a settlement signed under pressure. Trump has spent months framing Iran as a problem only he can solve; accepting a mediated compromise now, after American deaths, would hand his opponents the “wrapped up nothing” attack line just as the political season resumes leading into the US midterm elections. And Tehran, having absorbed strikes on its cities and the assassination of allies’ leadership, will not be able to sell a deal to its most hardline elements that reopens the Strait in exchange for sanctions relief that arrives, if ever, in installments.

Then there is the track record. Every truce this year has broken. The April 16 Israel-Lebanon ceasefire began fraying within days of its signing. The June interim agreement between Washington and Tehran lasted roughly a month. The 10-day proposal on the table now is smaller in ambition than either of those, and the parties are further apart than when those failed. Structurally, the Strait dispute has hardened from a security standoff into a sovereignty claim: Iran is no longer merely threatening transit, it is demanding that all shipping use a northern route through its territorial waters. Sovereignty claims do not get compromised away in ten days.

Could a real peace still happen? Yes, and that is why the number is 10% and not zero. Exhaustion is a strategy too, and both economies are bleeding (the Iranian economy is seemingly in substantially worse shape, but was also prepared to be on a war footing). But nobody’s September calendar rewards peace, which brings us to the argument at the center of my analysis.

Subscribe now

The September convergence

My core claim is that actors across at least five separate theaters are all working toward the same soft deadline: the start to middle of September. None of them coordinated this. That is precisely what makes it dangerous. Readers of my argument that World War 3 has already started know I treat these theaters as one connected conflict; what follows is that conflict’s calendar. Independent clocks, all striking at once, mean the next six to eight weeks are likely to see the most intense activity of this war, because everyone wants their position settled before the bell.

Let’s walk through the clocks one at a time.

The Israeli clock. Final candidate lists for the October 27 Knesset election must be filed by September 9. The Likud primary runs August 4. After September 9, every Israeli political actor is locked onto a list and campaigning on a record. Netanyahu needs that record to read “restored deterrence, degraded Iran” by the time lists close, which means Israel’s incentive is maximum military achievement in August and early September, followed by consolidation. Wars are hard to run and win during election campaigns.

The American clock. Labor Day falls on September 7. American political attention, dormant in the summer the way it always is, snaps back with the school year and the return from vacation. The 2026 midterm campaign begins in earnest at that moment. Trump’s incentive is to have Iran “handled” before the electorate starts paying attention again, so the story he carries into the fall is victory rather than quagmire, and so any economic damage can be blamed on the war’s end rather than compounding through its middle. American deaths make this clock louder, not quieter. Every additional week of casualties past Labor Day converts a foreign policy story into an election story.

The Gulf clock. The Gulf states need the Strait fully open by early September, and the reason is more domestic than most Western analysts appreciate. The school year starts then, and with it the annual decision point for expatriate families: return, or stay back. Gulf consumer economies are substantially impacted by those families. A Strait still closed when school starts means more empty seats in classrooms that may not refill for years. Behind that sits the tourist season, which begins in late October and requires venues, hotels, and supply chains to start provisioning in September. The Gulf states have absorbed Iranian strikes with restraint so far. Their tolerance is not infinite, and their calendars say September.

The Russian clock. Russia holds State Duma elections September 18-20, with United Russia polling around 34%, down from roughly 40% at the war’s start and having touched 29% in March. Ukrainian long-range operations have produced a genuine fuel crisis inside Russia, and Kyiv has been explicit that its strikes are timed as pressure on that vote. A Kremlin managing a domestic election under drone attack is a Kremlin with limited bandwidth for its Iranian partner. Moscow’s capacity to resupply, advise, or diplomatically shield Tehran is at its weakest exactly when Tehran needs it most. That window closes after September 20.

The market clock. September is the only month that has averaged a decline over nearly a century of stock market history: minus 1.2% on average since 1928, negative 55% of the time, home to nine of the forty worst monthly losses ever recorded. The crashes of 1929, 1987, and 2008 either landed or accelerated in September. Fund managers know this. Positioning ahead of September is defensive in normal years. This year, defensiveness meets a market priced for perfection in one specific corner, which we will get to below.

Five theaters, five clocks, one convergence. Put differently: if you are Iran, the US, Israel, the Gulf states, or Russia, your incentive is to shape facts on the ground in the next six weeks, because after mid-September your freedom of action shrinks. That is why I expect August to be the most violent month of this war regardless of which road we end up on.

So what do the two roads actually look like?

Subscribe now

Road one: Managed Fragmentation until September (50%)

Managed Fragmentation is the war continuing inside boundaries both sides tacitly maintain. It is what we have had, with the boundaries now under more strain than at any point since March.

Concretely, between now and September this road looks like the following. The Strait stays closed or throttled, with convoy transits under US escort continuing at a fraction of normal volume and Iran periodically attacking ships to remind everyone it can. Ceasefires start and stop; at least one more short truce that collapses, because truces serve both sides as repositioning pauses even when neither intends to keep them. The US and allies conduct limited ground operations: special operations raids against missile and drone infrastructure, possibly short incursions measured in days, but no seized and held territory. Holding territory is the boundary that separates this road from the other one, and both militaries know it.

Lebanon moves again. The July 20 handover of buffer zone sections to the Lebanese government looks like de-escalation, and partly it is, but it also clears the board for the next round. Hezbollah has been degraded, its supreme leadership assassinated, and more than a million people displaced, but the April ceasefire has been violated in both directions since the week it was signed. In a Managed Fragmentation world, Lebanon (and, to a lesser extent, Iraq) is where pressure gets vented when the Gulf gets too hot: strikes and counter-strikes that let both Iran and Israel show action away from the Strait.

Economically, this road is a grind that compounds. War-risk insurance premiums stay elevated, shipping stays rerouted, Gulf logistics stay expensive, and energy prices stay high with a floor that keeps ratcheting up after each incident. Inflation re-accelerates slowly rather than spiking. The key financial fact about this road: it is survivable for markets but corrosive, and it carries a smaller but real probability of a September market break anyway, because the market clock described above may not care which road produced its bad news.

The most important thing to understand about Managed Fragmentation at 50% is that it is metastable. It persists only as long as both sides keep choosing it nightly. Which is why the second road is only ten points behind.

Subscribe now

Road two: all-out war (40%)

All-out war means the boundaries that have contained this so far get crossed. The likeliest trigger is not a decision but an accident: a strike that kills too many Americans at once or an Iranian missile that hits a Gulf capital’s civilian center or a sunken warship. At 40%, I am saying there is a two-in-five chance one of those thresholds gets crossed before mid-September.

What does it look like? Start with duration. A full campaign against Iran is not Iraq 2003, and nobody serious thinks it is. Iran is four times Iraq’s size with over 90 million people, mountainous, and defended by a state that has spent forty years preparing for exactly this war. But the war aim would not be occupation. It would be the destruction of Iran’s military capacity, its nuclear program’s remnants, and quite possibly the state’s coercive apparatus, from the air and with limited ground forces at specific points. My estimate is a high-intensity phase of two to four months, which, if it starts in August, runs through the fall and makes September its pivot month, followed by a long, ugly aftermath measured in years.

Iran’s playbook in this road is asymmetric and explicit. Expect direct attacks on any landed troops, which Iran has both the doctrine and the missile inventory to attempt. Expect attacks on civilian infrastructure across the Gulf: desalination plants, power grids, airports, ports. Water is the Gulf’s soft underbelly; desalination supplies the overwhelming majority of drinking water in the peninsula, and those plants are large, coastal, and hard to defend. Expect soft-target attacks beyond the region through proxies and sleeper networks, because internationalizing the pain is Tehran’s only way to impose costs on Washington directly. And expect the Strait to become an active battlespace rather than a blockaded one, with mining, anti-ship missiles, and swarm attacks replacing the current calibrated harassment.

The war ends, in this road, when Iran’s ability to strike outward is broken or when Washington’s political clock forces a settlement, whichever comes first. Note that the September convergence cuts both ways here: the same American political calendar that pressures Trump to finish by Labor Day pressures him to escalate hard in August rather than accept a slow grind into the fall.

What each road means for companies and macro investors

Two roads, two very different economic worlds. I will take them in turn, sector by sector, with illustrative names where they help. These are examples chosen to make the mechanism concrete, not recommendations of any kind. This is the part I often keep behind the client wall and that has highly specific differential impacts on each company depending on their situation.

Subscribe now

If it’s all-out war

Inflation gets worse, broadly, and stays worse. Prices rise across energy, shipping, insurance, food, and everything transported. Brent above $120 is the entry fee; sustained Strait combat could take it well past that, and roughly a quarter of the world’s seaborne oil and a fifth of its LNG normally transit Hormuz. The second-round effects matter more than the spike: airlines repricing fuel surcharges (watch what airlines do to fares within weeks), chemical and fertilizer producers passing through feedstock costs, and food inflation with a six-month lag. The Fed’s soft landing dies in this road. Rate cuts stop; the conversation flips back to hikes into a slowing economy, the worst of both worlds.

The dollar’s erosion accelerates. Here is the uncomfortable mechanism: war spending lands on top of an already enormous US debt load, and this time the traditional wartime dollar bid competes with a long-term, structural story about US fiscal capacity. In every Middle East crisis of the past fifty years, the dollar strengthened as capital fled to safety. Watch whether that reflex weakens this time. Gold’s behavior since 2024 suggests central banks have already been positioning for exactly this scenario. If the dollar bid fails to materialize in the first weeks of full war, that is one of the most important macro signals of the decade. Instruments that express this: gold, TIPS breakevens, and the dollar index itself.

The AI bubble bursts in September. This is my most specific and most contrarian call, so let me “show” rather than “tell” to start. The AI buildout is on track to absorb more than a trillion dollars of data center spending in 2026, with four companies alone (Amazon, Alphabet, Microsoft, Meta) committed to roughly $725 billion. That spending assumes cheap energy, abundant capital, and government enthusiasm. All-out war undermines all three at once: energy costs spike, capital reprices as rates back up, and Washington’s fiscal and political attention rotates from subsidizing data centers to funding a war. Add the market clock: leading AI names trade at price-to-sales ratios above 30, analysts were already watching second-half earnings as the moment Wall Street’s patience snaps, and September is the month that has hosted more of history’s worst market losses than any other. A war-driven energy shock hitting the most extended valuations in the market during its historically weakest month is how bubbles actually end: an external pin, not an internal realization.

But probably no mass unemployment. This is the counterintuitive piece. Most of the losses in this road are paper losses: equity drawdowns, bond repricing, venture markdowns. The real economy of jobs holds up longer than the market does, because war spending is itself stimulus and because labor markets entered this crisis tight. The transmission channel to Main Street is credit, not layoffs: households and firms that borrowed against inflated asset values default as those values reset. Watch consumer delinquencies and regional bank credit provisions, not the unemployment rate. The pain arrives through balance sheets before it arrives through paychecks.

Who benefits: defense primes (Lockheed, RTX, Rheinmetall as the European expression but not necessarily through stock prices because of the more limited margins during war time), energy producers outside the Gulf (US shale, Brazil, Guyana), tanker owners on the routes that still function, and cybersecurity, because Iranian retaliation will be partly digital.

Subscribe now

If it’s Managed Fragmentation

More of what you have already seen, but compounded. Elevated energy with a rising floor rather than a spike. War-risk premiums as a permanent line item for anything touching the region; the Lloyd’s market has already repriced Gulf hulls the way it repriced Black Sea hulls in 2022, and that does not reverse while the Strait is contested. Shipping stays rerouted and expensive, through mechanics I mapped in my shipping industry deep-dive; Maersk’s quarterly guidance is as good a real-time indicator of this road’s cost as any official statistic. Gulf-exposed sovereign debt trades with a persistent risk premium.

A September market break is smaller but still live. Call it a meaningful minority probability within this road. The mechanism: even without full war, the September convergence delivers a dense cluster of headline risk (Israeli lists, Duma elections, post-Labor Day repricing, Q3 pre-announcements from AI-exposed names) into the market’s weakest seasonal window at extended valuations. Managed Fragmentation does not protect the AI trade, but it gives it a slower downfall. If second-half capex does not start showing up as software revenue while energy costs grind higher, the same repricing happens over months instead of weeks.

The corporate trap in this road is normalization. Companies are treating a closed Hormuz as a supply chain problem through longer routes, higher premiums, or delay buffers. A closure that persists past September becomes something categorically different, a structural reset of costs, suppliers, and trade relationships. Those are different problems and they require different responses. Supply chain problems get managed by logistics teams. Structural resets get decided by boards. Most companies I speak with are still staffing this at the logistics level and not thinking abotu the strategic implications.

For macro investors, the trade in this road is patience with protection. The grind rewards quality balance sheets, energy exposure, and volatility ownership into September, and punishes anything that needs cheap capital or calm seas. The mistake would be reading a quiet August week as resolution. In this road, quiet weeks are repositioning pauses. I said the same thing about the June ceasefire, for whatever that record is worth.

Subscribe now

How to stand when you can’t pick a side

I opened by saying this is the moment of maximum uncertainty, so let me close with what you do about it, because “wait and see” is a decision too (and usually the worst one).

You cannot cheaply hedge two near-equal outcomes. So stop trying to predict, and start building posture with three components:

First, know your tripwires. The September convergence gives you a ready-made dashboard: the Likud primary on August 4, the Israeli list deadline September 9, Labor Day September 7, the Duma elections September 18-20, the pace of US strikes and casualties week to week, and the Strait transit count every Sunday. Each of these moves probability weight between the two roads. When American casualties mount or a Gulf civilian center gets hit, weight shifts to war. When convoy transits rise and a truce survives past day ten, weight shifts back. Do not wait for my next update to move your own weights.

Second, find what can work for you on either road. Some positions and decisions work on both: energy cost discipline, reduced reliance on Gulf transshipment, war-risk coverage locked in before the next repricing, and defensive positioning into September specifically. Whatever your exposure, the September seasonal window plus this convergence is the wrong moment to be maximally extended. That statement is true at 50/40 in a way it would not be at 75/15.

Third, know your divergence points. Somewhere in your business or portfolio is a decision that goes one way under fragmentation and the other way under war: a Gulf expansion, a supplier contract, a hiring plan, a leveraged position. Make it clear now, and communicate it internally to those who need to know with the information on the trigger that flips it. The companies that came through 2022 best in Ukraine were the ones that had done this before the tanks moved, and in interview after interview since, their executives told us the same thing: the plan mattered less than having decided, in advance, what would change their mind.

Two roads, six weeks, one deadline. Chance favors the prepared mind.

If your board, executive team, or investments team want to talk about this might map against your specific footprint, reply to this email/DM me.

Then I told him the part that usually surprises people. Risk models cannot wait historians to sort out the labelling and my World War 3 statement (provocation?) is about understanding how these three conflicts are fundamentally intertwined. It doesn’t really matter whether you agree with me or if I’m wrong on this one. Whether you accept my framing or reject it, the connection between the conflicts should mean that geopolitical risk models the same possibilities. The states aligned around Russia, Hamas, and Iran are acting as one bloc. The states opposing them are acting as another bloc. Two blocs, at war, with no end in sight. The “World War 3” label (I think) carries that information but the realities are there regardless of anyone’s acceptance of the label.

That was my answer in the room. In this week’s post, I want to spend some more time with it because it informs a lot of my thinking about the risk that the current moment poses to companies. The biggest challenge is to see through the public pronouncements and to the actions of both blocs. The blocs are not acting as mirror images of each other. One coordinates supply but with domestic economic constraints. The other coordinates force but with domestic political constraints. Neither coordinates everything. And the coordination layers where each side is weak are usually the layers that produce all the headlines, which is why so much commentary gets both sides wrong at once and ends up being confused.

I’ll go further: the two blocs need names that capture what actually binds them, because “East and West” tells you geography and “autocracies and democracies” tells you ideology, and neither tells you how these blocs will behave under stress. So call them what they are. One bloc is the Depositors: the states whose elites deposit their wealth, their children, and their commercial disputes inside the other side’s institutions. The other is the Outsourcers: the states that outsourced their production, their energy, and their cost structures into the first bloc’s economies and spheres of influence. Each side has entrusted something it cannot replicate at home to a system it is now at war with. Hold that thought. It explains almost everything that follows and how companies need to model their risks and scenario planning accordingly.

Subscribe now

Opposing blocs and supply chains of violence

Start with what the “one war-three theaters” framing is based on because the fact pattern is not seriously contested.

Iran sold Russia thousands of drones, plus missiles and other hardware, for use in Ukraine, and transferred the production technology so that Russia now builds Iranian-designed drones domestically. Russia reciprocated with warplanes and captured Western equipment. In the current war, reporting indicates Russia has been sharing satellite intelligence, drone technology, and battlefield methods with Tehran, including hard-won lessons on drone navigation in jammed environments from three years of fighting in Ukraine. Chinese propellant precursors, including documented shipments of sodium perchlorate, have moved into Iranian missile production.

The pattern predates war in the Gulf. Captured documents verified by the New York Times show Iran knew of Hamas's plan in principle from at least July 2023, when a senior IRGC commander told Hamas that Tehran and Hezbollah welcomed it but needed time to prepare the environment. Israeli intelligence maintains that Hamas could not have carried out the attack without IRGC training, weapons, and funding. And yet no intelligence assessment has established that Iran controlled the timing or scope, and no evidence ties Moscow to foreknowledge at all. Build the capability, welcome the plan, deny the operation. That is the Depositor pattern in miniature, visible sixteen months before the treaty behavior confirmed it.

Follow the materiel. Capability built for one theater shows up in another within months. Drone tactics refined over Kharkiv now shape strikes and drone defense over the Gulf. Missile fuel loaded in a Chinese port becomes a threat to shipping in Hormuz. Financing structures built to evade Russia sanctions now serve Iranian oil sales.

For anyone building a risk model, the connections are obvious and need to be taken into account. Three theaters, two blocs each with its own supply chain. If your model treats Ukraine exposure and Gulf exposure as independent events, it is wrong by construction, whatever you call the war(s).

Subscribe now

What each bloc actually is

Here’s the rub. Both blocs are real, but they are held together by different things, and the difference determines what kind of war this is and how each theater is likely to play out.

The Depositors are a logistics network without a defense obligation. The comprehensive strategic partnership treaty Russia and Iran signed in January 2025 contains no mutual defense clause. Compare the pact Russia signed with North Korea in 2024, which obliges Moscow to join any conflict Pyongyang enters. The Iran treaty merely commits each side not to aid an aggressor against the other. Moscow knew the difference when it signed both documents.

The behavior matches the paper. Iran fought the 12-day war of June 2025 largely alone, with meaningful backing only from the Houthis, and it is fighting largely alone now in the Gulf. Russia and China condemned the American and Israeli strikes, requested emergency Security Council meetings, and sent condolences. They did not send air defense support.

Why not? Because this is where the Depositor constraint bites. China buys roughly 90 percent of Iran’s exported oil, and roughly 40 percent of China’s own crude imports transit the Gulf. Iranian strikes on GCC territory attack Beijing’s supply lines. Both Moscow and Beijing hold extensive economic positions across the Gulf monarchies that Iran has been hitting. And the bloc’s elites hold something even more personal inside the other side’s system: the frozen Russian central bank reserves, the sanctioned London properties, the children enrolled at American universities. A bloc whose wealth sits in the opposing bloc’s financial institutions fights with one hand occupied, counting its deposits. Putin has spent this war pressing Washington to end it quickly, not to lose it.

So the Depositors will sell you the precursors, share the targeting data, and buy the oil, but they will not fly the sorties or commit ground troops for anyone else’s war. That cap on commitment is priceable: the escalation tail is bounded by what the bloc’s own economics will tolerate, and what they will tolerate is a customer relationship, not a cause.

The Outsourcers are a defense network without real political consensus. This is where the fashionable analysis fails, and it fails because it reads the wrong signals.

The political fractures are real and loud. The G7 split over Palestinian recognition, with the UK, Canada, France, and Australia moving while Washington, Tokyo, Berlin, and Rome held out. European capitals split again over the Trump’s Board of Peace. NATO members issued statement after statement against Israeli conduct in Gaza.

Now look at what actually moved operationally. Basically nothing. Defense postures across NATO did not shift because of Gaza. Basing access in the Gulf held through the current war. And when Iranian missiles and drones came at GCC territory, the defense effort came from the Outsourcers, in force. UK warplanes have downed drones and flown defensive sorties for the UAE. France's naval base in Abu Dhabi has itself come under drone attack. Australia sent missiles and a surveillance aircraft to help Gulf defense. And in the most telling deployment of the war, Ukraine sent 228 military specialists to the UAE, Saudi Arabia, Qatar, Kuwait, and Jordan to run interception of the same Shahed drones it has fought over Kyiv for four years, at Washington's request. Iran's partners contributed satellite imagery and chemical precursors. One bloc sent interceptor crews. The other sent targeting data.

Why does the posture hold while the politics fray? Because the Outsourcer constraint is domestic and political, not economic. Voters in London, Paris, and Berlin want cheap energy, cheap goods, and no wars, and they punish governments that deliver expensive versions of any of the three. That makes public criticism nearly free and posture changes nearly impossible: unwinding a basing agreement or an interoperability standard means explaining to your electorate why their heating bills just went up and their security just went down. The result is a bloc that argues constantly and moves rarely.

The apparent exception proves the rule. Spain denied US forces the use of its bases at Rota and Morón for the Iran campaign, the kind of refusal my framework flags as a warning sign. But watch what the refusal actually changed: nothing. No US strike route ran through Spain, no mission was rerouted, and no coverage degraded. The denial came at no operational cost to Washington and considerable domestic benefit in Madrid. The signal to watch is a refusal that forces something to move and has real-world operational impact, not one that makes no difference except for the shaping of domestic political perception.

Put differently: the Outsourcers send force, but argue in public. The Depositors agree in public, but only send inputs. Posture is expensive like buying something online and having it shipped to you. Recognition votes are cheap like “likes” and reshares online. Watch the expensive signal to see what will actually impact companies and markets.

Subscribe now

The layer test

So the framing needs one refinement to be usable. Alliances have layers: political consensus, economic alignment, operational postures, etc. Each bloc in this war is coherent at some layers and fragile or nonsensical elsewhere.

The Depositors have deep materiel and economic integration, thin political trust (if any), and no operational commitment. The Outsourcers have deep operational integration, a fraying political consensus, and inconsistent economic alignment. Turkey sits inside NATO, but outside the sanctions regime. Saudi Arabia and Kuwait coordinate oil production with Moscow through OPEC+ while hosting American military forces.

Commentary misleads and misreads in both directions because it sees each bloc at its weakest/strongest layer. Treaty language makes Russia and Iran look like an alliance that they are not. Parliamentary speeches make NATO look like they are about to rupture then they’re.

And yet…

The open question, and I do not claim to know the answer, is whether Outsourcer operational coherence will continue structurally or is only inherited and will not survive these conflicts. Basing agreements, interoperability standards, and procurement cycles were built over decades and unwind slowly. It is entirely possible that posture held through 2026 because it was locked in years ago, while the political consensus that would renew it has already thinned if not disappeared entirely. If so, the fracture is real but is lagging, and it will show up first in boring paperwork like basing renewals, joint procurement commitments, and exercise calendars. A NATO or Gulf state declining a material request, rather than making a critical speech, is the signal that the political layer has finally reached the operational one. Watch for it over the next 24 months.

Subscribe now

The countries in neither column

One more problem with any two-bloc map, mine included: the largest share of some companies’ actual exposure may sits in countries that belong to neither bloc.

India’s imports from Russia rose roughly sixfold after 2022, passing $60 billion, even as New Delhi deepened security cooperation with Washington through the Quad. Turkey holds NATO membership, but declines the sanctions regime. Some Gulf states host American basing while coordinating with Moscow on oil, and the UAE’s conduct in this war, leaving OPEC, extracting conditional security guarantees from Washington, is what a hedger renegotiating its position looks like, not a state choosing a side. Brazil and Indonesia decline entanglement outright.

These states are not confused. They are deliberately hedging, and they have been rewarded for it (so far). Notice that the hedgers are also where the two blocs’ constraints meet: the Depositors park wealth in Dubai and the Outsourcers park supply chains in India, which is precisely why neither bloc can afford to force the middle to choose. If your scenario framework has only two columns, the countries holding your manufacturing base, your energy supply, and your growth markets may be in neither. They need their own column, with their own trajectories, including the scenario where a forced choice collapses the hedge. A Hormuz closure long enough to compel a decision would do it. Until then, the third column is where the war’s economic impacts actually reside.

Subscribe now

What this means for companies and investors

Four instructions follow from the structure, and none of them depend on whether you accept the World War 3 label.

Correlate your theaters. The same suppliers, financing channels, and targeting infrastructure now serve Ukraine, Gaza, and the Gulf. A risk model that treats these as independent draws underestimates tail risk by design (not to the model’s benefit). If your Ukraine scenario and your Hormuz scenario do not share variables, rebuild them.

Discount statements; upgrade actions. Build your indicator set on basing access, overflight permissions, procurement commitments, and exercise participation. Public statements on both sides are domestic political management: the Depositors’ condemnations cost them nothing with their publics, and the Outsourcers’ criticisms buy domestic cover for postures that have not moved. Discount statements, whatever their volume.

Plan for chronic, not acute. No mutual defense clause means no clarifying escalation. No defense obligation means no patron intervention that ends things decisively. And each bloc’s constraints reinforce the pattern: Depositors will not risk their deposits on escalation, while Outsourcers will not spend political capital on resolution. Both structures favor long, attritional, intermittent conflict. Insurance terms, contract language, and force majeure provisions should be written for a decade of intermittency, not a quarter of crisis. I made a version of this argument about supply chains in the pharmaceutical installment of the industry specific series: the firms that performed under shock were the ones that had built for duration before the shock arrived. The same logic now applies to the war itself.

Give the hedgers their own column. India, Turkey, many of the Gulf countries, Indonesia, and Brazil are running independent strategies and should be modeled that way. Include the collapse scenario, and define in advance what would trigger it for each.

Subscribe now

Decision architecture

Five indicators, and what each one changes:

Mutual defense language appearing in any Russia-Iran or China-Iran instrument. The Depositors would be converting a customer relationship into an obligation, and the escalation tail widens materially. Re-run tail scenarios.

Chinese transfers crossing from inputs to platforms. Propellant precursors are a customer relationship. Air defense systems or aircraft would be an alliance. Reprice accordingly.

A material refusal inside the Outsourcer bloc. A NATO or Gulf state declining basing, overflight, or munitions transit in a way that forces rerouting, degrades coverage, or delays operations. That last clause is the test: Spain's denial of Rota and Morón fired the formal indicator but failed the impact test, because no American operation depended on Spanish territory. A refusal Washington doesn't have to plan around is a statement. A refusal that moves aircraft is a posture shift. Only the second kind means political fracture has reached the operational layer; when you see it, localize continuity planning.

A Hormuz interruption long enough to force the hedgers to choose. The third column collapses, and the two-bloc map becomes all-encompassing for the first time.

Posture renewals lapsing quietly. Basing agreements, joint procurement, or exercise calendars expiring without replacement over the next 12 to 24 months. Outsourcer coherence confirmed as inherited rather than structural. Extend your planning horizon for allied fragmentation.

Subscribe now

Back to the room

The questioner at that talk wanted to know whether I really believed this was World War 3. I do. But grant the opposite fully, call it three regional wars with overlapping suppliers, and then notice what changes in your firm’s planning.

Nothing. The correlations are the same. The indicators are the same. The chronic-conflict horizon is the same. A firm that accepts my framing whole and a firm that rejects it whole should build the same model.

The label was never the risk. The risk is a war between a bloc that banks with its enemy and a bloc that shops with its enemy, where each side is strong exactly where the other is weak, where the loudest signals come from the layers that matter least, and where most of the world’s economic weight is held by states betting they never have to choose. History will name this period eventually. Your planning cycle ends before then.

Think about the number for a second before reaching for a villain. Almost none of this violated any law. Most of it ran inside the G7 price cap, and was papered, attested to, and legal. That is the part that should worry all of us, but should especially worry companies and their investors.

Subscribe now

Six countries, one ship

Think of a single tanker voyage. The ship flies the flag of one country. A management company in a second country runs the ship. Ownership sits in a shell company in a third country. A charterer in a fourth country hires it. An insurer in a fifth country covers the ship. A classification society in a sixth country certifies it.

None of this was intentionally designed to beat sanctions. It was built over decades for tax and liability arbitrage, and it is entirely legal and entirely ordinary. Every major shipping company on earth is structured this way, in peacetime, for peacetime reasons.

Here is the rub. A sanctions regime assumes a chain of control somewhere that a government can grab: a headquarters, a balance sheet, an accountable owner. Maritime was engineered, lawfully and long ago, so that no single jurisdiction ever holds the full chain. The Greek trade looks like a hole in the system only if you assume the system ever had a grip. It never did. The political ground simply shifted, and a fifty-year-old structure became visible to the general public.

The structure has political protection too. Ukraine’s anti-corruption agency listed Dynacom as an “international sponsor of war,” then removed it after negotiations, and the list itself was later pulled from public view. When a peacetime arrangement gets a wartime defense, you are looking at something systemic, and treating it as private opportunism will mislead you.

Readers of my pharma piece (the first in this series) will remember the argument that concentration is a strategic choice, one that only shows itself under stress. Shipping is the mirror image. Dispersion is the peacetime choice, and it defeats state control just as thoroughly. Sanctions against concentration fail at the one point everyone can see. Sanctions against dispersion fail everywhere all at once.

Subscribe now

Dispersion as a business

Now read the same structure as a market. When most of an industry exits a trade, someone gets paid to occupy the vacated link. Shipbrokers cited by the FT put the premium at 30 to 40 percent more for tankers willing to carry Russian crude. “There is money to be made there and no one else will go in and make that money,” said maritime intelligence analyst Michelle Wiese Bockmann.

That is a geopolitical risk premium being harvested by the owners best equipped to price it. Greek shipowners are the most risk-tolerant operators in the industry, and this is a considered decision to take these risks, not a mistake or a lapse of judgment. The same appetite shows up wherever exposure pays. Since the Gulf war opened on February 28 and Iran moved to close the Strait of Hormuz, Greek tankers, Dynacom’s above all, have been among the very few willing to run the strait, at roughly four times pre-war daily rates.

The premium and the tail risk are one trade. Whoever prices it well compounds quietly for years. Whoever prices it wrong is next quarter’s story.

Subscribe now

The rest of the industry made the same trade

The Greeks are only the end of the chain that got noticed. Look at what the rest of the chain did when the Russia trade repriced.

Owners across the sanctioning West sold rather than sailed. Shipowners from 21 of the 35 countries that sanctioned Russia sold at least 230 aging tankers, for more than $6 billion, into what became a shadow fleet, vessels that would otherwise have gone for scrap. Selling the asset into the trade is the same position as running it, one step removed and with the reputational risk laundered through a bill of sale. Roughly a thousand tankers now move Russian oil outside Western insurance, class, and oversight. Flag registries competed for the business as it arrived. Insurers of last resort appeared to cover what P&I clubs dropped. The EU and UK have blacklisted more than 550 of these ships, but the fleet kept sailing, because a blacklisted hull can reflag in days and the registries doing the reflagging are inside the same dispersed structure the sanctions cannot impact.

Under stress, an industry behaves according to how it structured itself in peacetime. This is the core idea behind these looks at specific industries. Maritime structured itself so that every link could reprice a politically toxic trade, and no link had to exit the machinery that makes the trade possible. Under pressure, that is exactly what every link in the chain did.

Subscribe now

What actually moved behavior

The cap never changed Greek behavior. Counterparty designation did. On October 22, Washington sanctioned Rosneft and Lukoil, the producers behind more than half of Russia’s seaborne crude. Within three weeks, only three of 65 Russian crude cargoes loaded on Greek-owned ships, the lowest Greek share since the invasion.

Force is arriving at the same conclusion. Ukrainian drone boats have been striking shadow-fleet tankers since late 2025, and in March one hit a laden tanker carrying Russian oil in the Black Sea. Designation and drones share a logic that the price cap lacked: both ignored the paperwork and landed on the few links that cannot be dispersed. A named producer. A physical hull. Cargo afloat. Enforcers, state and non-state, have worked this out, and behavior across the industry is now being set by the parties who target nodes rather than the parties who regulate chains.

Subscribe now

Insurance: the blind link

One link in the chain deserves particular attention, because the insurer is the point in the chain that is supposed to see the whole voyage. Increasingly, it cannot. War-risk, P&I, and hull cover are priced on assumptions about flag, route, and cargo, and the shadow-fleet era has degraded all three: flags swapped mid-life, AIS gaps and false signals, and cargo provenance resting on the same attestations the price cap relied on. That March strike by Ukrainian drones was the exact spill scenario hull and pollution cover exists for, on a vessel whose route and cargo history no underwriter could have fully reconstructed. A risk that is hard to see is worse than a risk that was mispriced, because mispricing gets corrected by experience, while blindness does not. Pricing a voyage you cannot fully see is the maritime version of running geopolitical risk without a geopolitics function, and as I wrote in the pharma piece, fewer than one in five global companies have one.

Subscribe now

What this means for shipping (and beyond)

Three patterns from the maritime record apply to anyone whose business touches a dispersed chain, which is most businesses.

Dispersion is a strategic choice that gets repriced under stress, just like concentration. Pharma concentrated its inputs for cost efficiency and had to pay the bill when routes closed. Maritime dispersed its accountability for tax and liability efficiency and is paying the bill now that governments want someone to hold accountable. Neither structure is wrong. But both are choices that were made decades ago, for reasons nobody currently in the C-suite intentionally decided, and stress is the moment the bill arrives. The board question is the same in both directions: do you know which choice your industry made, and what it costs under pressure?

Compliance and protection have come apart. For three years, the maritime chain filed its attestations and the trade ran undisturbed, then twenty days of counterparty designation did what the paperwork regime never could. When enforcement shifts from regulating chains to targeting nodes, a clean compliance file protects the filer from liability while doing nothing to protect the business from consequence. Those were the same thing for most of the post-war trading order. They are separating now, and compliance departments are still staffed and chartered for the era when they were one.

The premium and the tail are one position, and you may not have chosen it. The Greeks are long geopolitical risk deliberately, at 30 to 40 percent, with generations of pricing experience. A cargo owner whose goods move on chartered tonnage, an insurer whose book includes flags it underwrote at peacetime rates, an investor holding tanker equities or the commodity flows downstream: each may be long the same premium unknowingly and unpaid. Directional, as always, not a trade. The work is knowing which side of it you sit on.

Subscribe now

The decision architecture, maritime edition

The pharma piece argued that most companies improvise conflict decisions because the frameworks do not exist. The maritime version of that architecture has three elements.

First, a jurisdiction map of every voyage you depend on. Six links, six countries, per ship. The map shows which government’s law actually touches each link, who attests to what, and where the accountable party sits relative to the enforcers most likely to act. Charterers and cargo owners hold compliance certificates instead of this map. The certificate tells you the paperwork was filed. The map tells you where the consequences land.

Second, exit thresholds committed in advance and tied to enforcement events, not prices. The Greeks emptied their Russian book in twenty days because their threshold was clear: designation of a counterparty. Most companies exposed to the maritime chain have no equivalent. Define now which events move you from normal operations to contingency to exit: a counterparty designated, P&I cover withdrawn, a flag state delisted, a hull struck in a trade adjacent to yours. A threshold defined as “when things get bad” is a press strategy, and thresholds set during the crisis will be set by the news cycle rather than by you.

Third, look-through exposure accounting. Trace the freight, cover, and commodity flows you actually hold back to the chains that carry them, the same way the pharma piece traced generic drugs through India to Chinese inputs. The dispersion that hides the chain from regulators hides it from you too, and the parties now setting behavior in this industry, Treasury designators and drone operators alike, will not check whether your exposure was intentional.

Subscribe now

Back to the numbers

The $3.8 billion did not come from beating the system. It came from an industry built in peacetime to be ungraspable, meeting a sanctions regime that assumed someone, somewhere, was holding the wheel. Flag, manager, owner, charterer, insurer, class: each in a different country, each partially accountable, none wholly. For fifty years that was a tax structure. It is a geopolitical structure now.

If your business touches the maritime chain, you already know which link you are. What you may not know is whether your thresholds were chosen or inherited, and the difference between those two only ever becomes visible the way everything in this series becomes visible: under stress, at the worst possible time, in public.

You could not script a stranger split screen. The founding tale of the United States, on full display for the world, about the American Dream and opportunity for all who are willing to work hard for it, next to the oldest and most persistent flaw in that founding: the people who write the rules have more than occasionally owned the game.

It’s what political scientists often call grand corruption. Political and economic power fused into the same hands, with no check between them. Acemoglu, Robinson, and Johnson won a Nobel studying why some nations escape this trap and others don't. Institutions are either inclusive, with real constraints on elites and decision making that is widespread among different groups, or extractive, where one or a few narrow groups capture the state and rewrites the rules to keep the winnings. Extractive arrangements are inherently unstable and potentially violent because, once a ruling class discovers it can convert office into fortune, every faction that wants power or money starts fighting over the same prize. Inclusive institutions are inherently stable (and less violent) because groups can pursue power or money, but not both simultaneously and there are fair paths towards each. That’s why grand corruption is a leading indicator for coups, autogolpes, and civil war across the historical record.

The markets people have started saying this out loud. Matt King, formerly Citi's top credit strategist and now running Satori Insights, argues in a recent report that we are stuck with a "seemingly structural slide towards lawlessness," and that this is "not just a political science observation. It's an investment observation." Gillian Tett gave King's thesis a full column in the FT this week, noting that hedge funds have already moved past debating the diagnosis and are quietly copying the trades of political elites. She calls it the “tribal investment play”.

King and Tett have the direction right. But direction is the easy part. A structural slide toward lawlessness gives a risk committee nothing to act on. How far along is the slide? What could arrest it? What do you check next quarter? These are the questions that matter and to answer those, you need to know which specific checks on power have already failed, which are failing now, and which still hold.

That checklist exists, and American history wrote it. The country has faced the grand corruption problem twice before. It got the first one catastrophically wrong and the second one emphatically right. Four checks produced the second outcome. This year, I need to tell you plainly: on two of them, the checks have already been broken.

Subscribe now

When America Failed to Address Grand Corruption

The South after the American Revolution was dominated by plantation wealth. Those who controlled it were the same people who held political office and the Constitution itself tilted the field for them. The three-fifths clause counted enslaved people - the core of plantation owners’ wealth - toward congressional representation while giving them no vote, which handed the slave states extra House seats and extra Electoral College votes for seventy years. Ten of the first twelve presidents owned slaves. So did much of the Senate and federal judges. These men (they were all men) owned the asset whose value depended entirely on the rules they were writing. The purest expression came in 1857, when the Supreme Court ruled in Dred Scott v. Sandford that Congress could not restrict slavery in the western territories and that Black Americans could never be citizens. A Court whose majority included sitting and former slaveholders had constitutionalized the portfolio of its own class.

Watch what ordinary politics becomes when the fusion goes that deep. In 1856, Congressman Preston Brooks of South Carolina beat Senator Charles Sumner nearly to death with a cane on the Senate floor over an antislavery speech. Brooks’s admirers mailed him replacement canes. Once the ruling class concludes that debate threatens the asset, debate stops being the game. And when the election of 1860 finally produced a president hostile to the asset, the planter elite didn’t lobby, litigate, or hedge. It seceded. The result wasn’t a negotiated settlement. It was a war that killed at least 620,000 Americans, and by more recent estimates closer to 750,000, more than every other American war combined.

Subscribe now

When America Succeeded in Addressing Grand Corruption

After Reconstruction collapsed in the disputed election bargain of 1877, the Gilded Age produced its own elite fusion, industrial this time instead of agricultural. Boss Tweed’s Tammany Hall ran New York City government as a revenue stream, skimming somewhere between $45 million and $200 million in contemporary dollars. State legislatures, which then held the power to choose US senators, could be bought wholesale. Copper magnate William A. Clark bribed Montana’s legislature so openly to win his Senate seat in 1899 that the Senate refused to seat him. His defense has outlived him: “I never bought a man who wasn’t for sale.” Reformers of the era called the chamber a millionaires’ club, and they meant it literally.

This time the country built its way out, on four checks.

Civil service reform came first, and it took an assassination to get it. After a rejected office-seeker shot President James Garfield in 1881, Congress passed the Pendleton Act, which replaced patronage appointments with competitive exams and broke the machines’ grip on the federal payroll. Government jobs stopped being currency.

Antitrust came second. The Sherman Act of 1890 sat largely dormant for a decade until Theodore Roosevelt used it in 1904 to break up Northern Securities, J.P. Morgan’s railroad combine, and the Supreme Court dissolved Standard Oil into 34 companies in 1911. The target was never really consumer prices. Roosevelt’s argument was about power: no private fortune should be large enough to buy the state.

The press came third. Ida Tarbell’s nineteen-part investigation of Standard Oil in McClure’s documented, contract by contract, how Rockefeller had rigged the railroads against his competitors. David Graham Phillips’s The Treason of the Senate named the senators serving corporate paymasters one by one. The muckrakers made elite self-dealing too costly to ignore, and they sold magazines doing it.

And fourth, the 17th Amendment, ratified in 1913, took the power to pick senators away from legislatures that lobbyists could buy wholesale and gave it directly to voters. Phillips’s series gets much of the credit for pushing it over the line. The mechanism that made Clark’s seat purchasable was simply removed.

All of those checks worked because courts, a free press, and the electorate still had enough independent power to force it. Not because Teddy Roosevelt or any of the Progressives of that era were uniquely righteous but because the machinery for accountability still functioned.

Subscribe now

Where the four checks stand right now

Enforcement independence: broken. On June 29, the Supreme Court ruled 6 to 3 in Trump v. Slaughter that the president can remove commissioners at the FTC, and by extension most other independent multi-member agencies, without cause, overruling Humphrey’s Executor, the precedent that had protected them since 1935. The same day, in the companion case Trump v. Cook, the Court carved out one exception: the Federal Reserve, sparing Governor Lisa Cook by a 5 to 4 vote. That single-vote margin is the entire remaining check. This is not a future risk - it’s a NOW problem, a completed event. This runs in the opposite direction from the Progressive/Roosevelt era fixes. The courts dismantled a check that already existed and had worked for more than century.

Disclosure with teeth: hollow, and has been for over a decade. The STOCK Act’s standard penalty for a member of Congress who violates it is a $200 fine, and there have been zero prosecutions since the law passed in 2012, despite watchdog complaints covering tens of millions of dollars in undisclosed trades. The reform effort tells you even more than the enforcement record. Trump endorsed a narrow insider trading ban in his State of the Union, then attacked the broader bipartisan bill the moment it grew to cover the presidency. A discharge petition to force a House vote stalled at 82 signatures. The floor vote that House leadership promised for the first quarter of 2026 never happened. That’s the real lesson for anyone tracking this as a risk indicator: bipartisan agreement that a problem exists doesn’t matter. Bipartisan agreement plus enacted, enforced legislation does. Those are different things.

A press that can make things costly: heavily at risk. The Ellison family now controls CBS and Paramount, and on June 12 the Justice Department cleared their $111 billion acquisition of Warner Bros. Discovery, adding CNN, HBO, and a major studio, declining to challenge the deal after an eight-month review. Paramount had already settled with Trump for $16 million over a 60 Minutes edit, and CBS canceled Stephen Colbert’s show days after he called that settlement “a big fat bribe” on air. The FCC chairman publicly praised the incoming owners’ “commitment to serious changes at CBS.” The Gilded Age muckrakers had to be intimidated one story at a time. This is closer to buying the newsroom.

Direct accountability mechanisms: the one lever still standing, but untested. State attorneys general are the main independent actors still scrutinizing the media consolidation. California’s Rob Bonta says the Warner deal is “not a done deal” and remains under investigation by his office, and New York’s Letitia James is probing it too. Voters will also have their say. The 2026 midterms will be the first real stress test of whether elections themselves remain insulated from everything above them.

Subscribe now

What this means for risk models

Judicial independence over agency removal and press independence over broadcast and cable news aren’t tail risks anymore as they were before Trump’s second term. They’re now the realized base case. Corporate planning has to start from that and understand that half of the checks that underpinned political stability in the United States for the last century have largely been dismantled.

This is where the four checks earn their keep against King’s portfolio conclusions, which are still worth taking very seriously. He argues bonds get dangerous in a lawless world because their floor, the recovery rate, depends on legal enforcement that can vanish with political change, while equities with real assets, pricing power, or the right connections can do fine. He likes gold. And he’s clear-eyed about the tribal trade that Tett describes: with the right political connection your stock goes up tenfold, and when you or your patron falls from grace it gets decimated. His examples are 4iG in Hungary and Alibaba in China. Mine would add every Gilded Age railroad that owned a state legislature until it didn’t.

The tribal trade, in other words, is a bet on regime durability with no expiration date printed on it. The four checks are how you estimate the expiration date in the US context. The more of them that fail and the longer connection-driven repricing continues, the more violent - literally and metaphorically - the eventual reversal. Corporate political risk teams should read the same checks through a different lens: enforcement is now a function of presidential preference rather than agency independence, which means your regulatory exposure is a relationship variable, and relationship variables can flip overnight.

What’s actually left to watch, updated:

1. Whether the Fed exception holds, since it’s the only piece of the removal-power firewall still standing, and it survived by a single vote.

2. Whether state attorneys general and private civil litigation can generate real consequences in the absence of federal enforcement, since that’s now the main accountability check for corporate self-dealing, not the SEC or DOJ.

3. Whether investigative capacity migrates to outlets outside the consolidated broadcast and cable ecosystem, and whether that migrated capacity can still move markets or trigger resignations the way the old press could.

4. Loyalty selection in security and law enforcement not just at the leadership level about the level of middle management and street-level officers, which is the next domino after courts and press and hasn’t fully fallen yet.

5. Whether elite capital, on both sides of the aisle, hedges into gold, hard assets, or offshore structures at an accelerating rate. King’s gold call is one data point. Insiders usually act on this before they say it publicly.

6. Whether the 2026 midterms produce a procedurally unremarkable outcome. That sounds like a low bar. Given where the other three checks stand, it currently isn’t one.

   Subscribe now

Two Americas, One birthday

The country has lived both versions of this story before: the one that let elite capture run until it broke the union, and the one that caught it early enough to fix it without a war. The honest answer for where things stand at 250 is that two of the four checks that made the second outcome possible last time are no longer available in the form they took then. Markets have begun pricing the drift. King and Tett are evidence of that. But drift is not destiny, and the difference between the two American scripts was never the corruption itself. It was whether the machinery for accountability still worked.

If the checks fail, the fireworks to watch won’t just be in the sky.

So when I saw that researchers at Binghamton University just published a method for solving Wordle that wins 99% of the time, I had to read it (but only after I overcame my silent jealousy that they got to study Wordle methods as part of their jobs). This was 9% better than the best available other strategies.

Here is what makes it interesting. The researchers used Shannon entropy, a mathematical measure of uncertainty, to identify which guesses would eliminate the largest number of possible words fastest. The best move teaches you the most about what you do not yet know, even when it has little chance of being correct on its own. The strategy ignores (for the most part) which word is most likely to be the answer. It chooses instead the guess that gains the most information probabilistically.

One of the researchers put it in a way that should resonate with every corporate risk officer: the objective shifts to maximizing the expected reduction in uncertainty rather than the probability of being right.

That is corporate geopolitical strategy at its best. The strongest advice you can give a company about a country’s future comes as probabilities of likelihood, paired with an explanation of why those probabilities matter for the decision in front of them.

Subscribe now

What Shannon entropy actually measures

Shannon entropy comes from Claude Shannon’s 1948 paper that was critical for information theory. It measures how spread out the possible outcomes of a system are. High entropy means many outcomes are roughly equally likely. Low entropy means one or two outcomes dominate.

Consider two countries. In the first, one political outcome is overwhelmingly probable: the ruling coalition holds, elections proceed without incident, the regulatory environment stays broadly stable, but it’s unclear which political party will win the next election. This is an example of a low entropy environment. In the second country, four outcomes are roughly equally plausible: the coalition holds, or a snap election occurs, or a technocratic government takes over, or a prolonged political crisis unfolds with no clear resolution. This is an example of a high entropy environment.

A standard risk index might score both countries identically. Elevated political risk, monitor closely. But they are completely different operating environments. The first country is largely predictable. The second is not. Entropy captures that difference precisely. The more evenly spread the possible outcomes, the higher the entropy, the less your current analysis tells you about what happens next, the more uncertainty there is, the more contingencies you have to fund.

This is the gap in how many companies approach geopolitical risk. They measure the severity of possible outcomes. They rarely measure how spread out those outcomes are. Those are different problems.

Subscribe now

The scenario analysis trap

Scenario analysis became the industry standard because it is helpful and it feels rigorous. You commission a report, a consulting firm produces three futures, each with a narrative and a set of implications, and you present the findings to the board. Due diligence complete.

Here is the rub. Four scenarios with no probabilities attached reduce your uncertainty by nothing. You have named four possibilities. A list of possibilities is not analysis. The likelihood of each of those events is more important than the description of the events themselves.

Put differently: predictions without probability assignments is maximum uncertainty dressed up as strategy. The analyst has told you that multiple things could happen. You already knew that. What you need to know is how likely each outcome is, because the probability tells you where to point your attention.

And the relationship between probability and effort is counterintuitive. A 75% base case does not earn 75% of your planning resources. It earns your steady-state operations. If it holds, almost nothing surprises you, and the business runs roughly as it does today. The real work sits with the low-probability events. Those are the ones that would blindside you, so those are the ones that need contingency plans. Normalcy is built around the likely outcome. Contingency planning is built around the unlikely ones. Treating every scenario as equally weighted gets this exactly backwards, and the cost of getting it backwards is occasionally catastrophic.

When Russia invaded Ukraine in February 2022, more than 400 companies moved to withdraw within weeks. Many of them had scenario documents. They had analysts who had modeled a range of outcomes. What most of them lacked was an honest probability attached to the scenario in which Russian forces crossed the border and Western governments responded with asset freezes and sweeping sanctions within days. Renault, the Western carmaker with the greatest exposure to Russia, sold its controlling stake in AvtoVAZ to the Russian state for one ruble, with a buyback option if sanctions lifted. Three years later, its partner is demanding $1.3 billion before any return. An honest probability on the sanctions scenario in 2021 would have changed the investment calculus.

Subscribe now

What this looks like when you do it right

Earlier this spring, my team and I were asked, as you might expect, to project what might happen over the next 6 to 36 months in the Gulf and the Levant after Iran had started attacking the Gulf countries in response to US-Israeli strikes. We came up with three scenarios. We did not present them as equally weighted.

The first scenario, Managed Fragmentation, was the most likely. The war ends through a negotiated or imposed cessation of major hostilities, but no comprehensive resolution follows. The region settles into an unstable equilibrium of cascading low-intensity pressures rather than a single catastrophic event. We put this at 75% likely (the first ceasefire was announced days later). That number told our clients something specific: if a ceasefire is declared, it will likely hold and the operating environment stays roughly steady. The discipline to maintain is guarding against the normalcy bias that mistakes the absence of one catastrophic event for a “return to normal”.

The second scenario, Regional Escalation, was possible but far less likely. The war does not end in spring and instead widens through expanded Israeli operations, a US move on Iranian territory, or direct confrontation in the Hormuz region. Oil moves into the $150 to $200 range and a global recession becomes more probable. We put this at 15%. Low probability, highest downside, which is exactly why it earned the bulk of the contingency planning.

The third scenario, Reconstruction and Realignment, was the least likely. A comprehensive settlement holds, Iran’s new leadership accepts constraints in exchange for sanctions relief, and Gulf capital drives a reconstruction wave. We put this at 10%. Low probability, highest upside, so the planning attached to it was about which relationships built now would determine market access if it arrived.

75, 15, 10. One dominant scenario that defines steady state, two low-probability tails that define where the planning effort goes. We assigned the favorite a number, named what would move us off it, and accounted for the less likely events before they could surprise anyone. The point was never to pick a winner and stop thinking. The point was to weight the outcomes honestly and aim our attention where a surprise would actually hurt.

Subscribe now

Tetlock and the discipline of the number

Philip Tetlock spent decades studying political forecasting. His Good Judgment Project, which won a four-year forecasting tournament run by the US intelligence community from 2011 to 2015, produced one finding that practitioners resist more than any other: disciplined probability assignment by informed amateurs, even on messy geopolitical questions, consistently beats expert intuition that refuses to commit to a number. Tetlock’s “superforecasters” were over 30% more accurate than intelligence analysts who had access to classified information.

The resistance is understandable. Saying “there is a 65% chance of early elections in Country X” creates accountability. Saying “political instability remains elevated with several scenarios in play” does not. Consulting firms default to the second formulation because it is defensible. It is also, in most cases, close to useless.

The superforecasters shared one habit. They assigned a number, tracked it, and updated it when new information arrived. The number forced them to think. It also forced them to notice when they were wrong and learn from it. And they only needed publicly available information to reach this level of accuracy, not classified intel.

The Wordle researchers found the same thing. The entropy-maximizing strategy rarely guesses the right word on any given turn. It wins by gathering better information faster, updating faster, and converging on the answer before the other approach does.

Subscribe now

What this looks like in practice

Three things follow from this that companies can actually do.

The first is to force probability ranges onto every geopolitical risk scenario. Equal weights as a default will not do. Reach for an explicit split that reflects a genuine analytical view: 55% base case, 25% secondary scenario, 20% tail risk. The numbers do not need to be precise. They need to be honest. A 55/25/20 split that someone is willing to defend in a room beats three scenarios that all carry an implicit 33% because nobody wanted to take a position.

The second is to treat every new development as a Wordle clue. What does this piece of information eliminate? What does it confirm? Which scenarios does it make more likely, and by how much? The discipline lies in updating the probability distribution you already have rather than writing a fresh scenario document every time something happens. That is how uncertainty gets reduced. More narratives will not do it. Better weights, revised as the evidence changes, will.

The third is to point your planning at the tails. Build your steady-state operations around the high-probability case, then spend your contingency budget on the outcomes that would blindside you. The likely scenario rarely needs a plan, because if it holds, the business runs as it does today. The unlikely scenarios are the ones worth war-gaming.

Subscribe now

The 9% gap

The conventional Wordle strategy, built around common letters, solved about 90% of puzzles. The entropy-maximizing approach solved 99%. Nine percentage points.

That gap looks modest until you think about what it represents in a corporate context. A geopolitical risk process that guesses a comfortable word because it feels right operates at 90%. It will be wrong more often than it needs to be. And when it is wrong in a high-stakes market, the cost is a stranded asset, a broken supply chain, or a workforce evacuation.

The companies that manage political risk well do not hold the most accurate predictions. Nobody does. They get better information faster than their competitors, update their assumptions more honestly, and make decisions calibrated to the actual distribution of possible futures rather than the most comfortable story in the scenario document.

That is a method, and it works as well for Wordle as it does for geopolitical risk.

So the deal is signed and not signed. The Strait is open and not open. The sanctions are lifting and tightening in the same week. The MoU has become the Schrodinger’s cat of international agreements, alive and dead until somebody opens the box to look inside.

If you’re working in a company affected by the deal, how do you act on this? Maybe you schedule tankers, or you underwrite them. Maybe you trade crude, or your company might want a piece of the reconstruction Tehran is advertising. You cannot tell from one morning to the next whether you will be sanctioned or open for business, a criminal or a first mover. The uncertainty is the point. The tools of economic warfare have changed, and your company is now standing in the blast radius whether or not it chose to be.

I spent the early part of my career as a lawyer (insert lawyer joke here), and one thing you learn fast is that what is written down matters more than whatever intentions existed when the writing happened. A decade ago, an economic confrontation between two states might have had a fairly standard and relatively small playbook: a sanctions package built over months, maybe an embargo, an asset freeze here or there, all of it slow to assemble and politically expensive to change. That playbook still exists. But it has been joined by a much larger arsenal, most of which either did not exist or was never used this way ten years ago. Economic warfare used to come in a few slow flavors. Now the options are legion, and they are fast.

The drone-warfare parallel

I argued in my piece on the energy thesis that AI and cheap drones have rewritten the cost structure of actual warfare. A $500 quadcopter can disable a $50 million tank. The expensive, slow, centralized platform loses its monopoly to something cheap, fast, and easily replicable so victory shifts to whoever can coordinate the swarm, which increasingly means whoever has the better AI. The battlefield fills with tools that blur the line between civilian and military, that can be deployed and recalled in hours, and that proliferate to anyone with the will to use them.

Economic warfare is going through the same transformation. The old sanctions regime was like an aircraft carrier: powerful, slow to build, hard to turn, costly to retire, and not very many of them. The new instruments are like drones. Cheap to deploy, because a tariff or a designation takes a signature rather than a coalition. Fast, because they move quicker than the institutions built to constrain them. Easily reversible, which is the part most companies forget. And proliferating, because the moment one government shows it has a tool, every other government with a grievance and a functioning bureaucracy takes notes and tries to replicate their own version of it.

Drones did one more thing on the real battlefield. They made everyone a potential target. The same holds here. When the weapon is cheap and the line between civilian and military dissolves, your company does not get to opt out of being a combatant. Someone else makes that call for you. The rest of this piece is about who makes it, how, and what it costs you when they do.

Subscribe now

The security tariff: pressure with a downward gear

Start with the tariff, because it shows the whole pattern in one type of instrument.

A tariff used to be a relatively simple trade tool for protecting a domestic industry or correcting a trade balance. In the past two years governments turned it into a security weapon. Ecuador showed the template at regional scale earlier this year. President Noboa raised duties on Colombian imports to 100 percent, framed as cost recovery, not protectionism. Cross-border crime, his government argued, forces Ecuador to spend an extra $400 million a year on security, with 75,000 troops under emergency curfew. The revenue was tied publicly to that dollar figure and that troop count. Then, within days, Quito ratcheted the rate back to 75 percent. Up fast, down fast, each step pinned to a security claim. The Andean Community eventually ordered both countries to dismantle the tariffs, but the ruling landed a week after the job was already done.

If you imported Colombian flowers or exported Ecuadorian shrimp, here is how that felt from the inside. Your landed cost doubled then fell by a 25 percent inside a single quarter, and the trade body meant to protect you arrived after the damage was already priced in. You cannot hedge that. The whole point of the design is that it moves faster than anything you can do to absorb it.

Washington built the permission structure for countries to develop and use these instruments. Across two Trump terms, tariffs became multi-purpose coercion: Section 232 measures on national-security grounds, threats used as pressure and walked back when demands were met, revenue treated as a feature. The lesson the rest of the world absorbed is that the WTO dispute system, its Appellate Body paralyzed since 2019, cannot move fast enough to stop a determined actor. A government willing to eat short-term retaliation can use a tariff coercively for months before any ruling bites. Lest anyone think this is just a Trump phenomenon, remember the lesson from military technology: once the tool is developed, every future government will use them. The proof in this context is that the Biden administration did not ban the tools the first Trump term developed, but in fact expanded their use in response to the full-scale Russian invasion of Ukraine in 2022.

The Iran confrontation shows how the security tariff can work both ways and play into an active kinetic war. A February 2026 executive order authorized tariffs on any country trading with Iran, with 25 percent cited as the benchmark, and by April Washington was threatening 50 percent on anyone supplying it weapons. That is the upward ratchet. Its real target was every third-party firm and government that trades with Iran. The June MoU, with its phased relief and conditional asset release, is the downward gear. Same logic as Ecuador, at superpower scale: graduated pressure up, graduated relief down, each step conditioned on the other side’s behavior. Traditional sanctions are slow to build and politically expensive to unwind. A tariff or a waiver can be adjusted overnight.

This week’s other news is why the tariff story is not finished. The Ecuador-Colombia case rested on a particular alignment: a right-wing government in Quito pressuring a left-wing government in Bogota that rejected the security framing outright. On June 21, Colombia elected the far-right Abelardo de la Espriella over the leftist Ivan Cepeda by less than a percentage point, ending the era of a left-wing executive. A Bogota aligned with Quito does not retire the security tariff. It may pick the weapon up. The instrument outlives the government that built it, and it changes hands. If your supply chain runs through the Andes, the risk did not resolve on election night. It switched sides of the border.

Subscribe now

The administrative blockade: war without warships

The second tool usually arrives as paperwork, which is exactly why it works.

Earlier in the war, Iran stood up the Persian Gulf Strait Authority, an agency that required every vessel transiting Hormuz to submit ownership, insurance, crew, and cargo data and pay a toll, some reportedly up to $2 million, in yuan or crypto. The legal hook was that Iran signed but never ratified the UN Convention on the Law of the Sea, so in Tehran’s reading the transit rights the Convention guarantees do not extend to everyone. A military blockade is an act of war that invites a military answer. An administrative blockade is a filing requirement. It is much harder to challenge with a destroyer, and it leaves behind precedent, documentation, and institutional memory that outlast any single regime.

For the shipping line, the difference is everything and nothing. Your tanker is just as stuck behind a form as behind a mine, except now you are filling out the form, paying the toll, and legitimizing the agency that trapped you, all at once. That the reopening of the strait became the headline term of the June MoU tells you how well the paperwork worked. You do not negotiate hard to undo something that does not bite.

The dual-use designation: the label is the weapon

A third tool will largely define the next decade: dual-use designations. It is where the US and China have quietly arrived at the same move from opposite ends and something I’ve written about previously.

On June 9 the Pentagon added Alibaba, Baidu, and BYD to its 1260H list of Chinese military companies. None of them make weapons. One runs e-commerce, one runs search, one builds electric cars. The designation rests on the judgment that their technology aids China’s armed forces, a standard broad enough to catch almost any large civilian firm. The effect is concrete: the Defense Department cannot contract with them, and from 2027 cannot buy their products through third parties. China called it “overstretching national security,” which is what it is, and what makes it powerful.

China has largely created a mirror image of the American model. Instead of branding foreign civilians as military, it controls its own exports by end-use. Beijing’s dual-use rules cover roughly 1,100 items, including the heavy rare earth minerals that the drone manufacturing and chip industries cannot function without. Sales to military end-users are blocked. Civilian trade, Beijing is careful to say, has “absolutely no need to worry.” In January 2026 it ran the targeted version, banning all dual-use exports to Japanese military end-users over a Taiwan comment, then easing off days later. Two systems, one structural point.

The United States weaponizes the buyer’s identity. China weaponizes the buyer’s purpose. Both have the same ends in mind: withhold economic support from a perceived enemy by labelling products as something they may or may not actually be.

Changing a label is an extremely cheap option available to governments’ economic warfare arsenal. It can be applied or lifted by announcement, costs nothing to manufacture, and turns a civilian company into a defense contractor without the company lifting a finger. And for companies to overturn the designation takes legal and lobbying time and expense that may be well in excess of the value of the product or the market. As I argued in my piece on the sovereign defense boom, this is how an agricultural drone maker or a battery firm wakes up reclassified, subject to export controls and treated as a legitimate target, having chosen none of it.

One more layer, because it is where the economic war stops being a metaphor for the shooting war and becomes the same fight. A firm named on an entity list one week can find its networks probed the next, because the listing and the intrusion are answering the same question about who counts as a combatant.

Russian state hackers have been systematically hitting civilian companies that support Ukraine: logistics operators, IT providers, maritime firms, air-traffic systems. The selection rule is usefulness. If your logistics or your code helps the other side, you qualify, and it makes no difference that you have never built a weapon. That is a version of the dual-use designation again, executed in code instead of on a list. A tariff degrades your revenue. A designation degrades your market access. A cyber operation degrades the systems that run both.

Subscribe now

What this means for your company

I have spent twenty years talking to people who run companies where the rules change overnight. The lesson never changes. The firms that survive are the ones that understood early who they are, what their values are, and which communities they are really embedded in.

For the next several years, exposure to reversible instruments is going to depend almost entirely on those choices. And if a company is not open about these choices, governments will assume the answers to these questions. Most corporate risk models still treat sanctions as slow, durable, and binary: on or off, and once on, on for years. That assumption is now wrong in both directions. Pressure arrives in a week, by tariff or designation, not in a months-long package you can watch coming. And relief arrives just as fast, which means the firm that mapped the compliance conditions in advance is the one standing in the doorway when it opens. The Iran reconstruction money will flow to whoever is ready on day one, not day ninety.

So put two questions in front of your decision makers. The first is about exposure. Given who we are and what we stand for, which governments have reason to turn these instruments on us, and in each country where we operate, which side of the security-cost and end-use ledger do we sit on? The second is about timing. Where could a sudden de-escalation open a re-entry window before our competitors can move, and what is our plan for being in the doorway when it opens?

Answering these questions well will ensure, no matter what the result of Schrodinger’s MoU, your company is prepared. The economic aircraft carrier is still in the water. But the war is being fought with economic equivalent of drones now, and the deal-makers in Washington, Tehran, Beijing, Quito, and Bogota are already flying them.

Quick note: I had already written most of this piece before the “deal” happened. But as I mentioned yesterday, the deal doesn’t mean much about the Strait opening until safe transit occur consistently and safely. Part 1 of this series closed with a claim: the most important competition on earth right now is not for territory, but for energy. The Hormuz closure is what that looks like when it stops being a thesis and becomes a headline. I planned to use Hormuz as an illustrative risk scenario. I am leaving it where it was in the argument, because the argument is the same. What changed is that the illustration is now in the news.

On February 28, 2026, the United States and Israel launched Operation Epic Fury against Iran. Two days later, Iran’s Revolutionary Guard Corps announced that the Strait stood closed, and warned vessels against attempting passage. What followed is, according to the International Energy Agency, the largest oil supply disruption since the 1973 Arab oil embargo. Roughly 15.8 million barrels per day, or approximately 20 percent of world petroleum consumption, are stranded. The IEA has triggered a record release of 400 million barrels from strategic stockpiles. Bypass infrastructure through Saudi Arabia and the UAE, running at theoretical maximum, can handle at most 7 million barrels per day. The structural supply deficit exceeds 10 million barrels per day that cannot be closed by any alternative routing currently available.

A 10-cent increase in gasoline prices correlates with a 0.6 percentage point drop in presidential approval ratings. That is not a political curiosity. The connection between the energy supply and the legitimacy of those in power is one of the oldest structural relationships in political life. Understanding why requires going back much further than 2026, or 1973, or even 1859.

Part 1 of this series argued that we are living through a shift in which form of energy organizes everything else: from human and animal energy, to fossil fuels and industrialized production, to the electricity and computation that now underpin both military power and economic output. Part 2 is about what history teaches us when the dominant energy supply fails, by looking at what used to happen the previously dominant energy source, namely food. Those lessons largely come from the Third Horseman in the Book of Revelation: famine.

Famine is the pre-industrial world’s version of an energy crisis. Understanding how it worked, why it declined, and what actually caused that decline tells us more about the strategic competition over energy today - and the geopolitical risk for business associated with it - than almost anything written specifically about that competition.

Subscribe now

Famine as Pre-Modern Energy Crisis

Part 1 introduced the Domesday Book as an energy audit. William the Conqueror surveyed England in 1086 because he needed to know, with precision, how much productive capacity he had conquered: arable land, pasture, livestock, plowlands, mills. Everything that mattered was, at bottom, a measure of how much caloric energy the territory could generate and how it was controlled.

Fernand Braudel made the same argument for the early modern Mediterranean. The grand strategies of the Ottoman and Habsburg empires were strategies for controlling the agricultural hinterlands that could feed armies and navies. The Mediterranean was shaped by calories.

Here is what those histories tend to understate: food in the pre-industrial world was not doing one job. It was doing two.

The first is obvious. Food fed people.

The second is less often stated explicitly but was at least as important. Food powered the entire economy. The horse that pulled the plow needed to eat. The ox that hauled grain to market needed to eat. The soldier who carried a pike and marched forty miles in three days needed to eat. The craftsman whose labor supported the whole system needed to eat. The energy chain in a pre-industrial economy ran from sunlight and rain to crops to livestock to humans to labor. Every single step in that chain required food.

Which means that when the harvest failed, two things failed simultaneously. People starved. And the economy’s energy supply collapsed.

This is why historical famines look so catastrophic to modern eyes (that and modern medicine, but we’ll get there shortly). A bad wheat harvest in medieval England didn’t just mean hungry peasants. It meant fewer plow teams for the following season, because the horses starved too, or had to be slaughtered for food. It meant reduced labor capacity, because workers weakened by hunger could not plant the replacement crop effectively. It meant diminished military capacity and capability, because armies ran on the same caloric chain as everything else. The failure was self-reinforcing. Scarcity in one part of the system cascaded immediately into scarcity in every other part. There was no firewall between the food supply and the energy supply, because they were the same supply.

This is what famine actually was. It was what happened when the fuel for everything ran out.

Subscribe now

A Confession Before We Continue

First, a confession about how I made the connection in my head, and why it took so long to sink in.

About a decade ago a friend introduced me to Kim Stanley Robinson’s Mars trilogy. It contains a scene I have been thinking about since. Robinson’s Martian colonists, designing an economy for a new civilization from scratch, face a question every founding society eventually confronts: what should the unit of exchange actually measure? Instead of defaulting to a fiat currency, their answer, worked out by a group of scientists who called their system eco-economics, was the calorie. Because the calorie is an energy unit, and all economic activity is ultimately a transformation of energy from one form to another, it worked for a community literally building from nothing. Food, fuel, manufacturing, transport, computation - in thermodynamic terms, they are all just energy transfers. A monetary system denominated in energy content would be, in their view, more honest about what is actually being exchanged than any fiat currency. Robinson never explains why this is such a radical idea. But it is.

While I was reading a book on famine, given to me by a recently retired friend, Robinson’s eco-economists kept bugging me. There was something there that I wasn’t seeing, a connection that my brain was trying to make, but I couldn’t articulate. When the Strait of Hormuz was closed by Iran, something clicked and I began working on this piece.

What Robinson’s colonists saw as a design problem, looks, from the historical perspective, like a description of what the pre-industrial world lived without ever needing to theorize. Pre-industrial economies did not require eco-economics because they were practicing it. Food was the calorie-denominated unit of exchange that valued everything else: land, labor, military capacity, political power.

Which is to say: the energy transition that the rest of this piece describes is, in Robinson’s terms, a succession of eco-economic base units. Each era runs on whatever calorie-equivalent currently denominates productive capacity. The competition for that resource is, in every era, the most important competition. Right now, the Strait of Hormuz is closed. Twenty percent of world petroleum consumption, in energy terms the calorie base of the current eco-economy, cannot move. Robinson’s fictional colonists on Mars would have perfectly understood, at the level of system design, why this matters. What they didn’t see, looking forward from their fictional present, is that they were also describing why famine mattered so much for so long, and why its decline tells us something precise about the crisis unfolding today.

Subscribe now

The Good News, and the Bad News

Here is one of the most “good-news” stories of the modern era: catastrophic famine has largely disappeared.

Cormac Ó Gráda’s Famine: A Short History documents this with care. Famine mortality fell dramatically across the twentieth century, and especially from the 1960s onward. The number of famines fell as well. Ó Gráda is cautiously optimistic: famines are becoming less common, and the trajectory suggests they will continue to decline.

His explanation has four components, and each one is important.

The first is information. It is substantially more difficult to hide a famine in the information age the way you could in 1932 Ukraine, when Stalin’s government concealed the Holodomor from the outside world while somewhere between 5.7 and 8.7 million people died. The Great Leap Forward famine of 1959-61 killed between 15 and 55 million people, most often estimated at around 30 million, while Chinese Communist Party officials reported fictitious grain surpluses upward through the bureaucracy. Both of those famines depended on information control. Modern communications make that control much harder to sustain.

The second is economic development. Richer societies have more options: food reserves, purchasing power to import, logistics capacity to move food from surplus to deficit regions. They also have more diverse diets and supply chains, which makes them less vulnerable to the failure of any single crop.

The third is the end of the great despotic regimes. Ó Gráda is explicit about this. Stalin, Mao, and Hitler actively weaponized hunger as an instrument of political control and demographic engineering. Their regimes are gone. Many of the 20th century’s worst famines were, in a very real sense, policy choices made by people who could make them without facing consequences.

The fourth is medical and logistical improvement: better understanding of nutrition, faster response capacity, more effective international aid organizations. But in addition to this, we have a much better sense today of what types of things will kill you if eat them and how to maintain public health in the face of famine. To quote Ó Gráda directly, since the beginning of the 20th century, “in places where infectious disease is endemic during normal times, it wreaks havoc during famines”, but in economies where, through vaccinations and sanitations systems these diseases are not endemic (think the Leningrad famine of 1941-42), deaths are much lower as a result.

Amartya Sen’s contribution to this analysis has been canonical since he first made it in the 1980s: no substantial famine has ever taken place in a functioning democracy with a relatively free press. The reasoning is structural. Governments that depend on populations for their survival cannot afford to let those populations starve. Democratic accountability turns famine into a political crisis for those in power, and that changes the incentive calculation fundamentally. The empirical record supports this: across post-colonial Africa, 59 percent of famine events occurred under closed autocracies, 36 percent under electoral autocracies, and just 4.5 percent under electoral democracies. The exceptions, where famines have occurred under nominally democratic governments, tend to involve weak press freedom, which is exactly what Sen’s theory would predict.

This is the good news. Famine, as a civilizational-scale event, is largely history so long as the above elements are maintained (and they have largely held).

Here is the bad news.

The same vulnerabilities that made food the defining political problem of the pre-industrial world have migrated to energy. The countries most exposed to the current Hormuz closure are not starving. But they are watching the price of everything spike, their logistics chains seize up, and their governments scramble for emergency measures. What economists call “demand destruction” in this context means everything from mandated work from home policies, limits to air conditioner usage, shifting schools and universities to online learning and limiting travel. The political pressure on leaders whose populations feel that pain is, structurally, the same pressure that pre-industrial rulers felt when the harvest failed. The mechanism is identical. Only the commodity has changed.

Subscribe now

The “Other” Explanation of Famine Decline

Ó Gráda’s account is persuasive, but to me it is missing a layer.

The four factors he identifies (information, development, political transformation, and logistics improvement) are all real. But they are downstream of a larger issue: the industrial revolution made food more abundant and easier to move, sure; but it also changed what food is to an economy.

For the first time in human history, the energy that runs an economy no longer flows primarily through the caloric chain. Coal replaced the horse. Oil replaced the ox. Electricity supplemented and then replaced human muscle in factory after factory. The energy that powers a modern economy comes from burning hydrocarbons or splitting atoms or capturing wind and sun, not from growing grain. Food is no longer the economy’s fuel.

The consequences of this shift are profound.

In the pre-industrial world, feeding the population and powering the economy were the same problem. A harvest failure cascaded immediately into economic collapse because there was no separation between the two. In modern economies, food and energy are largely separate systems. A drought in Kansas reduces wheat yields. It does not cut electrical power to Chicago. A failed harvest in Bangladesh raises food prices. It does not stop the container ships. The feedback loops that made historical famines self-reinforcing have been severed by the energy transition.

In my view, this is the reason that is upstream of all of Ó Gráda’s proximate explanations. Yes, better information helped. Yes, development helped. Yes, democratic accountability helped. But all of those factors were operating in a world where food had already lost its second job. Food no longer powered everything else. So when food supplies were disrupted, the disruption stayed contained. It did not cascade through the entire productive apparatus of society the way it once did.

Put differently: famine declined partly because food became less important. The management improvements of better information, better politics, and better logistics were real and necessary. But they were operating on a system that had already become structurally less fragile because food was now just food.

That last point sounds almost cruel. It isn’t meant to be. People still starve, and at this point in human history that is a moral catastrophe wherever and whenever it happens. Sen is right that the remaining famines are entirely preventable. The capability to prevent them exists. The obstacle is political will. But the structural reason that famines no longer collapse entire civilizations is that the civilizational energy supply no longer runs through food and that matters for other reasons (see below).

The implication for today should be obvious. Whatever form of energy now organizes the economy and underwrites military power occupies the structural position that food once held. It has food’s old vulnerability profile. And it will, unless managed very carefully, have some of food’s old failure modes as well.

Subscribe now

Three Lessons Famine Teaches About Energy

History does not repeat itself. But it rhymes with some regularity, and the rhyme between famine and modern energy disruption is close enough to be useful. Here are three things the history of famine teaches us about the energy crises we are in and the ones that are coming.

The government toolkit for energy crises is the same toolkit that made famines worse.

When governments faced famine, they reached, almost reflexively, for the same set of interventions: price controls to prevent profiteering, export bans to keep food in the country, forced requisition to redirect supplies to priority populations, national rationing schemes. These interventions were politically intuitive. They were also, with remarkable consistency, catastrophically counterproductive.

Adam Smith observed this about the Bengal famine of 1768-69: East India Company price controls on rice, intended to protect consumers, disrupted the market signals that would have drawn food into the affected region from areas with surplus. The French Revolutionary government’s maximum price decrees of 1793, applied to grain during a period of genuine shortage, caused bakers from surrounding towns to travel to Paris to buy subsidized flour, creating worse shortages in the city the policy was meant to protect. The Irish famine of the 1840s was actively worsened by British policies that required Ireland to continue exporting food to Great Britain even as people starved.

The clearest modern evidence comes from the 2007-08 food price crisis. Roughly three-quarters of the spike in global rice prices during that period was caused not by production failures but by the export restrictions that major producing countries imposed in response to rising prices. Countries tried to protect their own consumers by banning exports. The bans removed supply from global markets, drove prices higher everywhere, and intensified exactly the crisis they were meant to contain.

Look at the standard toolkit deployed when energy prices spike: price caps on electricity and fuel, restrictions on energy exports, emergency nationalization of energy assets, rationing schemes for industrial users. The economic logic of why these interventions fail in energy markets is exactly the same as why they failed in food markets. They suppress the price signals that would draw additional supply into the market, reduce the incentive for investment in new production, and create distortions that outlast the crisis. The governments that resisted these interventions during famines, or that reversed them quickly, consistently produced better outcomes than those that doubled down.

Subscribe now

Transportation and storage are still the main constraints.

Here is the insight from famine history that I think matters most for the current moment: famine was rarely, if ever, about a global shortage of food. There was almost always food somewhere. The problem was getting it from where it was to where people were dying.

Ó Gráda documents cases where populations starved within a few hundred miles of food surpluses, because the infrastructure to move it did not exist, was too expensive, or was controlled by parties with no interest in deploying it. The Irish during the famine were within sailing distance of enormous grain stores in Britain and North America. Global supply was rarely the limiting factor. Local access, purchasing power, and the political will of those who controlled the distribution system determined who ate and who starved.

Falling transportation costs, better logistics networks, and faster communication were among the primary reasons famine declined in the 20th century, precisely because they attacked the binding constraint directly.

Now look at Hormuz.

The world does not face a shortage of oil and gas. Saudi Arabia has production capacity sitting idle. The United States is the world’s largest oil producer. Canada, Norway, and others have more. The problem is that approximately 20 percent of world petroleum consumption, the portion that was transiting the Strait, cannot currently move. The IEA has activated a record 400 million barrel release from strategic stockpiles. But bypass infrastructure, the Saudi Petroline pipeline and the UAE’s Fujairah route, can handle at most 7 million barrels per day at theoretical maximum. Normal Hormuz throughput was 15.8 million barrels per day. The arithmetic is unambiguous: even with every bypass route running at full capacity, the structural supply deficit exceeds 10 million barrels per day that cannot be addressed through alternative routing alone.

This is the Strait of Hormuz as pre-industrial transportation bottleneck. The oil exists. The road is blocked, the alternative roads cannot carry enough traffic, and the people who need it are on the wrong side of the obstruction.

The European natural gas crisis of 2021-22 makes the same point from a different angle. When Russia cut approximately 80 billion cubic meters of gas deliveries to the European Union, global natural gas supply did not disappear. LNG terminals in Spain and Portugal had substantial capacity. The problem was that France’s internal north-south pipeline network was insufficient to move gas from Iberian import terminals to central Europe. Gas that physically existed, in terminals that were operational, could not reach the consumers who needed it because the connecting infrastructure was not there. A plumbing problem, pure and simple.

The lesson is direct: in both food and energy systems, the strategic investment that matters most goes into the redundancy and resilience of the infrastructure that moves supply from where it is generated to where it is consumed. Production capacity without transmission capacity is energy that happens to exist somewhere you cannot always reach. That is a very different thing from energy security.

Subscribe now

Acts of God operate differently depending on what you are trying to protect.

Pre-industrial crop failure was, in one sense, egalitarian in its destructiveness. Drought hit an entire region. Blight moved across borders. The harvest failure that triggered a famine did not discriminate between one farm and another in its path. Everyone in the affected area was exposed.

Modern energy systems have a very different vulnerability profile, and the differences matter enormously for resilience planning. The 1859 Carrington Event, the largest geomagnetic storm in recorded history, disrupted telegraph networks across North America and Europe, set telegraph paper on fire, and gave operators severe electric shocks. It had essentially no effect on the coal-fired steam economy. The coal was still in the ground. The steam engines still worked. The industrial apparatus was untouched, because it did not run on the electromagnetic infrastructure the storm disrupted.

A Carrington-scale event today would be a different matter. Modern high-voltage electrical grids are acutely vulnerable to geomagnetically induced currents, which can cause transformers to overheat and fail. Some of the same issues arise with flooding, earthquakes, and wildfires. The cascade risk is severe: a disruption in one region shifts loads to adjacent regions, which then fail under the additional burden, and so on. Liquid fuels do not face this specific threat profile. A pipeline is not a power line. An ICE car or truck does not depend on grid stability to move. An EV does.

None of this means electrical energy systems are inferior to liquid fuels in general. They have enormous advantages. But it does mean the vulnerability profiles are different, and resilience planning that treats all energy systems as interchangeable is making a mistake that famine history would warn against. The question that actually matters is how robust your energy supply is to the specific failure modes that can occur. Volume is secondary to that.

Subscribe now

What This Means for the Next Twenty Years

Part 1 closed with the claim that the most important competition on earth is for energy and what it means directly for geopolitical risk. Here in Part 2 I’ve tried to add some precision to what that claim actually means in the context of shortages.

Famine was tamed by a combination of factors that all had to work together. No single one was sufficient. Better information mattered. Economic development mattered. Political transformation, specifically the end of regimes that actively weaponized hunger, mattered enormously. And the underlying structural shift, food’s disaggregation from the economy’s energy supply, provided the foundation on which all the other improvements could build. Remove any one of those factors and the improvement stalls.

The modern energy competition is similarly multi-causal. The question is who has the most resilient generation capacity across multiple failure modes; the most redundant transmission infrastructure connecting producers to consumers; the political arrangements that make the deliberate weaponization of supply politically costly; and the information systems to allocate energy efficiently during disruption.

Ó Gráda’s emphasis on information deserves special attention here, because it is probably the least discussed of the four factors in most energy security analysis. Better information was essential to famine’s decline for two reasons. First, it removed the cover under which deliberate starvation operated. Regimes that used hunger as a weapon depended on information control to do so. When that control became harder to maintain, the political cost of weaponizing food supply rose. Second, better information allowed markets to function more efficiently. The price signals that draw food from surplus regions to deficit regions only work if traders know where the surpluses and deficits are.

Neither of those functions has a clear equivalent in modern energy markets. Energy infrastructure is vast, complex, and largely opaque to the public and even to many policymakers. Supply and demand signals are transmitted through markets that are themselves subject to manipulation. The energy equivalent of “you cannot hide a famine in the information age” has not arrived. It should be a policy priority that it does. Real-time, transparent information about energy production, storage, transmission capacity, and consumption, accessible to markets and policymakers simultaneously, would do more to reduce energy vulnerability than almost any other single investment.

Subscribe now

At Least Four Reasons This Analogy Might Be Wrong

I made the same commitment in Part 1 that I am making here: honest objections, honestly engaged. Here are the four that I find most compelling.

The thesis misreads the famine literature. Ó Gráda doesn’t make the food-energy disaggregation argument. Neither does Sen. The thesis that food stopped being the economy’s energy source and that this structural shift explains famine’s decline is the your own interpolation, not a finding from the research. Sen’s entitlement approach explains famines as failures of the legal and economic means to acquire food; it says nothing about energy coupling. Ó Gráda’s four factors are all proximate to each other and he is explicit that famine prevention success is contingent and fragile. Alex de Waal, another leading scholar in the field, has spent decades arguing that what looks like structural progress in famine prevention is actually politically contingent, reversible, and dependent on international norms that are themselves under pressure. None of them saw the energy connection. The piece is building an analogy on a reading of the literature that the literature’s own authors didn’t see and might in fact be likely to reject.

This objection is really more about authority than anything else. Ó Gráda doesn’t make the disaggregation argument, and I am not claiming that he or any other famine scholar does. I am aware that I’m making a claim that connects two areas that are not in communication with one other though they could have been. My claim though is not a major leap I think: that Ó Gráda correctly identifies four proximate causes of famine decline, and that beneath those proximate causes is a structural shift in energy that his analysis doesn’t explicitly address. Whether that structural shift is the underlying cause of the ones he identifies is a fair question, but I don’t know that it could be empirically tested (another point against my argument). It certainly doesn’t contradict Ó Gráda he didn’t ask the question. On de Waal: the fragility argument applies to political will, which is Sen’s category, and I accept that. What the fragility argument doesn’t address is whether the structural coupling between food and the energy requirements of the rest of the economy is what it was in 1850. Even a deliberately engineered famine today operates in a world where food has lost its double task - food is just food, not energy for the whole economy. That limits how far it cascades regardless of political intent. The Holodomor killed between 5.7 and 8.7 million people and did not collapse the Soviet system. The Irish famine killed roughly a million people and devastated Ireland’s economy for a generation. The structural difference between those two cases is partly explained by industrialization, i.e., by how much of the economy had moved off the food-energy nexus by 1932 compared to 1845. That isn’t Ó Gráda’s argument and I’m okay with the fact that I’m making an argument that is an extension unintended by the original researchers. In my view, that’s the only way to view the world how it is and not in a reductionist manner. Put another way, just because the original empirical researchers didn’t see the forest, doesn’t negate its description; they are focusing on trees after all.

Subscribe now

Food production improvements are as entangled with energy as they’ve always been. The real explanation for famine’s decline, on this reading, has nothing to do with food-energy disaggregation. It is that the industrial revolution, and specifically the Green Revolution of the twentieth century, expanded agricultural productive capacity faster than population grew, permanently relaxing the food constraints. Norman Borlaug’s high-yield varieties, synthetic nitrogen fertilizer, mechanized irrigation: these broke the production ceiling of food that previously existed. If this is right, famine declined because humanity got dramatically better at producing food not because food changed its role in the economy. And if this is right, the dissaggregation claim breaks down because energy is a core input for food.

I actually think this makes my argument stronger, not weaker. If the same forms of energy are required to make food and power the economy but those are NOT food, that makes that form of energy MORE important than food, not less. The Haber-Bosch process, which synthesizes ammonia from atmospheric nitrogen and which sustains roughly half the world’s current population, runs on natural gas. Synthetic nitrogen fertilizer is a fossil fuel product. If the argument is that better food production technology explains why famine lost its civilizational destructiveness, that technology runs on exactly the energy supply I’ve been talking about. The food-energy disaggregation and the Green Revolution are certainly more entangled than I might be giving them credit for, but, in the pre-industrial era, that relationship was reversed because food was the energy for everything else. Now, energy is the food for food (if you get my meaning).

Subscribe now

People die without food. They don’t die without power.

The asymmetry between food and energy is biological and immediate. Starvation kills. A power outage is uncomfortable. The political urgency of famine is in its capacity to destabilize governments, generate mass population movement, produce civilizational restructuring and a moral catastrophe which all derive directly from that biological urgency. Energy disruptions generate economic pain and political pressure over months and years. They do not produce the same death timeline, which means they do not generate the same political forcing function. Famines take down entire civilizations. While energy crises might produce stagflation, political realignment, and some bruising electoral outcomes, no regime or constitutional order or civilization collapses. The stakes are categorically different, which undermines the claim that the same lessons transfer.

My overall point is that, when famine led to civilization collapse it was because of lack of food and energy. That an energy crisis has yet to cripple a civilization isn’t an argument against so much as a recognition that this is a fairly new phenomenon. We don’t actually know what happens when a country runs out of energy inputs.

The objection is correct though that the death timeline is different. In the modern food system, energy disruption does become food disruption faster than this objection assumes. The Haber-Bosch process alone accounts for the nitrogen fertilizer that sustains roughly half the world’s population and it runs on natural gas. Mechanized harvesting, refrigerated storage, food transportation: all of it runs on energy. The Texas winter storm of February 2021 is the clearest recent illustration of the cascade. The energy failure became a water failure within hours. Food access collapsed for millions of people within 72 hours. The most vulnerable Texans died not from starvation, but from the compounding effects of a system that had assumed continuous power. More slowly than famine, through more indirect pathways, but the cascade was structurally identical. The longer the energy disruption, the more it converges with exactly the dynamics the objection says it lacks. The biological asymmetry is real. The political and economic asymmetry is smaller than it looks.

Subscribe now

Food and energy are too different for this analogy to hold.

Food has one end use: human consumption and a specific biological urgency that only water and air share. Energy has dozens of end uses and no equivalent biological urgency at the point of supply. More importantly, energy is substitutable across sources in ways that food is not: when natural gas supplies tighten, utilities switch to coal or nuclear or wind; there is no equivalent switch from wheat to iron ore. Food production is geographically dispersed in ways that create a different geopolitical structures: there is no food equivalent of Saudi Arabia, no country that produces 20 percent of global calories the way a handful of states produce 20 percent of global hydrocarbons.

This is the objection I find hardest to fully answer, and I want to be honest about that. Two concessions and one rebuttal. Concession one: energy’s substitutability does mean that energy crises are structurally less severe than historical famines at comparable disruption levels. I probably haven’t fully accounted for how substitutability changes the dynamics, and I am still working through the implications. Concession two: the geographic concentration argument cuts both ways. The concentration of hydrocarbons in a small number of politically volatile regions is itself an argument that the energy analogy is more alarming than the famine analogy because food disruptions were typically regional, while a Hormuz closure affects global supply simultaneously. At the same time, in the past, we certainly understood the concept of “breadbaskets”, places that, because of their unique geographies, produced the vast majority of a region’s food (think the Nile River valley or large swaths of Ukraine). That is not a weakness in the framework; it is a reason the framework may understate the risk. My claim is that whatever resource is most tightly coupled to both population welfare and overall productive capacity will generate the same class of political dynamics: strategic competition for control, government intervention failures, transportation as the binding constraint, and information opacity as a force multiplier for bad actors. The operational differences between food and energy don’t determine whether the structural position they occupy generates those dynamics. The question is whether energy now occupies that structural position. The piece argues it does. I think the evidence supports that argument. I am genuinely uncertain about the degree.

Subscribe now

Why This Framework Matters for Geopolitical and Business Risk Analysis

I am not writing this as a purely academic exercise.

If the lessons of famine scholarship can transfer to modern energy (and this piece has argued they do), then several things follow for how executives, investors, and policymakers should assess risk.

First, transportation and chokepoint exposure is likely systematically underweighted in standard risk frameworks. Most geopolitical risk analysis focuses on production capacity, reserve levels, and supplier diversification. The historical record on famine is clear that transportation is almost always the main constraint, not total supply. Ten suppliers sharing a single transit chokepoint is a single point of failure with extra paperwork. The Hormuz closure is demonstrating this right now. There is no global shortage of oil. There is a blockage at one transit point, and production capacity sitting idle in Saudi Arabia, the UAE, and Qatar is irrelevant to the crisis. The right question to ask of any supply chain is “how many independent routes to market do suppliers share?”

Second, governments will often reach for the wrong tools first, and their interventions are likely to extend the disruption before they improve it. The historical consistency here is striking. From the Bengal famine to the French grain crisis to the 2007-08 food price spike to the EU gas crisis, the standard first-order government responses are price caps, export restrictions, and emergency requisitions, which all made the underlying problems worse. Companies that assume government stabilization as their primary risk mitigation mechanism are systematically underestimating their exposure. The more defensible planning assumption is that official intervention will amplify the short-term disruption, and that recovery will depend on market mechanisms that took longer to engage than anyone in authority was willing to wait for. Organizations that build that assumption into their continuity planning will be better positioned than those that expect government intervention to solve the problem quickly.

Third, electrical and liquid fuel systems have genuinely different risk profiles, and most corporate resilience frameworks treat them as equivalent. A company that has stress-tested its liquid fuel exposure may not have stress-tested its electrical exposure )and vice-versa). The failure modes, the recovery timelines, the available workarounds, and the cascade effects are different in kind, not just degree. Electrical systems can go to zero in ways that liquid fuel systems cannot, and they can stay there for reasons that have no analogue in liquid fuel supply chains like transformer lead times, single-source steel dependencies, and cascade failure dynamics across interconnected grids. The specific failure mode determines everything: who is affected, how long the disruption lasts, and what options remain. Resilience planning that treats energy as a single undifferentiated risk category will be caught unprepared in one direction or the other. Companies need to map out that exposure facility by facility. (As a brief aside, companies that have done extensive Scope 1, 2, and 3 carbon emissions studies and reporting likely have all of this data already; they just need to stress-test it).

Fourth, and this is the point I most want to leave with anyone managing exposure to politically volatile regions: there is no Sen theorem for energy. Sen established that no substantial famine has ever taken place in a functioning democracy with a relatively free press. The mechanism is political accountability: governments that depend on populations for their electoral survival cannot weaponize food without facing consequences that threaten the regime itself. No equivalent constraint operates on actors who weaponize energy supply. Russia cut gas to Europe and remained a functioning state. Iran has closed the Strait and the successor regime has stated the closure should continue. OPEC has manipulated supply for decades without facing the kind of political transformation that ended the weaponized famines of the twentieth century. The countries that currently hold the most leverage over global energy supply are not, as a group, moving toward greater democratic accountability. Standard political risk frameworks are implicitly calibrated to a world where the dominant scarce resource was subject to the famine-era accountability dynamic. Energy supply is not. The floor that Sen’s theorem implies does not exist here. On the flip-side, we have also yet to see electricity cuts as a means of internal political punishment for instance the way Stalin used famine to punish Ukrainians. Organizations whose operating models depend on stable energy access in regions where that leverage is concentrated are carrying a risk category that most standard frameworks rate inadequately.

Subscribe now

The Optimistic Reading?

Famine’s near-disappearance deserves more recognition than it usually gets. It was earned. It required structural changes in the economy, political transformations in how states relate to their populations, massive improvements in information and logistics, and the deliberate work of thousands of researchers, aid workers, and policymakers across several decades. None of it was inevitable. All of it required choices.

The question now is whether we apply those lessons deliberately to our core sources of (non-food) energy, or stumble into the equivalent outcomes of famine reactively.

The disaggregation of food from energy was a background condition of the industrial revolution. It happened. Nobody planned it. Societies discovered its implications over time and built institutions to manage them. The disaggregation of strategic military power from human bodies, which Part 1 argued is now underway, is also happening rather than being designed. The consequences are already visible.

What is different now, and what gives me something that functions as cautious optimism, is that we have a historical template. We know what made the previous version of this problem better. We know that information transparency mattered. We know that political accountability mattered. We know that market infrastructure, built in advance of the crisis, mattered. We know that transportation redundancy mattered. We know that the regimes most willing to use the dominant resource as a weapon were eventually constrained, though not quickly and not without enormous suffering.

The Hormuz closure is producing the kind of pain that forces choices. Strategic petroleum reserve releases. Emergency diplomatic efforts. Accelerated investment in alternative supply routes. Political pressure on the governments most exposed. Some of those responses will work. Some will not. What the history of famine suggests is that the ones most likely to work are the ones that build durable infrastructure and durable accountability, not the ones that suppress the price signal in the short term.

We have been through this before. Not with oil. Not with electricity. But with food, and with the same underlying structure. Those lessons are worth remembering.

It is not.

Here’s the rule for business in this crisis, and it’s simple: watch what happens not what anyone says will happen. The Strait of Hormuz is open when ships are passing through it safely and consistently. Not when the parties say they’ve agreed to a deal. Not when a deal is signed. Not when a press release is issued from Washington, Tehran, Doha or Geneva. Not when anyone announces that it’s done. When ships are actually transiting safely and regularly.

Until that moment, nothing changes operationally. Nothing.

This is the reality and the numbers back it up.

As of June 14, two ships were transiting the Strait daily. Under normal conditions, that number is 94. Roughly 1,550 vessels are stranded right now, with approximately 22,500 mariners trapped in regional waters. Six Maersk ships are sitting inside the Gulf waiting. The UAE’s state-owned oil company, with more at stake in this than almost any other actor, estimates full flows through Hormuz will not resume until 2027, even under a rapid deal scenario. Traders, who price these things rather than just discuss them, are putting 38% odds on traffic returning to normal by July 1.

So what’s actually in the MOU? It says the Strait reopens “immediately” upon signing and that prewar shipping conditions will be restored within approximately 30 days. That’s the document. The document also exists alongside Iran’s recent history of issuing new passage conditions as recently as April, creating enough confusion that Euronews ran a piece headlined “Confusion reigns over Hormuz shipping reopening.” Iran is currently pushing differing versions of the deal text. The U.S. Secretary of State has said any tolling system on the Strait “would make a diplomatic deal unfeasible.”

Again, signed “deals” and an opened Strait are two different things.

Here’s the rub. Business decision-makers right now face a specific and concrete choice: do we reverse the supply chain rerouting we’ve done over the past 100 days? Do we start building plans around Hormuz flows resuming?

My answer is no (at least not yet) for two reasons.

First, the rerouting you’ve done is a sunk cost and now functions as hard-won operational knowledge. Your teams now know the alternative routing that works. They understand the additional mileage, or the 10-14 extra days, or the 20-35% fuel cost premium. They’ve built relationships with transhipment hubs like Salalah. The major carriers like Maersk, CMA CGM, and Hapag-Lloyd still have suspended Hormuz transits and have not announced reversal. Before you unwind any of that, you need the Strait to demonstrate that it is open, not just declared open.

Second, the trapped ships matter more than the signed papers at this point. There are vessels inside the Gulf right now whose transit through the Strait will be the real test of whether this deal holds operationally. Watch those ships. If they move freely, consistently, over a period of weeks, that is your signal. A signing ceremony in Switzerland cannot tell you what the IRGC will actually do at the waterway. The first convoys of vessels transiting without incident will tell you more than any MOU.

I’ve spent twenty years studying how businesses operate in conflict zones and through crises. The pattern I’ve seen repeated in situations like this is that the announcement of anything between conflict parties (a ceasefire, a peace deal, a diplomatic breakthrough, whatever) creates enormous pressure to normalize for companies. To take the foot off contingency planning. To revert to whatever was happening pre-conflict.

Time only moves in one direction. There is no real “going back”. It will never be what it was. Whatever changes you make now that the change is within your control, make them consciously, with evidence, not in response to a press conference or social media post.

Your company should use two evidence-backed questions before your company makes any supply chain reversal decisions:

Question 1: Are ships, particularly those currently trapped inside the Gulf, transiting the Strait of Hormuz safely and without incident? Not one ship. Multiple ships, over multiple days, without coercion, tolling, seizure, or harassment.

Question 2: Have the major carriers, who make routing decisions based on insurance, risk assessment, and operational intelligence, officially announced resumed Hormuz services with specific sailing dates? Until Maersk or Hapag-Lloyd publishes a schedule with “via Hormuz” on it, the strait is not open for business.

A deal signed in Switzerland is good news. It’s genuinely, meaningfully good news, particularly for the 22,500 mariners currently trapped in this crisis who deserve to go home. But good news is does not necessarily meaningfully change the operating environment. The Strait of Hormuz is open when ships are transiting it. Until then, keep your supply chain where it is and watch the water.

A government contract here, a dual-use reclassification there, a request to modify an existing product for a military customer. Before long, you are building capability for a customer you did not strategically choose, in a risk environment you did not price, operating under legal and regulatory frameworks you are still working out. The European defense boom is accelerating this drift at a scale not seen in a generation.

The numbers tell the scale of the pull: €800 billion under the Rearm Europe plan, EU member states collectively spending €343 billion on defense in 2024 alone, up 57 percent from 2021. Greece alone has committed to a €25 billion modernization program over twelve years, the largest in its modern history. Politicians across the continent have discovered the word “sovereignty” and are using it to justify almost everything. Seemingly every company wants to be a defense company now in part because defense budgets are untouchable.

The conventional wisdom is that all of this is overdue. Europe underspent on defense for decades and is now catching up. That story is mostly right. But the geopolitical environment driving the boom is not simply a story of insufficiency being corrected. It is one of structural transformation. As I argued in my recent piece on the energy thesis, AI and drones are rewriting the cost structure of warfare in ways that make the boundary between civilian and military infrastructure nearly impossible to draw. The company that built your GPS receiver, your encrypted communications platform, or your agricultural drone fleet is increasingly indistinguishable, in the eyes of procurement offices and adversaries alike, from a defense contractor. That has real consequences.

Three of those consequences are not being talked about seriously by companies inside the boom. Here they are.

Subscribe now

Problem One: “Dual-Use” Is Becoming Meaningless, and That Has Consequences

The term “dual-use” has a precise legal meaning in European export control law. It refers to goods, software, or technology that can be used for both civilian and military purposes. The classic examples are GPS technology, certain chemicals, and high-performance computing. The logic is sound: some things genuinely sit on the boundary.

That was then. The EU’s current dual-use regulatory framework is, as analysts at the European Policy Centre put it, “rooted in a binary logic that recent developments have eclipsed.” In practice, the category is expanding fast, in both directions, and in ways that create serious problems.

On one side, things that are clearly military are being laundered as dual-use to access civilian research funding and avoid the political complications of pure defense contracts. The European Defence Fund received an additional €1.5 billion for 2024-2027 explicitly to support dual-use and defense R&D, and the message, as one analysis put it, is that dual-use innovation in Europe is “no longer balanced between civilian and military aims but is tilting firmly toward the defense end of the spectrum.” That is a polite way of saying the label is being used strategically to avoid harder due diligence.

On the other side, things that are clearly civilian are being pulled into defense pipelines under the same label. Drones originally built for agricultural surveys, encrypted communications software built for journalists, satellite imaging services built for climate monitoring: all of these are now routinely classified as dual-use, which means their developers can find themselves subject to export controls, mandatory government disclosure, or quietly reclassified as defense contractors without ever choosing to become one.

Classification is the rub because it determines everything: how you are regulated, what contracts you can sign, what your liability exposure looks like, and, critically, who considers you a legitimate target. The EPC is already calling for a move from “dual-use” to “omni-use“ frameworks to reflect this reality. But the regulatory reform is running far behind the actual reclassification happening on the ground.

The 2025 update of the EU Control List added new “500 series” classification numbers for quantum computers, advanced semiconductor manufacturing equipment, and high-temperature coatings. The underlying logic of all of this is reasonable. But every expansion of the classification list is also an expansion of the population of companies that are now, effectively, in the defense business whether they wanted to be or not.

Subscribe now

Problem Two: Your Civilian Customers Have Values. Your Government Might Not Share Them.

Greece is the perfect case study for what happens when government defense policy and civilian market values collide.

Start with the strategic picture. Greece is one of NATO’s highest spenders relative to GDP, sitting at 3 percent in 2025. Athens has committed to a €27 billion defense overhaul centered on high-tech warfare. It is acquiring F-35s, building the Achilles Shield air defense system, and deepening a trilateral defense relationship with Israel and Cyprus that includes joint air and naval exercises and deep industrial collaboration.

In April 2026, Greece signed a €650 million deal with Elbit Systems for PULS rocket artillery systems. The Greek government also agreed to cooperate with Israel on anti-drone systems and cybersecurity. The bilateral relationship, as the Jerusalem Post noted, had by 2025 moved “beyond joint exercises into deep industrial and intelligence collaboration.”

Now look at what is happening at the civilian level. Greek dockworkers have been blocking Israeli cruise ships from docking. Pro-Palestinian protesters threw red paint on mock bodies during demonstrations specifically targeting Souda Bay naval base, calling for Greece to sever its military ties. A protest campaign literally named itself “Target Souda Base for Palestine.” The Times of Israel observed drily that “Greece’s pro-Israel turn remains contested at home, with incidents revealing a persistent gap between government policy and popular sentiment.”

That gap is not unique to Greece. It is the central reputational challenge facing any European company that decides to enter the sovereign defense space right now.

Consider what happened to Elbit in the United Kingdom. After sustained direct-action campaigns by Palestine Action, Elbit reportedly shut down its Bristol facility, which had been repeatedly occupied and covered in red paint. More telling than the protests themselves: both Allianz and Aviva terminated their insurance contracts with Elbit’s UK subsidiaries. Insurance companies, whose entire business model is based on calculating and managing risk, looked at the situation and decided the reputational exposure was not worth it.

This is the structural problem for any civilian-facing company that wants to enter the European defense market. Your B2C brand, your ESG commitments, your sustainability reports, your university partnerships, your tourism marketing: all of these exist in a world where the Gaza conflict has made public attitudes toward Israeli defense technology, and by extension anyone associated with it, deeply polarized. The government signing the contract and the citizens buying your products are operating in fundamentally different moral universes.

Some companies will say: we are a defense company, and defense is what governments do, not citizens. Fine. But that position requires consistency and clarity from the start. The dangerous position is the middle one: accepting defense contracts while still trying to maintain a civilian brand identity, which is precisely what most companies caught in the dual-use expansion are doing. You cannot be a peacetime tech company and a wartime defense contractor simultaneously without paying a potentially serious reputational price, usually at the moment you can least afford it.

Subscribe now

Problem Three: When You Join the Defense Industrial Base, You Inherit Its Risks

There is a reason that serious defense primes have entire divisions dedicated to security, redundancy, and physical protection. It is because they are targets.

This is not hypothetical. Russian state-sponsored hackers have been systematically targeting civilian companies that support Ukraine’s defense effort. Everything from transportation companies and logistics operators to IT service providers and air traffic control systems have been at physical risk. The targeting criteria is not “are you a weapons manufacturer?” It is “are you doing something militarily useful for the other side?” If you are providing encrypted communications, satellite imaging, logistics software, or drone components, you have become a potential target.

The cyber exposure is the most visible form, but it is not the only one. Physical attacks on defense-adjacent facilities are increasing across Europe. Supply chain disruption as a form of hybrid warfare is now standard doctrine for Russia and its proxies. The attack surface for any company that enters the defense ecosystem is dramatically larger than it was before, and the attackers are sophisticated state actors, not script kiddies.

Then there is the financial exposure that nobody discusses in the enthusiasm of new defense contracts.

The history of wartime defense procurement is not a history of corporate windfalls. It is a history of profit caps, cost-plus contracts with government auditors sitting in your offices, emergency nationalization, and excess profits taxes. The current European defense buildup is happening in peacetime for Europe (sit with that thought for a second), and the current contract structures largely reflect peacetime economics. But Europe is explicitly planning for conflict scenarios. The EU Defense Readiness Roadmap 2030 is a mobilization plan.

In actual wartime, governments do not ask suppliers nicely to accept lower margins. They legislate them. The windfall profits tax framework that Europe applied to energy companies during the 2022 energy crisis, which generated €26 billion in emergency revenue, is the template. Those taxes have already been extended and expanded, now reaching oil and gas companies in 2026. The legal and political infrastructure to apply them to defense contractors in an escalation scenario is already in place. What that means for a mid-sized European technology company that has just restructured its business around a ten-year defense contract is, to put it mildly, uncertain.

Federal acquisition rules in the United States similarly cap defense contractor profits at 10 to 15 percent depending on contract type, and even that ceiling is contested. The European equivalents, now being built at speed, are likely to be no more generous. The companies entering the defense market today on the basis of current contract economics need to model what those margins look like if the political situation escalates and governments start treating defense procurement as a public emergency rather than a commercial transaction.

Subscribe now

A Preview: What Happens When Governments Decide You’re in the Defense Business

You do not have to imagine this future. It arrived today.

On June 9, 2026, the Pentagon published an updated “1260H list” of Chinese military companies. The new additions included Alibaba, Baidu, and BYD. Alibaba is China’s largest e-commerce platform. Baidu is its dominant search engine. BYD makes electric cars.

None of these companies manufacture weapons. None of them are defense primes in any conventional sense. The Pentagon’s designation rests on a different standard: the US government determined they were “aiding China’s armed forces,” a category broad enough to capture civilian companies whose technology, infrastructure, or business operations are judged to contribute to military capability.

The immediate consequences are concrete. The Defense Department is prohibited from contracting directly with listed companies. From June 2027, it will be prohibited from procuring their products or services through third parties. All three companies disputed the designation and pledged to seek their removal. China’s embassy in Washington condemned the listing as “discriminatory” and an example of the US government overstretching the concept of national security.

That last phrase deserves attention. “Overstretching national security” is exactly what the dual-use expansion looks like from the outside. The question is who gets to define the stretch.

Now read that story again and substitute European names. A European AI company that provides logistics optimization software to a NATO member’s defense ministry. A European satellite imaging firm whose data feeds into a government’s border surveillance system. A European drone manufacturer whose products are explicitly civilian but whose technology has been reclassified under the EU Control List’s new “500 series.” All of them are now in a position structurally similar to Alibaba’s: civilian companies whose technology a government has decided is militarily relevant, whether or not they chose to be in that business.

The BYD case is particularly instructive because BYD’s primary exposure is not in software or AI or surveillance, but in electric vehicles and batteries. The Pentagon’s logic is that EV battery technology has direct applications in military vehicles and equipment. If that logic holds, the same argument applies to every European battery manufacturer, every grid-storage company, and every semiconductor firm whose chips end up in both consumer products and military systems.

This is not a hypothetical risk for European companies watching events from a distance. It is the operating environment they are already in. The designation of Alibaba, Baidu, and BYD is the clearest signal yet that governments do not wait for companies to make a choice about whether to be in the defense industrial base. Sometimes they make that choice for you.

Subscribe now

What Managers Actually Have to Decide

I have spent twenty years talking to people who run companies in places where security is not guaranteed: conflict zones, fragile states, high-risk environments where the rules can change overnight. The lessons are largely the same. The companies that navigate these environments well are the ones that made the most honest decisions about who they were, what they believed in, and what they were willing to become.

The European defense boom is creating enormous pressure to answer that question without really asking it. Governments are waving contracts. Politicians are calling it sovereignty. Defense budgets are being described as economic development programs. All of that is true, and none of it resolves the underlying question for any company that is not already, unambiguously, in the defense business.

That means being honest about classification. If your product is genuinely dual-use, you have some maneuverability. If it is primarily civilian and you are accepting a military reclassification for the contract, you are taking on risks you may not have fully priced.

It means being honest about your customer base. If you are selling to consumers, to universities, or to civilian institutions whose values are in tension with the government defense policy you are now supporting, you need a clear answer to the question that will come.

It means being honest about the operational environment. If you accept a defense contract, you become a potential target. Your cybersecurity posture, your physical security, your supply chain resilience need to reflect that, not your previous peacetime baseline.

And it means being honest about the financial scenario. The contract you are signing today is a peacetime contract. Model what it looks like if the situation it was designed to prepare for actually happens.

Greece is instructive precisely because it is doing all of this at the government level with unusual clarity. Athens has decided, quite explicitly, that strategic security takes precedence over popular sentiment about its Israeli partnerships, that military modernization is worth the price of domestic protest, and that being in the defense business means being in the defense business fully. You can disagree with that calculation. You can disagree with those values. But at least it is a calculation and at least the values are clear.

For companies being swept into the sovereign defense wave without making an equivalent decision about it, the risks are piling up. The dual-use label is expanding beyond its intended scope. The reputational exposure is real and, in some cases, has already proven fatal to operations. And the financial assumptions underlying current defense contracts may not survive contact with actual conflict.

The boom is real. The enthusiasm is understandable. But enthusiasm is not a strategy.

This series examines industries one at a time through a single question: what does the way an industry behaves under geopolitical stress reveal about how it structured itself under peace? (the first of the other series on my “operating theses” can be found here)

Armed conflict is the clearest version of that stress, which is why this series begins there. But it is not the only version. An export licensing rule, a raw-ore export ban, a tariff regime, a court ruling that shutters a mine, a sovereign severing a supply route: each shifts the political ground under a business that optimized itself under other operating assumptions. War makes the failure obvious. It does not make it unique. The structural vulnerabilities that conflict exposes, geographic concentration, thin margins, deferred resilience investment, and the absence of decision frameworks for politically complex environments, are the same vulnerabilities that a licensing office in Beijing or a protest movement in Panama can expose just as completely.

There is a comfortable fiction that runs through a lot of corporate risk literature: that conflict is an exogenous shock, a temporary disruption to be absorbed before normal operations resume. Pharmaceutical companies are among the industries that have most visibly exposed the silliness of that fiction. They operate in conflict zones because they must. Their products sit at the intersection of commerce, ethics, and international law in ways that no other sector quite replicates. And the decisions they make under fire reveal structural vulnerabilities that most C-suite leaders have been actively avoiding.

This is the first in a series examining how specific industries navigate armed conflict. Pharmaceuticals come first because the stakes are highest, the dilemmas most instructive, and the gaps between stated policy and operational reality most illuminating.

Subscribe now

The Humanitarian Shield and Its Limits

When Russia invaded Ukraine in February 2022, hundreds of multinationals faced the same binary: stay or leave. The financial sector largely exited. Technology companies followed. Airlines pulled routes. Big Pharma (mostly) stayed.

The argument was principled, and largely defensible. Drugmakers, medical device manufacturers, and healthcare companies were exempted from U.S. and European sanctions, and contended that international humanitarian law required them to keep supply chains open. A 2025 study in BMC Medical Ethics analyzed public statements from 19 major pharmaceutical companies and found that the majority adopted a “scale back, don’t exit” posture, citing treatments for cancer, heart disease, and diabetes as the basis for continued presence. Eleven of the 19 offered explicit values-based reasoning for their decisions.

But the humanitarian shield is a double-edged sword. The same exemption that allowed Novartis to keep insulin moving to Russian diabetics also allowed revenue to flow into a war economy. Yale’s Jeffrey Sonnenfeld argued that pharmaceutical companies citing humanitarian necessity were “being misguided at best, cynical in the medium case, and outright deplorably misleading and deceptive,” noting that banks and technology companies also provide essential services.

The pharma sector’s response was genuinely differentiated, not uniformly cynical. Companies including Pfizer, Bayer, and Sanofi deferred new investments and development in Russia. AbbVie suspended its aesthetic products operations and paused new clinical trials. AbbVie, along with Merck, Pfizer, Eli Lilly, and GlaxoSmithKline, announced they would donate profits from all essential medicines sold in Russia directly to humanitarian relief efforts in Ukraine.

What the Russia-Ukraine episode produced was a taxonomy of positions when faced with conflict (at least one side of it): full continuation, selective scale-back, revenue redirection, or staged exit. No major pharmaceutical or medical device maker announced plans to shutter plants or halt sales entirely inside Russia.

For executives in other industries watching this unfold, the lesson was not that pharma got it right. The lesson was that pharma had no several options, none of which were fully satisfying to all stakeholders, and the decision framework they used mattered more than the specific outcome.

Subscribe now

Gaza: When Access Itself Is the Weapon

The Ukraine case involved a commercial decision complicated by ethics. The Gaza case involves something structurally different: the deliberate use of medicine denial as an instrument of conflict.

On March 2, 2025, Israel intensified its blockade on Gaza and almost completely closed the border crossings, including Kerem Shalom, Zikim, and Erez. Warehouses became unable to receive new shipments of medicine or medical equipment. By late 2025, Gaza’s Ministry of Health reported that 321 essential medicine items had reached zero stock, representing a 52% shortage on the essential medicines list. Medical consumables were worse: 710 items at zero stock, a 71% shortage. The most critical gaps hit emergency services, particularly life-saving intravenous fluids, IV antibiotics, and painkillers.

A Lancet study published in September 2025 measured the damage empirically. Researchers assessed availability of 25 essential medicines across 14 functioning facilities in Gaza between May and July 2025. Mean availability in northern Gaza was 37%, versus 54% in southern Gaza. Humanitarian centers and private pharmacies in the north had availability of just 20% and 23% respectively. Insulin was unavailable in every private pharmacy surveyed. Antidiabetic and antiepileptic drugs were the hardest-hit categories. These figures represented a significant deterioration from the 55% mean availability recorded across Gaza in early 2023, itself a poor benchmark.

An estimated 200,000 people were at risk of losing access to emergency medical care. Another 100,000 faced disruption of surgical services. Seven hundred intensive care patients were directly affected by the shortages.

The implications for commercial pharmaceutical companies are less direct here but no less significant. Gaza represents the extreme end of a spectrum that every multinational operating in contested geographies must understand: supply chain access can be severed not by logistics failure, but by political decision. No amount of inventory buffering, route diversification, or scenario planning protects against a border closure enforced by a state actor. This is not a supply chain problem. It is a political economy problem wearing supply chain clothing.

For executives building conflict-readiness frameworks, Gaza introduces a category of disruption that most scenario planning ignores entirely: weaponized access denial. The standard corporate playbook assumes that if a company wants to deliver products and has the logistical capacity to do so, access is a solvable problem. Gaza demonstrates that access can be made structurally impossible by a belligerent party, and that no amount of corporate goodwill, humanitarian exemption, or logistical ingenuity changes the calculus when a sovereign actor decides the border stays closed. Companies operating in or adjacent to conflict zones need to ask a question that rarely appears in risk assessments: not “can we get product to market?” but “can access itself be denied as a tool of war, and what is our exposure if it is?”

Subscribe now

The Gulf Conflict and the Dubai Chokepoint

The current conflict in the Gulf has produced the most operationally significant pharmaceutical disruption in years, and it has exposed a dependency that the industry spent a decade building without adequately stress-testing.

Dubai functions as a critical pharmaceutical re-export hub, where medicines arrive, are warehoused, and then shipped globally. CEVA Logistics, DHL, and DP World have significantly expanded the Emirate’s pharmaceutical and healthcare distribution footprint, building out re-export and cold-chain capacity. DHL alone announced a $2.35 billion investment in life sciences logistics in 2025. Dubai’s air-cargo capacity runs close to 4 million tons annually.

When that hub came under pressure, the consequences were immediate. Following U.S. and Israeli strikes on Iran and Iranian retaliation in early 2026, major airports in Dubai, Abu Dhabi, and Doha were affected. Commercial activity through the Strait of Hormuz dropped to roughly 10% of pre-war levels. Global air-cargo capacity in the Gulf region fell 79% in a single week, driving a 22% reduction worldwide.

The pharmaceutical impact was direct and measurable. Conflict in the Middle East disrupted the flow of critical medicines to the Gulf, imperiling supply routes for cancer drugs and other treatments requiring refrigeration, and forcing companies to reroute flights and find overland alternatives. One pharmaceutical executive noted that internal teams had been assembled to prioritize patient-critical shipments, warning that some temperature-controlled shipments could miss connections unless proper storage and handling were secured.

Europe-to-Asia cargo that typically transits through Dubai or Doha has been redirected through China or Singapore, routes that add days to already tight cold-chain windows and significant cost to already thin-margin generic medicine logistics. Companies are trucking medicines overland through Saudi Arabia, Oman, and Turkey, and relying on dry ice to maintain temperature integrity over extended routes. Shipping costs on some affected lanes have risen 200% to 400%.

The preparedness data is damning. A survey of over 800 C-suite leaders for the Operations Outlook 2026 report found that only 22% of global pharmaceutical firms were actively conducting risk assessments for geopolitical threats at the end of 2025, with the same proportion developing formal scenario planning capabilities. This is not a gap. It is a posture. The Strait of Hormuz has appeared in geopolitical risk literature for the better part of four decades. The scenario was not unimaginable. It was merely inconvenient to plan for. And then it happened.

Subscribe now

The Dependency Beneath the Dependency

Beyond the logistics issues you can see a structural vulnerability that conflicts expose. The global pharmaceutical supply chain runs through just two countries for its most critical inputs.

The U.S. imports roughly 47% of its generic drugs by volume from India. India, in turn, relies on China for approximately 65% to 70% of its key starting materials, intermediates, and active pharmaceutical ingredients. After adjusting for this secondary dependency, researchers estimate that 46% of the U.S. generic drug supply has Chinese origins, far exceeding the 13% that direct import figures suggest. China remains the world’s largest producer of active pharmaceutical ingredients, and this dominance has, in critical sub-sectors, intensified rather than diminished.

The transparency problem is actually worse than the concentration itself because drug companies are not legally required to list the API country of origin on product labels. Regulators, procurement officers, and patients have no visibility into the actual geographic origin of the molecules inside the bottle.

The global pharmaceutical supply chain currently operates under a pervasive cognitive dissonance. On one side exists the political rhetoric of de-risking, friend-shoring, and supply chain sovereignty. On the other lies the immutable economic and chemical reality: China’s dominance over the API market has not only endured but, in critical sub-sectors, intensified.

U.S. tariffs on pharmaceutical imports have compounded the pressure. Duties of 25% on APIs from China and 20% from India took effect in 2025, with far higher rates on certain categories of finished pharmaceuticals, including tariffs of 100% on branded and patented imports from China effective October 2025. The market now faces input price inflation, supply disruptions, and longer lead times at precisely the moment when conflict-driven logistics costs are already elevated.

The generic medicine market is particularly exposed. Nearly a third of oral medicines currently in shortage are priced at less than $1 per unit, and 44% of injectables in shortage are priced below $5. Petroleum-derived inputs are essential to the production of key starting materials, so rising energy prices further compress margins. At those margins, no company has the capital buffer to absorb prolonged conflict-driven cost shocks. The economics do not bend. They break.

Subscribe now

What This Means for Executives in Other Sectors

Pharmaceutical companies operate at the most visible intersection of commerce and human life. That makes them a useful diagnostic for the conflict-readiness of industry more broadly.

Four patterns emerge from the evidence.

Humanitarian exemptions create governance complexity. The ability to stay in a conflict market because your product saves lives does not resolve the question of whether staying is the right decision. It defers that question to a more ambiguous, reputationally exposed stage. Companies in financial services, energy, and technology face a version of the same choice: staying may be legally permissible and commercially rational, while simultaneously being reputationally corrosive. There is no stable neutral position in a high-visibility conflict. Pharmaceutical companies have been living this reality for years. Other sectors are catching up.

Supply chain concentration risk is not primarily a logistics problem. It is a strategic choice that accumulates over decades of cost optimization and only becomes visible when a conflict or tariff regime makes rerouting necessary. The China-India API dependency existed before COVID, before the Red Sea Houthi attacks, and before the Strait of Hormuz closure. Each crisis revealed the same underlying structure. The question for boards is not whether they have a contingency plan for the next disruption. It is whether the underlying geography of their inputs has changed since the last one.

Access denial is an emerging category of conflict risk. Gaza demonstrates that supply chains can be severed by deliberate sovereign action. The standard risk matrix treats logistics disruption as a probability-times-impact calculation. Access denial by a state actor is categorically different: it is binary, it is political, and corporate mitigation options are near zero once it occurs. Executives whose operations touch conflict-adjacent geographies should be modeling this scenario explicitly, not treating it as an edge case.

The preparedness gap is a choice. Only 22% of pharmaceutical firms were conducting geopolitical risk assessments at the end of 2025. Fewer than 20% of global companies have a dedicated geopolitics function. These numbers, measured against decades of available scenario-planning literature on the very disruptions now unfolding, indicate that under-preparedness is not an information deficit. It is a choice to defer resilience investment in favor of near-term cost efficiency. The companies that made that choice are now paying the difference in rerouting costs, stockout risk, and reputational exposure. The executives who recognize this pattern in the pharmaceutical sector and see it reflected in their own industries have a narrow window to act before the next disruption renders the question academic.

Subscribe now

The Decision Architecture That Does Not Exist

Many of the problems I hear about from executives in anonymous interviews are that the decision frameworks they use are often improvised under pressure. The Russia-Ukraine case, the Gaza blockade, and the Gulf disruption each demanded different kinds of choices, but many companies seemed to approach all three with roughly the same ad hoc process: convene a cross-functional working group, assess legal exposure, gauge reputational risk through media monitoring, and issue a statement calibrated to the news cycle. That is crisis communications, not strategic decision-making.

A genuine conflict-zone operating framework would require three elements that many companies lack.

The first is pre-committed decision thresholds. Before a crisis hits, boards should define the conditions under which operations shift from normal to contingency to exit. The U.S. Chamber Foundation and Harvard Business School’s 2025 “High Stakes” report found that while most American companies now rank geopolitical risk among their top concerns, fewer than one in five have translated that awareness into structured governance processes. BCG’s Center for Geopolitics has documented a similar gap: less than 20% of global companies have created a dedicated geopolitics function, even as the number of active interstate and civil conflicts has reached its highest level since the Second World War.

The second is a product-level ethical triage. Not all products carry the same moral weight. Insulin is not Botox. Cancer treatments are not elective aesthetics. AbbVie’s decision to continue essential medicines while suspending aesthetic operations represented an implicit triage. A pre-existing framework would classify product lines by humanitarian criticality and assign different operating rules to each tier before the crisis forces the question.

The third is revenue accountability mechanisms. The donation-of-profits model adopted by AbbVie, Merck, Pfizer, Eli Lilly, and GSK was a creative improvisation, but none of these companies had third-party auditing of those commitments built into their governance structures. A framework that requires companies to pre-commit revenue from conflict-market operations to humanitarian or escrow accounts, with independent verification, would convert an ad hoc reputational gesture into a structural commitment.

The absence of these frameworks is not a knowledge problem. The U.S. Chamber report, BCG’s geopolitics research, EY’s Geostrategic Outlook and my own interviews all converge on the same diagnosis: executives recognize the risk but have not invested in the decision making and governance machinery to respond to it.

The gap between awareness and preparedness is where reputational and operational disasters originate.

That is the version of this story that Western media told last week, and it is almost entirely wrong. That version is comfortable because it reduces a deliberate political act to an operational failure. It asks companies to draw the wrong lessons. And it leaves them completely unprepared for what is actually coming.

Here is what actually happened.

Subscribe now

Three signals, one campaign

On May 18, 2026, the 46th anniversary of the Gwangju Democratization Movement, Starbucks Korea launched a promotion for a 503ml tank-shaped tumbler. The campaign slogans were “Tank Day” and “Slam on the Desk.”

Each element carried specific meaning to any Korean adult.

May 18 is the day the military dictatorship of Chun Doo-hwan deployed tanks against pro-democracy protesters in Gwangju. Official figures record 165 civilians killed, with 65 listed as missing and hundreds more later dying of injuries. Many details remain officially unresolved.

The tumbler’s 503ml capacity has drawn scrutiny of its own. In Korean political shorthand, 503 is the detention inmate number assigned to former President Park Geun-hye after her 2017 imprisonment, a figure that became a coded symbol in conservative circles. Critics and academic analysts argue the choice was deliberate. Starbucks has said that the 17-ounce size (which rounds to 503ml) was in use before Park’s detention and reflects standard sizing. A professor at Chonnam National University’s May 18 Institute reviewed the campaign’s numerical choices and called them “too calculated to be a coincidence.”

“Tank Day” is the derogatory right-wing nickname for May 18, used by those who want to minimize or mock its significance.

“Slam on the Desk” refers to the 1987 torture death of student activist Park Jong-chul. When police were caught, they told the public that an investigator had simply slammed his hand on a desk and the young man died from the shock. The phrase became one of the most notorious lies in Korean democratic history.

Three signals. One campaign. One date.

All together this seems less like an accident and more like a message.

Subscribe now

Who sent the message?

The operational CEO of Starbucks Korea, Sohn Jeong-hyun, was fired within hours. Western coverage largely stopped there.

It should not have. The CEO is an agent of the company. The principal is Chung Yong-jin, Chairman of Shinsegae Group, a Samsung-family scion and publicly self-identified supporter of the Korean political right. Shinsegae’s E-mart affiliate holds a 67.5% controlling stake in Starbucks Korea, acquired in 2021 for 474.3 billion won ($341 million at the 2021 exchange rates). Chung issued a personal apology and has since faced calls for his own resignation.

The Korean public understood the distinction immediately. The message was not from Starbucks. The message was from Chung, sent through Starbucks.

The cost, so far

The response was swift and severe.

E-mart shares fell 6% in a single session. The 10,000-member Delivery Platform Drivers’ Union announced a full boycott. Multiple government ministries cancelled contracts. The Ministry of National Defense began reviewing its welfare partnership with the chain. Police opened a formal investigation into whether the campaign constituted defamation and insult of the Gwangju victims.

Starbucks Global headquarters called it an “unacceptable marketing incident,” apologized to the people of Gwangju, and announced a full investigation with strengthened internal controls.

Then came the financial clause in the licensing agreement.

The agreement between Starbucks Coffee International and Emart contains a buyback provision: if the agreement is terminated due to Emart’s liability, Starbucks can repurchase the 67.5% stake at a 35% discount to fair valuation. Starbucks Korea generated 3.24 trillion won ($2.36 billion USD at 2025 exchange rates) in revenue last year. It is Shinsegae’s single most important cash asset.

Chung Yong-jin may have used his most profitable business to send a political signal in a market he can now lose at a steep discount.

Subscribe now

There was no neutral lane

The most important lesson here concerns every company operating in any country (not just Korea). In a hyperpartisan, polycrisis world, every country has a local version of hot button issues that stakeholder groups demand companies have an opinion on. Here’s what they were in Korea.

In December 2024, South Korean President Yoon Suk-yeol declared martial law in an attempt to suspend the National Assembly. The gambit failed within hours. Yoon was impeached. The Constitutional Court upheld the impeachment in April 2025. South Korea had just survived an autogolpe (an attempted self-coup by a sitting president) and the country’s collective memory of what military rule looks like, and who paid for it, was freshly activated.

May 18, 2026 was the first May 18 after a sitting president had tried to deploy the military against his own parliament. The political meaning of Gwangju was not historical background. It was live commentary on the present.

Every brand operating in Korea faced a binary on that date. Honoring May 18, even briefly, was a political act. It said: we stand with this country’s democratic history. Staying silent was also a political act. In the post-autogolpe moment, silence read as indifference to, or quiet alignment with, the forces that had just tried to end Korean democracy.

In research published in Harvard Business Review, Jason Miklian and I documented how businesses that think they can stay above the fray of contentious political issues are woefully misguided. In a hyperpolarized environment, customers interpret inaction as a position. In a historically charged one, silence carries a specific meaning that the public assigns whether or not the company intended it (Katsos and Miklian 2021).

Starbucks Korea did not choose silence. It chose a third path: actively mocking the democratic dead while the principal sent a message to half the country. Multinationals watching from the sidelines who assume their own silence would have protected them should understand: it would not have. They would simply have paid a different political price.

There was no neutral lane. There never realistically is our current era of polycrisis. The only question is whether your company chooses its position or has it chosen for you.

Subscribe now

The three options Chung actually had, and why he chose the worst one

Once we accept that political neutrality was not available, the question becomes: what should a business owner actually do with a mass-market brand that serves everyone if they have certain political commitments?

There are three options. Chung Yong-jin chose the third, and it is the only one that was guaranteed to destroy value.

The first option is to suppress your political identity entirely when operating a mass-market brand. Do not signal, do not align, do not mix personal politics with commercial operations. Many executives do this successfully. The cost is personal: you accept that your business will not serve your political agenda. The benefit is commercial: your brand does not become a battleground.

Suppression only works if it is genuine and sustained. Partial suppression, i.e., staying quiet most of the time and then sending coded signals to your political allies, is not a strategy. It is a fuse. When the signal is eventually decoded, as it will be in a media environment where every consumer is also a broadcaster, the damage is compounded by the betrayal. Customers who bought Starbucks Korea assuming it was apolitical feel more deceived, not less, than they would have if Chung had simply been open about his views from the start.

Suppression also creates a specific problem with employees, suppliers, and government. These stakeholders are making decisions about the relationship based on what they believe your company stands for. A government ministry that has a welfare contract with Starbucks Korea is not doing so because it endorses the political right. When the suppression breaks down and the signal is decoded, the ministry does not just cancel the contract. It feels used. The reputational damage extends far beyond the consumer boycott.

The second option is what might be called transparent political identity with commercial universality. Be clear about what you believe, operate your business as a service to everyone, and let the market decide.

This is the Chick-fil-A model, applied with intelligence. Everyone knows where that company’s ownership stands. It has not hidden it. It has also not turned its stores into political instruments. Its employees serve every customer. Its products are available to every consumer. The market response has been mixed, but the company has maintained a clear operating identity. It has not deceived anyone, and it has not used its commercial platform to send coded signals about who deserves to live in a democracy.

This path requires courage. It accepts that some customers, suppliers, and government partners will not work with you. In a deeply polarized market, it may mean accepting a smaller market share. The reward is the only thing that sustains a mass-market brand over time: the trust of the customers who do choose you.

The third option is what Chung actually chose: use the commercial platform as a covert political instrument, signaling to allies while maintaining plausible deniability with everyone else. This is the worst possible option.

Covert signals are only valuable if they are understood by the intended recipients. Which means they can be decoded. Which means they will be decoded. The precision of the Starbucks Korea campaign was not subtle. It was legible to anyone who grew up in the country. The campaign was designed to be read. It was read. And once decoded, it carried none of the credibility of a declared position and all of the toxicity of a revealed deception.

If you are going to hold a political view, holding it covertly while running a mass-market business is the version that has the highest probability of catastrophic failure. Every major stakeholder — customers, employees, suppliers, government, global partners — is making ongoing decisions about your company based on an assumption of authenticity. The moment that assumption collapses, so do the relationships built on it.

Subscribe now

What Starbucks Korea should do now

The choices facing Starbucks Global have narrowed considerably, and they are harder than they would have been had stronger governance mechanisms been in place. What was once a preventable exposure is now an active crisis requiring a decision between two costly paths.

The first option is to stay with the existing shareholders. This preserves the relationship and avoids the complexity of a buyback. It also means continuing to operate a 2,115-store flagship market under a principal whose political commitments have been publicly decoded by millions of consumers. The boycott is already real as sales have dropped significantly and in a market structured around habitual consumption and prepaid card loyalty, damaged trust compounds. The risk is a permanent recalibration of what the brand means in the country where Starbucks has its third-largest global store network.

The second option is to exercise the buyback clause, acquire Emart’s 67.5% stake at a 35% discount to fair valuation, and attempt to resell the license to a clean owner at a higher value. The financial logic is sound: Starbucks acquires a distressed asset cheaply, rehabilitates the brand under its own management, and exits to a better-aligned partner at a profit. On paper, this is the sharper move.

What that calculation does not capture is the cultural weight that rehabilitation carries in this specific market.

Honor in Korean social life operates through relational concepts (like chemyeon (체면) and han (한)) bound up within a hierarchy of relationships within families, regions, school cohorts, political parties, etc. and specific accumulations of grievance and unresolved injustice give moral weight to victimhood narratives and sustains them across generations. These create a politics where public humiliation carries enormous cost, genuine apology requires accepting that cost unconditionally, and perceived insult to collective dignity can mobilize constituencies with unusual speed and force.

If Starbucks Global exercises the buyback clause, it completes a potentially lucrative financial transaction. It also publicly strips one of South Korea’s most prominent business families of their most profitable asset, at a penalty, because of conduct the principal directed. The cultural and political implications run in two directions simultaneously. Conservative constituencies may interpret the buyback as foreign capital exploiting a political crisis to dispossess a Korean conglomerate, a narrative that maps onto longstanding honor claims about foreign economic power. Progressive consumers will watch whether Starbucks’ rehabilitation is genuine or merely a rebranding exercise, measured against similar cultural and political standards: forgiveness will not be extended cheaply and acknowledgment will have to be proportionate to the wound.

That wound runs deep. The Gwangju Uprising is not a settled historical event in Korean political life. It is the defining grievance of the Honam region — the Jeolla provinces, including Gwangju itself — and the animating moral claim of the progressive political tradition going back to the Park Chung-hee era. Regional voting patterns in South Korea still track the Yeongnam-Honam divide hardened by decades of military rule and economic marginalization. The Tank Day campaign deliberately activated a wound on the one day of the year when it is most raw, in a market where the political meaning of that activation is understood by everyone.

Brand rehabilitation in this environment requires more than a new owner and a cleaner governance structure. Genuine acknowledgment means accepting the full cost of the apology without qualification. Carefully worded statements that preserve enough face will suggest the gravity was not fully understood. Korean audiences have considerable experience reading the difference between corporate contrition and genuine reckoning. So has the Korean press.

If Starbucks Global takes command of the shares, it becomes directly responsible for that reckoning. It cannot outsource the political rehabilitation to a new management team and call it done. The brand’s credibility in South Korea will remain tied to whether the Tank Day promotion is addressed thoroughly which may also mean slowly.

Subscribe now

What this means for the board in Seattle

Starbucks Global’s governance failure runs deeper than marketing approval. The core problem is a licensee political risk framework with no mechanism for what happens when the controlling principal uses the brand as a deliberate political instrument.

That failure is now more expensive to correct than it would have been to prevent, and the Korean cultural context makes explicit why.

Standard political risk modeling treats the environment as something that acts on a company: elections shift policy, social unrest disrupts operations, regulatory changes impose costs. The questions this case exposes is a different one: what happens when the company, through its local partner, acts on the environment? When the actor is a someone embedded in a politically polarized elite, operating in a society with deep honor-laden historical grievances?

In Korea, that choice carries cultural amplifiers that most boardroom risk frameworks do not model. Chemyeon means the manner of response matters as much as its content. A conditional apology, one that preserves any face for the apologizing party, reads as insufficient regardless of its substantive commitments. Han means the Gwangju wound is not a background factor but an active moral claim with generational constituencies, and those constituencies will measure Starbucks against a standard that runs in years, not news cycles.

The broader political structure deepens the constraint. The conservative-progressive divide in South Korea is organized around competing honor claims (who was wronged by history, who paid the price, and who must be vindicated) rather than primarily around policy preferences in the Western sense. These are loyalties, not policy disagreements that a rebranding campaign can move past. Any new licensee will not enter a neutral market. It will enter one where part of the country watched Starbucks be used to mock its defining historical grievance, and the another part spent weeks using Starbucks as a conservative symbol. The first May 18 under new ownership will be the most consequential single day on that licensee’s calendar, and there is no quiet way to get through it.

The governance question for every global franchisor is the same one Starbucks failed to ask in time: do we know the political identity of the individuals who control our licensees, do we understand the social context in which our brand operates, and does our licensing agreement contain the governance teeth to respond before a crisis, rather than after?

The prepared mind asks that question before the answer arrives in a tumbler.

The scenario your risk team is not running

Most multinational political risk frameworks model exogenous events: elections, regulatory shifts, geopolitical escalation, social unrest. They model what governments and publics might do to a company.

They rarely model what a franchise partner might do with the company’s brand, and what the exposure looks like when that action is deliberate rather than negligent.

That is a non-zero probability scenario in every market where your brand operates within a politically polarized elite, in a society with unresolved historical trauma, in an environment where even silence carries political meaning.

Sound familiar?

For companies treating the Gulf as a single operating zone, that divergence is the part that matters most, and the part least accounted for.

Subscribe now

What Saudi Arabia Is Doing

Riyadh is trying to position itself as the region’s de-escalation actor. The Saudi-Iran rapprochement brokered in Beijing on March 10, 2023, was the formal signal. Since then, Saudi Arabia has maintained diplomatic channels with Tehran, coordinated explicitly with Egypt, Turkey, and Pakistan through an emerging regional framework aimed at de-escalation and a negotiated settlement, and consistently resisted pressure to sharpen its alignment with the United States and Israel in ways that would foreclose future accommodation with Iran.

The strategic logic underneath this posture is straightforward. Saudi Arabia has concluded that Iran, in some form, survives this period. Whatever emerges from the current conflict, the regime in some form will still be across the water. The kingdom will need to live alongside it. Burning the diplomatic architecture now for the sake of alignment with a US posture that may shift, or an Israeli military campaign whose endpoint is unclear, is a bet Riyadh has (seemingly) decided not to take.

Sovereignty preserved through managed constraint. A long bet that stability on negotiated terms beats confrontation with a neighbor that is not going away, even when that neighbor is locked in an escalating standoff with your own security guarantor.

There is a historical precedent for exactly this kind of bet, and it held for four decades, but with some unexpected consequences.

Subscribe now

What Finland Did

After the Second World War, Finland found itself in a position that rhymes but doesn’t perfectly fit with what Riyadh is managing now. It had been attacked by the Soviet Union and was sitting on the border of a nuclear superpower with a long memory. The Soviets were in open confrontation with Finland’s Western partners who were more natural allies for Finland. The options available to Helsinki were stark. Full alignment with the West risked Soviet retaliation and potentially absorption. Full subordination to Moscow meant the end of Finnish independence. Finland chose neither.

What emerged was a doctrine of deliberate, institutionalized accommodation. Finland signed the 1948 FCMA Treaty with the Soviet Union, which committed Helsinki to resisting any use of Finnish territory as a base for attacks on the USSR. Finland kept its distance from NATO. Finnish officials exercised consistent self-censorship on foreign policy statements that might antagonize Moscow. At the same time, Finland maintained its institutions, integrated economically with Western markets, and built a prosperous, functioning state.

The West German politicians and political scientists who coined and weaponized the term “Finlandization” meant it as an insult. The Finns wore it as badge of strategic honor, calibrated and maintained over four decades. Helsinki made a clear-eyed bet that independence on constrained terms was worth more than confrontation with a neighbor who was an existential threat. The constraints and restraint were chosen deliberately, built into law and institutions, and held.

Finnish companies that operated in this environment learned to treat the constraints as structural features of the market. They built commercial relationships on both sides of the Iron Curtain. They did not wait for strategic clarity that was never coming.

The doctrine held until the Soviet Union dissolved in December 1991 and the rationale for it disappeared. Finland replaced the FCMA Treaty with a new agreement with Russia the following month.

The word “Finlandization” has been applied to Gulf strategy toward Iran by analysts at institutions including the Washington Institute for Near East Policy. No senior Saudi official has publicly adopted the label. The underlying strategic posture is extensively documented, and the parallel holds regardless of whether Riyadh uses the word.

Subscribe now

A Region Reading From Different Maps

Saudi Arabia’s posture toward accommodation is not the only strategic reading available in the Gulf, and Riyadh is not the only capital making consequential decisions.

Across the region, Gulf states are contending with the same facts and arriving at different conclusions. Iranian missile and drone activity has damaged critical Gulf energy infrastructure and created asymmetric threat perceptions across capitals that were, until recently, assumed to share a unified security outlook. Different states have drawn different lessons from the same events. Some have leaned toward de-escalation and diplomatic channel maintenance. Others have recalibrated toward closer alignment with external security partners and harder postures on Iranian conduct.

These are sovereign choices, each internally coherent given the threat environment each capital is managing. For outside observers, the relevant observation is that the strategic calculi have diverged, and the divergence is now producing different policy outputs across the region simultaneously.

The most commercially legible data point is the UAE’s decision to exit OPEC, effective May 1, 2026. Energy Minister Suhail Al Mazrouei described it as a sovereign strategic decision following a comprehensive assessment of national production policy and future capabilities, based solely on the national interest and the UAE’s responsibility as a reliable energy supplier. He stated explicitly that the decision was not driven by political considerations. The UAE’s actual output had been held to roughly 3.2 million barrels per day under OPEC quotas, well below an installed capacity already approaching 4.85 million. ADNOC’s target is 5 million barrels per day by 2027. Whatever the full set of motivations behind the decision, its structural effect is to reduce the framework through which Gulf states have historically coordinated on energy policy, at a moment when such coordination carries more strategic weight than usual.

For companies, the policy rationale matters less than the operational implication: the Gulf’s largest economies are now running different energy strategies, different security postures, and, increasingly, different regulatory and commercial environments. Capitals that once moved in close alignment are now making unilateral calls on questions that affect the entire regional operating environment.

Subscribe now

Where the Cold War Parallel Holds

Three features of the Finnish experience map directly onto what is unfolding in the Gulf, and each has practical implications for companies if the Gulf becomes a “neutral” space a la Finland.

The dual-track economy is the business model. Finland traded heavily with the Soviet Union while integrating with Western markets. Gulf states have been running this playbook for years: sovereign wealth funds in Western equities, Chinese-financed infrastructure, US military bases, Iranian trade running through indirect commercial channels across the region. Saudi accommodation normalizes this structure rather than creating it. Companies that have been treating Gulf commercial ambiguity as a transitional state are planning for an environment that is not coming. The dual-track is the destination.

Chosen constraint looks like weakness until it holds. Western analysts who dismissed Finlandization as capitulation were consistently wrong about its durability. The Finnish bet held for four decades because it was calibrated to actual power realities rather than preferred outcomes. The same analytical error is available in the Gulf. Saudi accommodation will read as weakness to commentators who expect confrontation. For companies, the relevant question is how long the posture holds and what it means for the operating environment while it does.

Ambiguity is not a transition. Finnish companies built for a market where the constraint was permanent, not for the day the Soviet Union would become a normal neighbor. Gulf-exposed companies waiting for the strategic situation to clarify before making decisions about capital allocation, operational structure, or workforce positioning are making a version of the same mistake.

Subscribe now

Where the Parallel Breaks Down

The Cold War comparison earns its limits in four places, and the limits matter more than the similarities for practical risk analysis.

The Gulf has no unified actor. Helsinki managed one dominant constraint with one great power neighbor. The Gulf presents a more complex structure, with multiple capitals now operating from different strategic premises simultaneously. The differences are producing different regulatory environments, different reputational exposures, and different commercial risk profiles across the same region. Companies with exposure across multiple Gulf jurisdictions are operating in a cross-jurisdictional environment that requires explicit mapping, not a single-market assumption.

Energy infrastructure is the target. Finland had no hydrocarbons. Finnish neutrality required nothing of the Soviet Union regarding Finnish oil fields, because there were none. Gulf accommodation is being negotiated over the most strategically consequential energy infrastructure on earth, while regional production expansion plans are running on their own separate trajectory. Those two dynamics sit in direct tension with each other.

The nuclear variable changes weekly. Finland accommodated a nuclear power, but the accommodation was not being negotiated inside a live proliferation crisis. The current Gulf de-escalation posture exists within an active US-Iran nuclear negotiation whose status changes the value of any accommodation framework from one week to the next. The strategic ground is moving beneath the analysis in a way that had no equivalent in Helsinki’s calculations.

There is no institutional lock-in. The FCMA Treaty gave Finnish Finlandization a legal architecture and domestic legitimacy. The Saudi accommodation posture is being pursued without an equivalent formal architecture. A shift in leadership calculus, in Riyadh or in Washington, changes the posture without notice. The durability that made Finnish Finlandization a viable business planning assumption is not present in its Gulf equivalent.

Subscribe now

What This Means for Companies

Three categories of exposure deserve explicit attention from companies right now.

Companies treating the Gulf as a single market. The strategic divergence across Gulf capitals means the compliance and reputational environments of the region’s major economies are now moving in different directions. What is commercially straightforward in one jurisdiction is beginning to create complications in another. This is already visible in banking and financial services decisions across the region. The first action is audit: where do your operations, partnerships, capital relationships, and regulatory exposure create cross-jurisdictional risk that your current framework is treating as a single exposure?

Energy, logistics, and supply chain companies. Saudi accommodation means Saudi Arabia will not automatically align with US escalation against Iran. Hormuz remains a live risk. If the US acts without Gulf cooperation, states across the region are positioned between their security patron and their accommodation posture, and logistics infrastructure is caught in that gap. Build the non-cooperation scenario into your contingency planning. Do not assume Gulf infrastructure remains operationally available under escalation conditions, because the governments controlling that infrastructure are managing more complex alignment calculations than they were two years ago.

Financial services and sovereign wealth fund partners. Finnish companies that operated across the Iron Curtain learned that dual-track exposure required dual-track governance. Companies with Gulf SWF capital, board relationships, or joint ventures need to understand that their Gulf partners are managing postures where certain commercial associations carry reputational implications in multiple directions simultaneously. This is already shaping private conversations across the region. Map the exposure before it surfaces in a context where you have no time to manage it.

Subscribe now

The Lesson from the Finns

Finnish companies that thrived during Finlandization were not the ones that waited for the Cold War to end. They were the ones that built institutions, commercial relationships, and governance structures calibrated to the actual operating environment rather than the one they preferred.

The Gulf is not going to restabilize on a timeline convenient for corporate planning cycles. Capitals that once moved in close alignment are now pursuing different strategic bets on how this period ends. The fracture is now visible in energy policy, alliance choices, and commercial regulation simultaneously.

The companies that get through this will be the ones that stop treating Gulf ambiguity as a problem to be waited out. The ambiguity is the operating environment. Plan for it accordingly.

The argument deserves a direct response.

Subscribe now

The Central Claim and Its Problem

Ray’s piece rests on a governing assertion: that the strategic risks confronting the United States and the operational risks confronting American firms are “increasingly one and the same.” If that premise holds, everything else follows. Firms that update their risk frameworks to reflect geopolitical exposure are, in effect, doing their fiduciary duty by updated means. Government and industry converge. Everyone rows in the same direction.

The premise does not hold.

Ray is right to note that the prior architecture has broken down. For decades, many companies operated under the assumption that broad geopolitical risk fell outside the ordinary scope of fiduciary responsibility. For some firms, that assumption has clearly shifted. But geopolitical risk was never actually outside the scope of fiduciary duty. It was underweighted, poorly measured, and structurally easy to ignore during a period when the international order seemed stable and self-reinforcing. The error was epistemic, not definitional. Boards should always have been pricing this. The current environment simply makes the cost of not doing so visible.

More fundamentally, Ray’s claim that corporate and national strategic interests are converging does not survive contact with the evidence he himself provides.

His central cautionary tale is the telecommunications industry. American firms engaged with China through the WTO architecture, expanded aggressively into Chinese supply chains and markets, and treated the commercial opportunity as aligned with both their own interests and U.S. government policy. They were correct on both counts at the time. Washington actively encouraged that engagement. China’s WTO accession in December 2001 was a deliberate policy choice, supported by successive administrations and by the U.S. business community acting in good faith on the signals it was receiving. The House voted 237 to 197 to grant China Permanent Normal Trade Relations, and the deal opened Chinese telecommunications markets to foreign participation for the first time since the 1940s.

Now the U.S. government has concluded that the resulting architecture was not in the national strategic interest. American companies are being told to decouple, to reshore, to treat the relationships they built as liabilities rather than assets. The government’s position has changed. The companies’ position, built on decades of capital allocation, has not changed nearly as quickly as the political context demands.

This sequence does not demonstrate that corporate interests and national interests are converging. It demonstrates the opposite. Companies were doing exactly what the U.S. government wanted, under the assumption that what was commercially rational was also strategically sound. The subsequent realization by U.S. political actors that this was not in the long-term strategic interest of the United States does not retroactively place the companies in the wrong. It places the policy framework in the wrong. Ray’s own best example, examined carefully, dismantles his own premise.

Subscribe now

The Trillion-Dollar Gap

Here is what Ray’s framework does not reckon with directly: unwinding two decades of supply chain integration is a capital problem of extraordinary magnitude, not a strategic planning problem.

A board of directors does not reroute a supply chain because a retired Air Force general writes a compelling essay. It reroutes a supply chain when the alternative (staying the course) becomes more expensive than the transition. For most multinationals with significant China exposure, that calculation has not yet cleared. And for good reason.

The costs of decoupling are immediate, concrete, and quantifiable. Production runs on existing equipment, often Chinese-owned or Chinese-situated. The workforce is local. The logistics infrastructure has been built over decades and priced accordingly. Replacing it requires not just new factories, but new supplier ecosystems, new transport networks, new training pipelines, and new quality management systems, often in countries with no existing industrial base for the relevant products. Bank of America analysts estimated that repatriating foreign, non-Chinese manufacturing operations currently in China alone would cost $1 trillion over five years. A joint report by the Semiconductor Industry Association and Boston Consulting Group estimated that building fully self-sufficient semiconductor supply chains would require at least another $1 trillion in upfront investment, with recurring annual cost increases of $45 to $125 billion. Across all sectors, the total figure runs well into the trillions. Some sectors, notably semiconductors, have begun to get public investment at scale through the CHIPS and Science Act, which appropriated $52.7 billion (including $39 billion in direct manufacturing incentives). Most sectors have not.

Ray writes that “Washington will eventually have to build that framework, but firms cannot afford to wait for it to be finished before they begin to act.” That sentence, read carefully, asks private enterprises to absorb the costs of a national security transition that the government has not yet funded, on a timeline the government has not yet specified, toward an industrial structure the government has not yet designed, while remaining legally obligated to their shareholders.

That is a request dressed as strategy.

Subscribe now

Friend-Shoring Is Not a Solution

Ray, along with most of the current economic statecraft literature, treats “friend-shoring” as a meaningful middle path between the current China dependency and a full domestic reindustrialization that remains politically and economically out of reach. Move supply chains to allied or friendly nations. Reduce exposure to adversarial coercion.

The problem is that “friendly nations” is not a fixed category.

Countries shift. Governments change. Strategic alignments that look durable from Washington can dissolve inside a single election cycle in the host country, or inside a single trade negotiation with Beijing. The countries most frequently cited as friend-shoring destinations (Vietnam, India, Mexico, parts of Southeast Asia) are not geopolitically anchored to the United States in any treaty sense. Several of them maintain active and deepening economic relationships with China.

Moving a supply chain from Shenzhen to Ho Chi Minh City exchanges one form of exposure for another, while adding transition costs and frequently lower productivity and quality in the near term. If Vietnam’s government shifts toward the Chinese sphere of influence in the next decade (and there are plausible scenarios in which that happens) U.S. companies will have spent billions to recreate a dependency problem with a different flag on it.

There is a second dimension to this that Ray’s framework omits entirely. The dependency argument cuts in more than one direction. The multinationals best positioned to weather the geopolitical storms ahead are the ones most diversified in their supply chains, their financing, their market access, and their political exposure. That last category increasingly includes exposure to U.S. political volatility as well as to Chinese or Russian pressure. A company that has relocated its entire supply chain back into the United States, or into jurisdictions tightly coupled to U.S. foreign policy, has not reduced its dependency. It has concentrated it differently. The strongest firms will be those that have distributed their exposure broadly enough that no single government’s decisions, including Washington’s, can impose existential cost.

Subscribe now

The Fiduciary Duty Problem

Ray’s argument for “redefining corporate self-interest” is, on the surface, appealing. He explicitly says firms should not subordinate commercial judgment to government direction. What he proposes instead is that boards and executive teams should “update their own definition of prudent stewardship.”

Fiduciary duty is not a concept boards get to update by philosophical reflection. It is a legal obligation, enforceable in court, with substantial and well-developed case law behind it. A board that voluntarily accepts lower short-term returns and higher transition costs on the grounds of geopolitical prudence, without a clear framework of government support, policy durability, or competitive protection, is legally exposed the moment those decisions damage shareholder value. Activist shareholders and institutional investors do not grade on a national security curve.

This has already played out. Companies that have moved faster on China decoupling than their competitors have, in multiple cases, faced shareholder pressure, analyst downgrades, and in some instances board challenges. The market has not consistently rewarded geopolitical prudence. Ray acknowledges this: “The market, for now, does not consistently price these risks.” His proposed remedy is for firms to act anyway. That answer, while morally coherent, is legally and commercially untenable as a general proposition.

The government needs to restructure the incentive environment so that geopolitical prudence is commercially rewarded rather than commercially punished. That requires durable policy commitments, enforceable industrial subsidies, and protected market conditions. Asking boards to accept more risk as an act of enlightened self-interest, without that structural change, solves nothing.

Subscribe now

The Counterparty Problem

There is a more fundamental issue that Ray’s framework does not address, and its omission is the most significant gap in the piece.

Ray’s entire argument assumes that the U.S. government is a stable, trustworthy counterparty for the alignment he is proposing. Firms expand their definition of fiduciary duty to incorporate national security risk. Congress provides durable policy and capital commitments. The military translates operational requirements into commercial terms. The three parties move in disciplined coordination toward a shared strategic objective.

That alignment depends on something Daron Acemoglu and James A. Robinson identified in their research on institutions and prosperity, work that contributed to the 2024 Nobel Prize in Economics (shared with Simon Johnson): the structural distinction between political power and economic power. Their research across centuries and continents, synthesized in Why Nations Fail, is consistent on this point. Societies in which political and economic institutions remain structurally inclusive, where the state cannot be captured by private interests and where private enterprise cannot be weaponized by political actors, tend to build sustained prosperity. When that distinction breaks down, the result is extractive institutions, accelerating inequality, and the kind of systemic fragility that typically precedes serious national decline.

The United States has historically been among the stronger performers on this dimension. The structural barriers were never perfect, but they were real: disclosure requirements, conflict of interest rules, the norm of divestiture from private holdings upon entering public office, independent regulatory agencies, and a legal architecture that treated the public fisc and private enrichment as categorically distinct.

Several of those barriers have been substantially weakened in recent years. The current administration has reduced the functional separation between political decision-making and private commercial interest to a degree that would have been structurally anomalous in prior periods of American governance. The presidency has become, to a meaningful and observable extent, a platform for commercial activity. Political decisions with direct consequences for asset prices, tariff structures, and market access are being made in contexts where the decision-makers have visible private financial interests in the outcome.

This is not primarily a partisan observation. It is a structural one, and it carries direct implications for the argument Ray is making.

If the strategic risks facing U.S. companies and the strategic interests of the U.S. government are to converge, if firms are to make long-term capital commitments on the basis of government signals, those firms need confidence that the signals reflect a durable national interest rather than a transient political or personal one. The systemic risk Acemoglu and Robinson describe is precisely the risk that those signals become unreliable, that policy becomes a function of proximity to power rather than a function of strategic analysis. Multinational corporate interests are not necessarily aligned with a government operating under those conditions. In some cases, they are directly exposed to it.

The companies best positioned to navigate this environment are not the ones most closely coupled to current U.S. government priorities. They are the ones that have maintained enough independence, geographic diversification, and institutional resilience to absorb shifts in political direction without catastrophic exposure. Long-term strategic deleveraging, in the current environment, likely means reducing concentrated exposure to any single political jurisdiction’s decisions, including Washington’s.

Subscribe now

The Huawei Lesson Cuts Both Ways

Ray uses Huawei’s rise as his central cautionary tale. It is a good one. But the lesson he draws from it is not the only lesson available, and the alternative lesson is more uncomfortable for his argument.

Huawei did not outcompete Western telecommunications companies because those companies were strategically naive. It won because the Chinese government treated telecommunications as a strategic industry and funded it accordingly. The Wall Street Journal reported that Huawei benefited from as much as $75 billion in grants, credit facilities, tax breaks, and other forms of government assistance. State-owned lenders offered preferential credit terms to Huawei’s foreign buyers. China Development Bank and the Export-Import Bank of China made more than $30 billion in credit lines available to Huawei’s customers. The domestic market was protected. Huawei won a subsidized competition, not a commercial one.

The lesson that case points to is this: the U.S. government needs to be willing to compete the way China competes, with state-backed capital, industrial policy, and protected conditions for strategic industries. The CHIPS Act is a partial recognition of that logic. It is also, in the context of the full decoupling problem, a fraction of what would be required.

The Huawei case is an argument for massive, sustained government intervention in industrial structure. Ray deploys it as an argument for corporate self-reflection. The evidence and the conclusion are badly mismatched.

Subscribe now

Congress Is Not a Corporate Board

Ray’s “three-party problem” analogy, with Congress as a corporate board setting conditions for the government-industry partnership, is the weakest section of an otherwise disciplined piece.

Corporate boards are accountable to shareholders. They face direct financial consequences for bad decisions. They operate under fiduciary obligations that are legally enforceable. They have access to material non-public information relevant to the decisions they make. They can be removed when they fail.

Congress does none of these things. It operates on electoral cycles structurally misaligned with the long-term investment horizons Ray is describing. It lacks systematic visibility into supply chain architecture, capital allocation, and strategic risk across the private sector. It is subject to lobbying pressures that often run directly counter to the national security interests Ray identifies. And it has demonstrated, repeatedly, that it cannot sustain durable industrial policy commitments across administrations.

The B-21 example Ray cites, a program on schedule and within cost discipline, is notable precisely because it is exceptional. The F-35, now expected to cost over $2 trillion through its lifecycle (400% more in inflation-adjusted dollars than the 2007 GAO estimate), with average delivery delays of 238 days per aircraft in 2024. The Littoral Combat Ship, whose per-unit cost more than doubled from $220 million to over $600 million, with the Navy decommissioning ships less than five years old because they couldn’t perform their intended missions. The Expeditionary Fighting Vehicle, cancelled after consuming $3 billion in development with another $12 billion still needed to build. The Coast Guard’s Deepwater program, which ballooned from $17 billion to over $24 billion before losing authorization entirely. The list of defense acquisition programs that have failed on cost and schedule is long. Ray knows this. He was in the system. The B-21’s success reflects specific conditions (a constrained contractor field, deliberate program management, sustained executive attention) that do not transfer to a broader industrial policy framework.

Subscribe now

Politics > Economics is the real argument

There is one more thing worth naming directly, because Ray largely declines to name it himself.

Despite the careful disclaimers (firms need not subordinate commercial judgment to government direction, national objectives are not a substitute for fiduciary duty) what the article actually lays out is a programmatic argument for bringing publicly traded U.S. companies into alignment with the political, military, and economic priorities of the U.S. national security state. The four risk categories Ray identifies, the three-party framework, the resilience-as-deterrence logic: taken together, they constitute a case that private enterprise should organize itself in response to government-defined strategic priorities. That is, by any reasonable reading, an argument for subordinating economic interests to political ones.

This is worth saying sympathetically rather than critically, because the underlying concern is legitimate. The geopolitical risks Ray describes are real, and the post-Cold War era produced a private sector operating under strategic assumptions that genuinely no longer hold. The argument that purely extractive commercial logic, profit accumulation without regard to systemic or political consequences, produces bad outcomes for the societies that permit it is one with serious intellectual grounding. Acemoglu and Robinson would recognize it. So would most serious students of political economy.

The problem is that this argument requires honesty about what it is. Framing the subordination of economic logic to political logic as an “update to fiduciary duty” obscures the nature of the ask. Asking companies to bear the transition costs of a policy failure while calling it enlightened self-interest is not a strategy. And asking firms to align themselves with government priorities at a moment when the separation between political and economic elites has demonstrably weakened creates the very systemic risk that the argument is ostensibly designed to prevent.

An honest framework has to begin with honest premises.

The first honest premise is that economic interests and national security interests often (though not always) diverge structurally. Acknowledging this is the prerequisite for designing policy that can actually close the gap.

The second honest premise is that closing that gap requires the government to carry costs it is currently asking the private sector to carry unilaterally. Someone has to fund the transition. The government has to be honest about that number and commit to paying it, or accept that the transition will not happen at the pace or scale the strategic environment requires.

The third honest premise is that (genuine) resilience, the kind that actually deters, requires institutional integrity as much as industrial capacity. A country that has weakened the structural separation between its political and economic elites has introduced a category of systemic risk that no amount of reshored manufacturing or friend-shored supply chains can offset.

Resilience emerges from sustained public investment and from institutions stable enough to make that investment credible over time. Asking boards to update their definition of prudent stewardship, in the absence of those foundations, places the burden of a structural problem on the entities least positioned to solve it.

John Katsos is a Visiting Professor at the Indiana University Kelley School of Business and Professor at the American University of Sharjah. He is co-author of CRISIS: Responsible Management in Crisis (Cambridge University Press) and writes The Mind Prepared on the intersection of geopolitical risk and corporate strategy.

The United Nations Convention on the Law of the Sea (UNCLOS) has 170 parties. It is the closest thing the world has to a constitution for the ocean. It defines territorial waters, exclusive economic zones, transit rights through international straits, and the rules governing who can extract what from the seabed. Most of the global economy floats on its provisions, whether or not anyone in the C-suite has read them.

But UNCLOS has a structural vulnerability that two governments exploited this week, from opposite ends of the same maritime corridor. Iran established a new bureaucracy to tax and vet ships transiting the Strait of Hormuz. Turkey began drafting a bill to codify its maritime claims across the Aegean and Eastern Mediterranean. Neither country has ratified UNCLOS. And both are now using that non-ratification as the legal basis for domestic laws and institutions that directly contradict the Convention’s framework.

I spent the early part of my career as a lawyer, and one thing you learn fast in that line of work is that the most dangerous party in any dispute is not the one making threats. It is the one drafting statutes (usually through lobbyists). Threats can be withdrawn. Statutes bind the government that wrote them, and every government that follows.

Iran: The toll booth at the mouth of the world

Subscribe now

On May 5, Iran announced the creation of the Persian Gulf Strait Authority (PGSA), an agency with a mandate to vet and tax every vessel seeking passage through the Strait of Hormuz. The PGSA will require ships to submit detailed ownership, insurance, crew, and cargo information before paying a toll. Some vessels have reportedly paid up to $2 million to the PGSA’s ad hoc predecessor, in Chinese yuan or cryptocurrency, for the privilege of transiting waters that international law says they already have the right to use.

The legal position is not ambiguous. Under UNCLOS, the Strait of Hormuz qualifies as an international strait subject to the transit passage regime, meaning all vessels enjoy an unimpeded right of passage that coastal states cannot suspend. Iran signed UNCLOS in 1982 but never ratified it, and has long maintained that only states party to the Convention benefit from the transit passage standard. Everyone else, in Tehran’s reading, falls under the older and more restrictive innocent passage framework, which coastal states can regulate and even suspend.

What the PGSA does is take that legal position and give it an office, a form to fill out, and a fee schedule. The blockade stops being a military operation and starts becoming an administrative process. That conversion matters. Military blockades are acts of war that provoke military responses. Administrative regimes are harder to challenge with destroyers. They create precedent. They generate documentation. They build institutional memory.

The human cost of this conversion is already severe. Approximately 1,500 ships and 20,000 crew remain trapped in the Persian Gulf. Traffic through the strait has collapsed to roughly 5% of pre-war levels. Insurance premiums on Gulf-transiting vessels have climbed from 1% to 10% of hull value. And for the 170 million barrels of oil sitting aboard tankers with nowhere to go, the PGSA represents not a path to resolution but a new layer of friction between them and open water.

Turkey: The bill that binds the next government

At the other end of the corridor, Turkey is preparing to submit legislation to parliament that would formalize its maritime jurisdiction claims in disputed Aegean and Eastern Mediterranean waters. The bill would codify Ankara’s rejection of UNCLOS-based boundaries and assert rights over potential offshore natural gas reserves in areas that Greece, Cyprus, and Israel consider settled.

Turkey, like Iran, is not a party to UNCLOS. Its maritime claims have until now existed as diplomatic positions, advanced in bilateral negotiations and occasionally enforced by naval presence. Moving those claims into domestic statute changes their character entirely. A diplomatic position can be traded away by a future foreign minister in exchange for something else. A statute constrains every government that follows until parliament repeals it.

That constraint function is the point. Devlet Bahceli, the nationalist coalition partner whose support keeps the Erdogan government in power, signaled in parliament that Turkey would respond forcefully to any initiative that ignores its sovereign rights or the rights of Turkish Cypriots. Codifying the claims in law locks Bahceli’s maximalist position into the legal architecture of the Turkish state, regardless of who holds power next.

The target is not abstract. The Great Sea Interconnector, a 1,208-kilometer subsea power cable linking Greek, Cypriot, and Israeli grids, has already been delayed by Turkish interventions. The EU has pledged EUR 657 million to the project, and a route survey must begin before mid-2026 if the cable is to be operational by 2030-2031. A Turkish statute asserting jurisdiction over the waters the cable would cross creates a domestic legal obligation for Turkish authorities to challenge any infrastructure project that proceeds without Ankara’s consent, regardless of who is in power in Turkey. Energy companies with Eastern Mediterranean exploration assets and infrastructure investors backing the Interconnector face a different category of regulatory risk than they did a week ago.

Subscribe now

The UNCLOS gap

Set these two moves side by side and the structural vulnerability becomes clear. UNCLOS works through voluntary compliance by ratifying states. It has no enforcement mechanism with teeth and no way to bind a country that never signed up. Iran and Turkey are not violating a treaty they agreed to. They are building domestic legal orders in the space that UNCLOS cannot reach, and daring the international community to do something about it.

Both moves serve the same strategic function: they raise the cost of reversal. Future Iranian leadership that wants to reopen the strait now has to dismantle a government agency. A future Turkish prime minister who wants to negotiate maritime boundaries with Greece now has to repeal a law. The status quo shifts from “this is our position” to “this is our law,” and laws are stickier than positions.

Iran and Turkey are not the only countries operating in this gap. The United States signed UNCLOS in 1994 but has never ratified it, which means the world’s dominant naval power invokes the Convention’s transit passage rights without accepting its dispute resolution mechanisms. Israel has neither signed nor ratified, a fact that complicates the very Eastern Mediterranean energy partnerships that Turkey’s bill is designed to disrupt. Venezuela, which withdrew from UNCLOS predecessor agreements and never ratified the Convention, has used that status to press maximalist claims against Guyana’s offshore resources in the Essequibo region. Syria, Eritrea, and Peru sit outside the Convention with active or latent maritime boundary disputes of their own.

The pattern is replicable anywhere a non-ratifying state sits on a contested waterway and has a functioning legislature. The South China Sea, the Bab el-Mandeb, the waters off the Horn of Africa. What Iran and Turkey demonstrated this week is that the move from military posture to administrative process only requires the political will to convert a claim into a statute and creates stricter boundary conditions for future attempts to resolve problems.

For the companies, insurers, and infrastructure investors whose supply chains cross these waters, the traditional risk model assumes military escalation as the primary chokepoint threat. What these two governments are showing is that the next escalation may arrive as a filing requirement on a government website rather than a kinetic strike. Both are equally capable of stopping commerce in its tracks. One of them is much harder to spot in real time.

Part One of the Frameworks Series

Every analyst has biases. The honest ones tell you what theirs are.

This is the first in a series of pieces where I’m going to do exactly that. Not to convince you to share my worldview, but to give you the intellectual scaffolding underneath my analysis. If you understand my overarching ideas about the world, you understand why I read specific events the way I do. You can agree with my short-term analysis even if you reject my long-term thesis. You can also do the reverse. Either way, you deserve to know where I’m coming from.

These frameworks are different from predictions. They operate at a different level of resolution entirely.

When I write about the next six months in the South China Sea, or what the next Federal Reserve decision means for emerging market exposure, those are specific claims. They can be falsified quickly. I’m happy to be wrong about them (and I’m wrong with some regularity).

The frameworks I’m describing in this series are different. They operate over decades. They are less amenable to rapid falsification and much more amenable to constant updating. The signal I care about is not whether a single event contradicts the framework. The signal I care about is whether repeated patterns of events are pointing in a different direction. If they are, then the framework has to change. That is a much bigger deal than any single wrong call.

Before you start the below, here is the warning: this post is MUCH longer than my normal posts (roughly 4x the size of a normal post and this is only one section of it). That is because it is meant to be something that provides context for all of my analysis and so is necessarily longer.

With those caveats fully on the table, here is one of my foundational frameworks for my analysis. It isn’t necessarily the most important as I’m not writing these as a ranking, but it’s one that has been on my mind quite a bit lately because of recent events.

Everything is about energy.

The Core Claim

When I look at the full sweep of human history, what I see is a series of eras defined not just by politics or culture or ideology, but by whatever the dominant energy problem was of that time.

Before the industrial revolution, one of the core recurring problems was feeding people.

That sentence sounds simple. It is not.

To feed people, you need to create and feed the systems that produce food. You need to feed horses and cattle and other livestock that plow fields and haul grain and carry soldiers and supplies. You need to feed the farmers and the craftsmen who maintain those systems. The energy that powers everything in a pre-industrial economy is organic energy. It moves through a chain: sunlight and water to crops to livestock to humans to labor.

Every political structure, every military strategy, every economic arrangement in the pre-industrial world was shaped by the constraints of that chain.

Land was the source of everything because land is where food comes from. The people who controlled land controlled energy. Feudalism - one such pre-Industrial Revolution system - was not primarily a social arrangement. It was an energy management system.

This is not an abstraction. When William the Conqueror commissioned the Domesday Book in 1086, what he produced was (functionally) an energy audit of England. The survey cataloged arable land, pasture, woodland, livestock, and the labor force attached to each manor. It recorded plowlands and plow teams. It counted mills. William was not interested in culture or religion or the happiness of his new subjects. He needed to know, with precision, how much productive capacity he had conquered. In modern terms, how much energy the land could generate, and who controlled it. The Domesday Book is one of the most remarkable administrative documents in European history, and its entire logic is caloric.

Ibn Khaldun saw the same dynamic around the Mediterranean three centuries later. His Muqaddimah, written in 1377, is often read as a theory of civilizational cycles. It is that. But underneath the cycles is also an energy argument. Nomadic peoples, in Khaldun’s framework, possess asabiyyah, a group solidarity forged by scarcity and hardship. They conquer settled civilizations that have grown soft through surplus. Then the conquerors themselves settle, enjoy the surplus, lose their cohesion, and fall to the next wave of nomads. The engine of the cycle is the relationship between a population and its energy base. Settled societies accumulate energy surplus in the form of agriculture and trade. That surplus breeds political complexity, inequality, and eventually the kind of fragility that makes conquest possible. Khaldun did not use the word “energy.” He did not need to. The logic is the same.

Fernand Braudel made an equivalent argument for the early modern Mediterranean. In his landmark 1949 study, The Mediterranean and the Mediterranean World in the Age of Philip II, Braudel argued that the political and military history of the sixteenth-century Mediterranean was subordinate to its material constraints: climate, soil, growing seasons, the caloric yields of wheat and olives and livestock. The grand strategies of the Ottoman and Habsburg empires were, at bottom, strategies for controlling agricultural hinterlands that could feed armies and navies. The Mediterranean was not shaped by kings. It was shaped by calories, which themselves are a measure of energy.

Even technological change within the pre-industrial era operated through this energy lens. The adoption of the heavy plow and the horse collar in northern Europe between the ninth and twelfth centuries did not merely increase crop yields. They shifted the center of European economic and political gravity. The heavy plow could turn the dense, clay-rich soils of northern France, the Low Countries, and the Rhine Valley that the older scratch plow could not handle effectively. The horse collar allowed horses to replace oxen for plowing, roughly doubling the speed of field preparation. Together, these technologies unlocked the energy potential of northern European land. The result was a population boom, urbanization, and a shift in political power from the Mediterranean rim northward. The rise of Carolingian and then Capetian France, the growing wealth of Flanders and the Hanseatic towns, the agricultural surplus that funded the cathedral-building campaigns of the twelfth and thirteenth centuries: all of it traces back to a change in how much energy could be extracted from a given acre of land.

The point is not that culture and politics and individual leadership do not matter. They do. The point is that they operate within constraints set by energy. No amount of political genius can overcome the caloric limits of your territory. No military strategy survives the exhaustion of the food supply that sustains it. The pre-industrial world was a world where energy was organic, local, and fundamentally limited by photosynthesis and the ability of soil and rain to convert sunlight into food. Everything else was downstream of that.

Subscribe now

The First Shift: When Destruction Got Cheaper Than Production

The story gets more interesting when you look at what happened in the century or two before the industrial revolution proper.

The introduction of gunpowder weapons changed the cost structure of warfare.

Here is the critical thing for me: for most of human history, the tools of destruction and the proper knowledge of their use required a substantial amount of labor and crafting (which itself required a lot of energy). A sword capable of killing someone in battle required significant metallurgical skill and substantial time to produce. A longbow required years of training to use effectively. The English longbowmen who devastated French cavalry at Crécy in 1346 and Agincourt in 1415 represented enormous investments: boys began training at age seven, and the physical demands were so extreme that the skeletons of longbowmen show measurable bone deformation in the left arm and spine. Armor required specialized smithing. A full set of plate armor for a mounted knight in the fifteenth century cost the equivalent of a small farm. The arms of war were expensive to produce and expensive to wield.

What gunpowder changed was the ratio between killing capacity and production cost.

A bullet is far easier to produce than the sword it replaces in lethality. A musket ball required no years of apprenticeship to fire. The Spanish tercio formations that dominated European battlefields from the early 1500s through the mid-1600s were built on this arithmetic: you could train a peasant to load and fire an arquebus in weeks, whereas training a competent cavalryman or longbowman took years. Gunpowder scales better than most pre-gunpowder armaments. A cannon is expensive, but one cannon replaces dozens of siege engineers and years of fortification work. The cost of destruction, per unit of damage inflicted, dropped.

The fall of Constantinople in 1453 was the proof of concept. The walls of Constantinople had stood for over a thousand years. They had withstood siege after siege. They represented the accumulated defensive investment of the Eastern Roman Empire across centuries. Mehmed II brought a Hungarian engineer named Orban, who built a cannon capable of firing stone balls weighing over half a ton. Those walls fell in weeks. The cost asymmetry was staggering: a thousand years of fortification rendered obsolete by a few months of cannon production.

The military historian Geoffrey Parker formalized this insight in what scholars now call the “Military Revolution” thesis. Parker argued that gunpowder weapons triggered a cascade of changes in European warfare between roughly 1500 and 1800. Fortifications had to be redesigned from the ground up. The medieval castle, with its tall thin walls designed to resist escalade and battering rams, was useless against cannon fire. The replacement was the trace italienne: low, thick, angled bastions designed to absorb and deflect artillery. These new fortifications were fantastically expensive. Only states with large tax bases could afford to build them. The trace italienne effectively priced smaller political units out of independent defense, concentrating military power in the hands of territorial states large enough and wealthy enough to build the new defenses and field the new armies. The modern European state system was, in significant part, a product of the cost structure of gunpowder warfare.

The economics at the individual level were just as decisive. Maintaining a single mounted knight, fully armed and armored, with warhorses, squires, and the land grants required to sustain the whole arrangement, cost roughly the equivalent of maintaining a small village. For the same expenditure, a ruler could equip and train a company of musketeers. The musketeers were less individually impressive on the battlefield, but the math was overwhelming. One knight or forty musketeers. The knight class did not disappear because society became more egalitarian. It disappeared because the cost-per-kill ratio shifted.

This had an important limit, though. Even with guns, you could still overwhelm a smaller, better-armed force through sheer numbers. Bodies still worked as a military instrument, even if the math of how many bodies you needed was changing. The Ottoman Empire, which fielded some of the largest gunpowder armies in the early modern period, relied heavily on sheer scale of manpower alongside its Janissary elite and artillery. Well into the early nineteenth century, numerical advantage could defeat technological advantage often enough to matter.

That limit disappeared with industrialization.

Subscribe now

The Industrial Inflection

Once you can mass produce weapons, the problem inverts.

You are no longer constrained by how many weapons you can make. You are constrained by how many people you have to use them. The factory produces rifles faster than the army can absorb them. The shell foundry outpaces the artillery units. The bottleneck moves from production to deployment.

This created a genuine strategic problem for the rulers of the industrializing world. They needed large numbers of people not just to produce, but to fight and die in the numbers that industrialized warfare demanded.

The casualty figures tell the story with brutal clarity.

At the Battle of the Somme in July 1916, the British Army suffered 57,470 casualties, including nearly 20,000 dead. That single day exceeded the total British casualties from the entire Boer War. At Verdun, the ten-month battle between February and December 1916 produced over 700,000 combined French and German casualties. On the Eastern Front, the scale was even greater: Russia mobilized over 12 million men during the war and suffered casualties estimated between 7 and 9 million. The American Civil War, half a century earlier, had already demonstrated the pattern: approximately 620,000 dead in a country of 31 million, a proportional bloodletting that the United States has never come close to repeating (at roughly 2% of the total population, that would be a modern equivalent of over 6 million people).

These numbers are the logical output of a system that tried to throw massive numbers of human bodies at industrially produced killing capacity on both sides.

But here is the political consequence that I think gets underweighted in most historical analysis.

You cannot ask millions of people to fight and die for a system that gives them nothing. You cannot sustain conscript armies drawn from populations that have no stake in the political order.

The pivotal moment was the French Revolution’s levée en masse of 1793. Faced with invasion by a coalition of European monarchies, the revolutionary government declared that every citizen owed military service to the nation. This was a new theory of the relationship between the individual and the state. The citizen-soldier replaced the mercenary and the feudal levy. The bargain was explicit: you fight because this is your republic, because you have rights within it, because you are a part of the state. And it worked. The armies of revolutionary and Napoleonic France defeated larger, better-trained professional armies across Europe for two decades. The citizen-soldier, motivated by political participation, fought harder than the paid professional or the pressed conscript of the old regime.

The rest of Europe learned the lesson. Slowly, but they learned it.

The expansion of democratic rights in the nineteenth and twentieth centuries was not simply an expression of Enlightenment ideals working their way through history. It was, at least in significant part, a strategic necessity. The British Reform Acts of 1832, 1867, and 1884 progressively expanded the franchise, and the timing is instructive. Each expansion correlated with a period of either active military engagement or rising anxiety about military preparedness. The 1867 Reform Act, which roughly doubled the electorate by extending the vote to urban working-class men, came in the aftermath of the Crimean War and amid growing concern about Prussian military power. The 1884 Act extended the vote to rural laborers, the very population from which the army drew most of its recruits. The pattern was not coincidental. Disraeli, who championed the 1867 Act, was explicit that a wider franchise would produce a more loyal and militarily useful population.

Bismarck understood the same logic from a different angle. His introduction of social insurance in the 1880s, including health insurance, accident insurance, and old-age pensions, was not an act of progressive generosity. It was a calculated effort to bind the German working class to the state and undercut the appeal of socialism, at a time when Germany was building the largest conscript army in Europe. The welfare state was a loyalty-purchasing mechanism. You could not ask men to die for Kaiser and country if they had no stake in either.

The sociologist Charles Tilly captured the deeper pattern in a phrase that has become foundational in political science: “War makes states, and states make war.” Tilly’s argument, drawn from his analysis of European state formation from 1500 onward, was that the demands of warfare drove states to develop extractive and administrative capacity, and that this capacity in turn required bargains with the populations being extracted from. Taxation required representation, or at least the credible promise of it. Conscription required citizenship, or at least the credible promise of it. The modern democratic state was not a gift from enlightened elites. It was a bargain, driven by the need to mobilize populations for industrial-scale warfare.

The same logic played out in the United States. The service of roughly 180,000 Black soldiers in the Union Army during the Civil War was a direct catalyst for the Thirteenth, Fourteenth, and Fifteenth Amendments. The argument was simple and politically powerful: men who had fought and died for the republic could not be denied citizenship within it. The same dynamic recurred after both World Wars. The GI Bill of 1944, which provided education, housing, and business loans to returning veterans, was the twentieth century’s most significant single act of democratic wealth distribution. It was not charity. It was the fulfillment of a bargain: you fought, so you get a stake in the society you defended. Women’s suffrage, too, was accelerated by war. Women’s entry into industrial production during World War I, filling the factory jobs vacated by men sent to the front, eroded the argument that women had no role in the economic and military life of the nation. The Nineteenth Amendment in the United States and equivalent legislation across Europe followed within a few years of the armistice.

The pattern repeats with such consistency that it begins to look less like coincidence and more like a structural relationship.

Democracy, in other words, had a structural underpinning in the industrial era. It was not just a moral achievement. It was operationally necessary. Rulers who wanted to win wars needed populations that were willing to fight, and populations became willing to fight when they had genuine stakes in the outcome.

The feudal bargain stopped working the moment you needed serfs to die by the hundreds of thousands in industrial warfare. A new bargain had to replace it.

Give people the vote, give them property rights, give them the idea that they are fighting for something that belongs to them, and they will fight. Deny them that, and you have a revolution or a collapse before the enemy even arrives. The Russian Empire learned this in 1917. The Ottoman Empire learned it across the same period. The Austro-Hungarian Empire learned it the same year. Three of the four great empires that entered World War I on the basis of the old monarchical bargain ceased to exist by 1918. The one that survived, the British Empire, had spent the preceding century progressively extending the democratic bargain to its domestic population.

That structural underpinning is now eroding. Fast.

Subscribe now

The Current Shift: When People Become the Vulnerability

Here is where the framework arrives at the present moment.

Drones and AI-enabled targeting systems are doing to the people-in-warfare equation what the industrial revolution did to the craftsmen-in-production equation. They are making human bodies operationally unnecessary in new and important ways.

This is not speculation. We are watching it happen.

In September 2020, Azerbaijan launched a military offensive to retake Nagorno-Karabakh from Armenian forces. The conflict lasted 44 days. It was, by the assessment of virtually every serious military analyst who studied it, the first war in which drone warfare played a decisive role. Azerbaijani forces used Turkish-made Bayraktar TB2 drones and Israeli-made loitering munitions to systematically destroy Armenian armor, artillery, and air defense systems. Armenian forces, which had fortified the region over nearly three decades and which possessed a conventional military advantage on paper, were unable to adapt. The losses were devastating: estimates suggest that Armenian forces lost the majority of their tanks, armored vehicles, and artillery pieces, many of them destroyed before they could engage.

The cost ratios were the critical detail. A Bayraktar TB2 drone costs roughly $1-2 million. An Israeli Harop loitering munition costs a few hundred thousand dollars. The Armenian T-72 tanks they destroyed cost $2-3 million each. The air defense systems they neutralized cost tens of millions. The arithmetic was running in one direction, and it was not the direction of the defending army.

Ukraine has turned this from a proof of concept into a laboratory. Since February 2022, the conflict has seen drone warfare evolve at a pace that has stunned military planners worldwide. FPV (first-person view) racing drones, modified with explosive payloads, cost between $400 and $2,000 to produce. They are being used to destroy armored vehicles, fortified positions, and supply lines. A $500 drone destroying a $3 million infantry fighting vehicle is not an anomaly. It is the new cost structure of ground warfare. Both sides are now producing these drones in quantities that would have seemed absurd two years ago. Ukraine’s stated production target for 2024 was over one million FPV drones. Russia is scaling similarly.

The Houthi drone and missile strikes on Saudi Aramco facilities at Abqaiq and Khurais in September 2019 demonstrated the same cost asymmetry in a different domain. A combination of drones and cruise missiles, costing perhaps a few million dollars in total, temporarily knocked out roughly half of Saudi Arabia’s oil processing capacity. The targeted facilities were protected by billions of dollars worth of air defense systems, including American-made Patriot batteries. The defense failed. The attack succeeded. The ratio between the cost of attack and the cost of the assets destroyed or disrupted was on the order of 1:1000 or worse.

And then there is the AI targeting layer.

Israel’s use of AI-assisted targeting systems in Gaza, reported in detail by independent journalists and confirmed in broad outline by Israeli military officials, represents the next stage. Systems reportedly capable of generating target recommendations at a pace that exceeds human analytical capacity. The human role in the targeting cycle compresses toward approval rather than identification. The speed of the system shapes the tempo of operations.

The United States military’s Project Maven, initiated in 2017, was designed to apply machine learning to drone surveillance footage, automating the identification and tracking of objects and individuals across vast volumes of video data. It was one of the first major military AI project, and it signaled where the trajectory was heading: toward systems that see, identify, and recommend faster than any human staff can process.

Think about what a drone swarm actually represents at scale.

A drone swarm can move street by street, building by building, and identify and engage individual targets with precision that was not achievable at this cost point even a decade ago. It can do this without risking a single soldier. The production of the drones is industrial. The targeting is increasingly algorithmic. The human role shrinks to supervision, strategy, and maintenance.

Now scale that. Tens of thousands of drones, produced at industrial scale (perhaps in a factory primarily made of robotics?), guided by AI targeting systems operating faster than any human command structure, require something close to what we needed for industrial-era warfare, but with one critical difference: energy, not people.

What do data centers run on? Energy. What do drone production facilities need? Energy. What do the charging and logistics systems for autonomous weapons require? Energy. But the energy they run on is not the food-based energy of the past, but rather electricity generated in whatever way is most convenient, be it coal, oil, natural gas, solar, wind, or nuclear. The strategic resource is shifting from human bodies to kilowatt hours and the infrastructure to produce those hours.

We have no sense yet of the split between civilian and military uses of these data centers, but already the numbers are already staggering. Global data center electricity consumption was approximately 460 terawatt-hours in 2022, roughly 2% of total global electricity demand. The International Energy Agency projects that figure could more than double by 2026, driven almost entirely by AI workloads. Microsoft signed a deal to restart a reactor at Three Mile Island to secure power for its AI operations. Amazon has been buying nuclear-powered data center capacity. Google, Meta, and others are locked in a competition for energy supply that increasingly looks like a strategic resource contest rather than a business procurement exercise.

AI requires energy at a scale and reliability that pushes the entire competition toward whoever can secure the most, the most reliably, at the lowest cost. The same AI that enables military targeting runs on the same infrastructure that powers commercial AI. The dual-use implications are direct and unavoidable.

China’s dominance in rare earth processing, which is necessary for the permanent magnets used in both drones and data center cooling systems, adds another layer. The semiconductor fabrication plants that produce the chips powering AI systems consume enormous quantities of energy and ultrapure water. A single advanced TSMC fabrication facility uses as much electricity as a small city. Control over energy supply chains, chip fabrication capacity, and rare earth processing is not a commercial competition. It is a strategic arms race conducted through industrial policy.

The Gulf states are reading this correctly. Their pivot from positioning themselves as oil exporters to positioning themselves as energy-for-compute hubs, investing in data centers, AI research parks, and sovereign AI capacity, is not a branding exercise or an investment fad. It reflects a strategic calculation: the value of their energy reserves is shifting from fueling cars and factories to fueling the computational infrastructure that will underpin military and economic power in the coming decades.

And here is the consequence that follows directly from this shift.

Population centers, which were always an asset in the industrial era because they represented recruiting potential and productive capacity, are now also a profound vulnerability. This vulnerability is not necessarily new: air power during World War II exposed the same vulnerability. Dense human populations are easier to target with the tools being developed now though because, unlike World War II, the cost of air power and targeting has plummeted. The people who were once the core asset of a nation’s military power are now also its most exposed liability and anti-drone and anti-missile defenses can only provide so much protection.

The side with more people does not necessarily win anymore. The side with better energy access, better drone production (with better anti-drone and missile defense), and better AI targeting capability wins. And that side may have a much smaller population than its adversary.

Subscribe now

What This Means for Democracy

This is the part that I want to be careful about, because it is easy to misread as a claim about what should happen. It is not. It is a claim about what the structural logic points toward.

Democracy is not eternal. It emerged in its modern mass form because of a specific set of conditions: industrialized warfare required mass armies, mass armies required politically engaged populations, politically engaged populations demanded political rights. The causal chain ran from military necessity to political structure. And all of that happened much more recently than we generally like to think about.

If that causal chain is now being severed at its first link, the downstream political consequences should concern us.

Consider the pattern from the other direction. Where mass military participation has not been necessary, democracy has generally not emerged or has remained fragile. The resource-rich Gulf monarchies are the clearest contemporary example. States that derive their revenue from oil and gas extraction rather than from taxing a productive population face a fundamentally different set of political incentives. They do not need the population’s labor in the same way an industrial economy does. They do not need the population’s military service in the same way an industrial military does. The political bargain, accordingly, looks different: generous public services and material comfort in exchange for political quiescence rather than political participation. The so-called “rentier state” model has been extensively studied in political science, and its core finding is consistent with the energy thesis: when rulers do not depend on the population for revenue or military power, the structural incentive to extend political rights weakens dramatically.

Democratic institutions have been under pressure for years, and most of the analysis of that pressure focuses on cultural explanations: polarization, misinformation, the failures of political parties to deliver for their constituents. Those factors are real. But underneath them is something structural that receives much less attention.

When those in power (i.e. governments) no longer need large populations to fight their wars, one of the core strategic incentives for maintaining broad democratic participation weakens. This is not a short-term phenomenon. It plays out over decades. But the direction of the logic is not reassuring.

The governments that are most aggressively investing in drone and AI-enabled military capability are not, as a group, the world’s most committed democracies. At best, we can say they are a mixed bag. Ukraine, China, Israel, the UAE, Turkey, Iran, Russia: the list of states leading in military drone deployment and autonomous weapons development reflects a hodge-podge or democratic, authoritarian, and hybrid regimes. This reflects a strategic intuition, perhaps not fully articulated, that the new tools of power do not require the old political bargains.

At the same time, the democracies investing in these technologies face their own version of the problem. The United States, the United Kingdom, France, and others are developing autonomous weapons systems within political frameworks that still assume mass military participation as a baseline. The all-volunteer force, the draft registration requirement, the cultural expectation that citizens may be called to serve: these persist as political and legal structures even as the operational reality moves away from them. If drone and AI capability is what wins wars now, and if a hodge-podge of regime types are all converging on the same technological investments, then the distinguishing advantage of democracy in warfare is no longer the willingness of citizens to fight. It is something else, or perhaps it is nothing. That question has not been answered yet, and the gap between the old political framework and the new military reality creates a kind of structural dishonesty that erodes democratic legitimacy from the inside.

If you do not need the population to fight, you still need them to work and to consume. That is true (for now). The economic argument for democratic participation persists even if the military argument weakens. But the historical record suggests that the military argument has been the sharper of the two drivers. Elites throughout history have proven willing to tolerate enormous economic inequality and limited economic participation as long as they could maintain political control. What they could not tolerate was military weakness resulting from a disengaged or rebellious population. Remove that specific pressure, and the balance of incentives shifts.

Subscribe now

Why This Framework Matters for Geopolitical and Business Risk Analysis

I am not writing this as a purely academic exercise.

If this framework is correct, or even partially correct, several things follow for how decision-makers should think about risk.

First, the traditional demographic analysis of national power is becoming less reliable. Population size and growth rates have long been inputs into assessments of national strategic capability. The assumption that larger populations translate into greater military potential, which has been broadly true since at least the Napoleonic era, is now open to question. A nation of 10 million with superior energy infrastructure, advanced drone production capacity, and cutting-edge AI capability may pose a greater strategic challenge than a nation of 100 million without those assets (incidentally, Israel’s population is just under 10 million, while Iran’s is around 93 million). The 2020 Nagorno-Karabakh war was, among other things, a demonstration of this principle: Azerbaijan’s population advantage over Armenia’s Karabakh forces was real, but it was the technological and industrial asymmetry in drone capability that proved decisive.

Second, energy infrastructure is becoming a primary strategic asset in ways that go beyond what most corporate geopolitical risk frameworks currently capture. Control of energy supply chains, the vulnerability of energy infrastructure to disruption, and the competition for energy access are not simply economic issues. They are security issues, and they are escalating as such. The targeting of energy infrastructure in the Ukraine conflict, where both sides have systematically attacked power generation and transmission systems, is a preview of what strategic competition looks like when energy is the binding constraint on military capability.

Third, the political risk associated with democratic backsliding needs to be assessed against a more fundamental question: what structural conditions are required to sustain democratic governance, and are those conditions still being reproduced? If the energy thesis has any merit, then democratic erosion is not simply a function of bad political actors or weak institutions. It reflects a deeper shift in the strategic incentives that originally produced democratic governance.

Fourth, and perhaps most directly relevant to the executives who read this: the firms most exposed to geopolitical disruption in the next decade are likely those whose supply chains, customer bases, or operational footprints are concentrated in regions where the tension between the old political bargain and the new strategic logic is sharpest. That is not a small set of regions.

A Note on Being Wrong

I want to come back to something I said at the start.

These frameworks are hard to falsify in the short run. That is both their strength and their weakness. Because they operate over long time horizons, I cannot point to a single quarter’s events and declare them proven or disproven. What I can do is hold them up against the accumulation of evidence, revise them when the pattern of evidence demands it, and be honest with you about where the evidence is putting pressure on the thesis.

Right now, I think the evidence is largely consistent with the energy framework. The strategic investment patterns of the world’s major powers, the trajectory of drone and AI development in military contexts, the observable pressures on democratic institutions across multiple continents, and the emerging competition over energy infrastructure all point in directions the framework would predict.

That does not mean the framework is right. It means it has not been wrong yet in the ways that would matter.

If large-scale democratic militaries with conventional mass armies prove decisively superior in the conflicts that define the next two decades, I will update. If the demographic and industrial models of national power prove more durable than the energy-autonomy model, I will revise. I will tell you when I do.

Subscribe now

At Least Five Reasons I Can Think of for Why I’m Wrong

I promised at the start to be honest about my biases. That means being honest about the weaknesses in my own framework. Here are at least five arguments against the thesis I just laid out, along with my responses. I am not being generous with these objections for rhetorical effect. Each of them has real weight.

Democracy came before the mass army, so the causal arrow is wrong. The English Bill of Rights in 1689, the American Revolution in 1776, Athenian democracy two millennia before any of this: democratic governance predated industrial warfare by centuries. If democracy existed before mass conscription, then mass conscription cannot be the cause.

This objection is weaker than it looks. The examples it cites are not democracy in any sense that a modern citizen would recognize. They are examples of landed elites wresting power from a much smaller group of elites. The English Bill of Rights transferred authority from the Crown to Parliament, which was a body of aristocrats and wealthy landowners. The American Revolution produced a republic in which only propertied white men could vote. Athenian “democracy” excluded women, slaves, and non-citizens, meaning the vast majority of the population. In each of these cases, what was being distributed was not political power to the population. It was political power within the elite.

Anything resembling full suffrage did not appear until the industrial era. And it appeared at moments of industrial shift. Women got the vote when they became a larger part of the industrial workforce, not when philosophers wrote compelling arguments for their inclusion. The expansion of the franchise to working-class men tracked with the expansion of conscript armies. The timeline supports the thesis better than the objection suggests.

Drone warfare still requires massive human infrastructure. Engineers, maintainers, intelligence analysts, logistics personnel, cybersecurity teams, satellite operators, supply chain workers, and the industrial labor force that produces the components. Ukrainian soldiers are hand-building FPV drones in field workshops. Wars still require occupation, stabilization, and governance, none of which drones can do.

This is correct. But it misreads the argument as an absolute claim rather than one about degree. Of course people are still required. Just not in the same numbers. Degree matters. We still use horses, after all, just not for anything in the food system. In developed economies, horses are almost exclusively a plaything for the rich. They went from being the foundation of the entire pre-industrial energy chain to being a luxury recreational asset. The shift from mass armies to smaller, more technically skilled forces does not need to be complete for the political consequences to begin. It only needs to be directional. And the direction is clear.

Democracies outperform autocracies economically, and that matters independently of military necessity. Rule of law, property rights, contract enforcement, regulatory predictability, and peaceful transfers of power all support economic performance. Authoritarian regimes face brain drain, capital flight, corruption, and succession crises. Singapore, South Korea, and Taiwan democratized for economic reasons, not military ones.

This objection has a data problem. We do not have models for understanding what economies or growth look like except in a world where the largest and most powerful countries are democracies. Essentially all of our data on the economic performance of democratic vs. authoritarian systems is post-World War II data, generated in an international order designed and maintained by democratic great powers. The institutions that enforce property rights, facilitate trade, and stabilize currencies were built by and for democracies. We do not know what economic incentives would look like in a world where the dominant powers are not democratic, because we have not lived in that world during the era of modern economic measurement. That is a problem with the background data, not with the thesis.

Nuclear weapons already broke the mass-army model, and democracy expanded anyway. If the argument is that rulers who do not need mass armies will not sustain democracy, the strongest version of that shift happened in the late 1940s. Nuclear weapons made mass conventional warfare between great powers obsolete. And yet the period from 1945 to 1990 saw the greatest expansion of democratic governance in human history.

Nuclear weapons are, for all intents and purposes, unusable. Mutually assured destruction took them off the table as instruments of actual warfare almost immediately. They are a deterrent, not a weapon. The expansion of democracy during the Cold War was, to a significant degree, a product of what the great powers wanted. The United States actively promoted democratic governance as a strategic tool against Soviet influence. Decolonization proceeded along lines that the great powers shaped, sometimes toward democracy and sometimes not, depending on the strategic calculation of the moment.

More importantly, the shift from mass armies to smaller, more capital-intensive forces did not start with drones. Fighter jets, tanks, precision-guided munitions, and strategic bombers all had similar effects on the ratio of technology to manpower in military operations. Drones and AI are the most recent and most dramatic examples, but the trajectory has been running in one direction since at least 1945. Again, this is about degree and direction over decades. Not single moments.

Norms, institutions, and path dependence are sticky. Democracy does not dissolve just because the original conditions that produced it have changed. Courts, legislatures, free press, civil society, and professional military officers trained in democratic norms all create constituencies that defend democratic governance. Political institutions outlast the conditions of their creation.

Path dependence is real. Norms and institutions are real. I would argue they are the primary reason democratic institutions persist as long as they do even as structural conditions shift underneath them. But persistence is not permanence. The Soviet Union and Communist China both demonstrate that political institutions, norms, and social structures can be remade on a vast scale, and that widespread internal death ends up being, from the cold perspective of regime survival, a blip. The Soviet system killed millions through collectivization, famine, and purges, and persisted for seven decades. The Chinese Communist Party presided over the Great Leap Forward, the Cultural Revolution, and Tiananmen, and remains in power. Those transformations were accomplished with mid-twentieth-century technology. It is at least worth asking what a regime with twenty-first-century surveillance, autonomous enforcement, and algorithmic social control could do with the same willingness.

The one argument none of these objections addresses.

There is one counterargument that I find genuinely difficult to contend with, and none of the five above quite captures it and I’m not sure what I think it about it yet: mass consumerism as an economic engine.

The modern global economy runs on consumption. Consumer spending represents roughly 60 to 70 percent of GDP in most advanced economies. That consumption requires a population with disposable income, and disposable income on a mass scale has historically been a product of the democratic bargain: wages, labor rights, social safety nets, and the expectation of rising living standards. If you hollow out the population’s political power, you risk hollowing out their economic participation, and that collapses the demand side of the economy that the entire system depends on.

This is not a military argument. It is an economic one. And it may be the strongest structural support democracy has left. The question is whether it is strong enough to sustain democratic governance on its own, without the military necessity that originally produced it. Henry Ford understood the logic a century ago when he paid his workers enough to buy the cars they built. Whether that logic survives an era of autonomous production and algorithmic resource allocation is, I think, genuinely open. But of the arguments against my thesis, this is the one I take most seriously.

In the next piece in this series, I will take up the energy question directly: what the shift from organic to industrial to what comes next actually looks like, and what it implies for where strategic competition is headed in the next twenty years.

The short version is that the most important competition on earth right now is not for territory. It is for energy. Not a new idea (I know), but perhaps with different implications when taken with the above.

But that requires its own article.

This is the first in an ongoing series exploring the overarching frameworks that shape my analysis of geopolitical risk and long-term trends. The goal is not to give you predictions. The goal is to show you the map I am using, so you can judge for yourself whether it is the right one.

There is an old military saying that captures the rhythm of every conflict, every deployment, every operation that runs on political time rather than operational time: “hurry up and wait”.

The soldiers sprint to the landing zone, fully loaded, ready to move. Then they sit. Not because the mission has changed. Because the orders haven’t arrived yet or the time hasn’t yet arrived for the mission itself.

That is where the global economy is right now.

The US military has staged the largest force posture in the Middle East since 2003. What began on February 28 as a joint US-Israeli air campaign has expanded, by the final week of March, into the largest regional troop deployment since the Iraq War: Marines, 82nd Airborne, amphibious assault ships, two carrier strike groups. The United States has not crossed the threshold into a ground war. But it has positioned itself to do so with limited delay. That is a deliberate ambiguity.

Meanwhile, the Strait of Hormuz has opened and closed three times in sixty days. The last opening lasted less than 24 hours.

And this morning, Brent crude rose 3 percent to $111.49 a barrel, even as Iran floated a proposal to reopen the strait because traders no longer believe the proposals.

The world has hurried up. Now it waits. And waiting, done badly, is how organizations get surprised.

Subscribe now

The Phantom Ceasefire

Two months after the war started, peace talks are on hold, with control of the Strait of Hormuz and the future of Iran’s nuclear program as the two main points of contention. The recent timeline helps understand why markets keep getting this wrong.

April 7: Trump warns that “a whole civilization will die tonight” if Iran doesn’t reopen the Strait of Hormuz. Hours later, a two-week ceasefire is announced.

April 11: US and Iranian officials meet in Islamabad for 21 hours of face-to-face talks.

April 12: JD Vance announces the sides have not reached an agreement.

April 13: The US implements a blockade of Iranian ports. April 17: Iran announces the strait is open.

April 18: Iran announces the strait is closed again, blaming US “breaches of trust.”

Today, Iran offered to end its chokehold on the strait if the US lifts its blockade, but would set aside discussions on Tehran’s nuclear program for later.

The Trump administration appears unlikely to accept. And an Iranian army spokesperson said this morning that Iran is still in a “war situation.”

This is not a negotiation approaching resolution. This is a negotiation that cannot resolve because neither side has agreement within itself about what resolution looks like. The Iranian government is fractured between hardliners demanding maximalist terms and pragmatists who understand the economic bleeding. Trump appears wary of restarting the war and would prefer a diplomatic resolution, but the ceasefire extension is not indefinite, and the US military is standing ready to resume strikes.

When both principals are simultaneously signaling war and deal, you are watching two organizations trying to manage internal dissent through external positioning - not a negotiation. No deal closes under those conditions.

What the Physical Market Knows That the Futures Market Doesn’t

Here is the number that should be on every CFO’s screen right now.

Futures markets have drifted back toward $97 per barrel for Brent in the expectation of eventual resolution. But the physical oil market stands at $132 per barrel as of April 28. That $35 premium reflects genuine scarcity, not market sentiment.

The market is not pricing the conflict. It is pricing Trump’s social media feed, and trading algorithms are doing the rest.

The physical market is priced by people who actually have to move barrels. Before the war, an average of 129 tankers per day transited the strait. As of last week, just three. Shipments through the Strait in early April averaged around 3.8 million barrels per day, against more than 20 million in February. Goldman Sachs estimates that the blockade and attacks on regional energy infrastructure have reduced global oil production by 14.5 million barrels per day.

The futures market is priced by people betting on when the headline changes.

If your procurement, hedging, or operational planning is built on futures prices, you have a $35-per-barrel blind spot in your model. At scale, for energy-intensive businesses, that is an existential risk exposure, not a rounding error.

Subscribe now

The Wave You Haven’t Felt Yet

Oil shocks do not arrive all at once. They transmit through supply chains, through input costs, through freight rates and insurance premiums and fertilizer prices with lags ranging between weeks and quarters. The disruption started February 28. The full economic weight of it has yet to land.

This is a classic stagflationary shock. Oil price increases hit consumers and businesses almost immediately. But the knock-on potential effects of slower investment, compressed margins, reduced consumer spending, compound over time. Central banks find themselves pulled in opposite directions: higher inflation implies tightening, but slowing growth implies easing.

The Federal Reserve, ECB, Bank of England, and Bank of Japan are all scheduled to announce policy decisions this week. Every one of them faces the same impossible question. None of them has a good answer.

The IMF has already laid out the math. In the reference scenario (a short-lived conflict) global growth falls to 3.1% and inflation rises to 4.4%. In an adverse scenario, growth drops to 2.5% and inflation climbs to 5.4%. In a severe scenario, growth reaches 2% (the IMF’s definition of global recession) while inflation exceeds 6%.

Today’s shock echoes the 2022 commodity price surge following Russia’s invasion of Ukraine. In that episode, the subsequent synchronized tightening and disinflation without a recession is widely seen as a major policy success. There are reasons to doubt we get the same outcome now. In 2022, the labor market was tight and balance sheets were flush. Today, the buffers are thinner and inflation expectations are already sensitized by that prior episode.

There is also a channel most corporate planning models are not capturing. More than 30% of global urea — the primary agricultural fertilizer — is exported from Gulf countries through the Strait of Hormuz. Food inflation is not a downstream risk. It is already baked in, regardless of when the strait reopens. The planting season runs on its own clock, and it does not wait for peace talks.

The Hard Floor on Normalization

Even if a deal is struck tomorrow, assume nothing returns to normal quickly.

Iran lost track of mines it planted in the Strait of Hormuz and is therefore unable to fully open it even if it wants to. The Pentagon told Congress that clearing the mines could take up to six months. Mine-clearing operations cannot begin in earnest until hostilities fully end. That means the earliest plausible return to pre-war transit volumes is late 2026 AT BEST.

Iran has entered negotiations demanding formal recognition of Iranian sovereignty over the Strait of Hormuz and a $2 million transit fee per vessel, a levy that would add approximately $2 per barrel to the cost of Gulf oil. If any version of that demand survives into a final agreement, it would be a permanent structural repricing of global energy that will function as a new tax, embedded in the cost of every barrel that moves through the Gulf.

No corporate risk model built on pre-February assumptions captures that scenario.

Subscribe now

What Executives Should Be Doing During the Wait

The military understands that waiting is not passive. A unit staged at the landing zone is conducting radio checks, confirming routes, running contingencies, rotating rest. The wait is active. The organization that treats the wait as downtime is the one that gets surprised when orders arrive.

Three things deserve attention right now for your organization as you wait for the inevitable.

First, even in the best-case scenario, build the six-month mine-clearing floor into every supply chain and procurement planning window. Normalization will take months. Mines need clearing. Insurance premiums will not fall until shipping companies believe passage is genuinely safe. Anything shorter is optimism, not planning.

Second, model the split market for your company. The $35 spread between physical and futures Brent is not an anomaly to be explained away. It is the market telling you that people with actual exposure are pricing something different than people with paper exposure. Know which one your business sits in.

Third, watch the ceasefire architecture, not the ceasefire headline. What matters is not whether a deal is announced, but whether it contains a verified timeline for mine clearance, a US naval blockade drawdown, and a durable framework on Iran’s nuclear file. Iran’s latest proposal would reopen the strait while leaving the nuclear question for later negotiations. The US has not accepted. Until those three elements are present in a signed agreement, any “opening” of the strait should be treated as temporary.

Subscribe now

The Closer

Sixty days in, this crisis has produced one clear strategic failure: the assumption, made by markets, by some boardrooms, and apparently by some in the Trump administration itself, that the other side would blink quickly.

In briefings held before the February attacks, the Joint Chiefs warned President Trump that Iran might close the strait. Trump acknowledged the warnings but dismissed the possibility, telling his team Iran would capitulate instead.

They did not capitulate. And now the global economy is waiting: staged, loaded, and unable to move for a resolution that keeps not arriving.

The soldiers at the landing zone know what to do with a wait. They stay ready. They run their contingencies. They don’t assume the orders are coming in the next five minutes.

That is the right posture for every executive managing exposure to this crisis right now. Hurry up. Get your scenarios stress-tested, your procurement hedged, and your board briefed.

Then wait, but actively. Because the order to move will come without warning.