It was 2 a.m. in Prague when the number landed. Not the two million barrels. The twenty.
Iraq had just moved two million barrels of crude through the Strait of Hormuz โ but not the way anyone expected. Ship-to-ship transfer, the first time the country had done it in decades. And to make buyers show up, it shaved twenty dollars off every barrel.
Twenty dollars. On crude that trades somewhere in the seventies. That's a quarter of the value, vaporized, just to convince someone to take the cargo.
I have seen this exact math before. Not on a tanker โ on a dashboard. In 2020 I helped launch a yield aggregator out of my Prague apartment, and we paid triple-digit APYs to anyone who would deposit. The moment the incentives dried up, the TVL evaporated like a Prague frost in April. Different number. Same logic. You don't need to blockade a chokepoint. You just need to make it expensive enough that people pay to go around it.
The Strait of Hormuz is the most important piece of plumbing on Earth. Roughly a fifth of global oil squeezes through a channel that, at its narrowest, is about 21 miles wide. Two-thirds of that water is Iranian. That geography isn't an accident of nature. It's a strategic fact Iran has spent decades learning to weaponize.
Here's what the reporting gets right, and what most crypto people will miss: the blockade never actually happened. There was no formal closure. No announcement. What happened is subtler โ a soft blockade. Insurance rates crept up. Shipowners got nervous. Risk became unacceptable, and when risk becomes unacceptable, the market does the blockading for you. Iran didn't have to fire a shot at a tanker. It just had to make the strait feel dangerous enough that nobody wanted to be inside it.
I've been thinking about this in the context of decentralization for years, because the same logic keeps surfacing in our own stack. We treat "decentralization" like a checkbox. But decentralization is really about one thing: eliminating chokepoints โ places where a single actor can extract rent, censor, or coerce. Hormuz is a chokepoint. So is a centralized exchange. So is a single sequencer. So is the one RPC provider your wallet quietly depends on every time you sign.
And here's the uncomfortable part. The soft blockade proves you don't have to own a chokepoint to weaponize it. You just have to make it risky.
Let me get concrete, because this is where the story stops being about oil and starts being about us.
Iraq's problem is structural. The country exports roughly 3.3 to 3.5 million barrels a day, and more than 90% of government revenue comes from oil. It has no meaningful pipeline capacity to bypass the strait โ no equivalent of Saudi Arabia's East-West line or the UAE's ADCOP route. So when Hormuz gets risky, Iraq has exactly one move: pay up, or stop shipping.
That is a single point of failure at national scale. And the $20 discount is its price tag. It's the market telling Iraq: your route is worth less than everyone else's.
Now map that onto Layer2.
For two years, the industry has sold "decentralized sequencing" as the endgame. But if you've actually traced a transaction through most rollups, you know the truth. There is usually one sequencer, run by one team, on infrastructure that looks a lot like a single node. The roadmap says it's temporary. The roadmap has said that since 2022. In the meantime, that sequencer is a chokepoint โ a place where transactions can be ordered, delayed, or censored, where one operator decides your transaction's fate.
You don't need to blockade that sequencer to make it dangerous. You just need people to worry about it. A rumored outage, a regulatory subpoena, a botched upgrade โ and suddenly the "decentralized" L2 carries a risk premium. Users route around it. Bridges get skittish. Liquidity migrates to wherever perceived risk is lowest. The market does the sequestering for you. This is the soft blockade, digitized.
I'll give you a personal example. Based on my audit experience, the most dangerous thing in any system is never the code you can read. It's the dependency you never thought to question. In 2020, at VaultPrime, I was so busy celebrating 300% APYs that I never traced the oracle underneath them. When that oracle got manipulated, two million dollars drained in a single block. The chokepoint wasn't the smart contract. It was the single price feed everyone trusted and nobody audited.
We didn't lose that money because someone blockaded us. We lost it because we'd built a system with one throat to choke โ and then acted surprised when it got choked.
Same story in 2021. Two hundred people, one loft, one mint contract with a gas limit nobody stress-tested. The floor spiked, the contract failed, the chain congested, and I spent the next month reimbursing friends out of my own pocket. One contract. One chokepoint. One very expensive lesson in the idea that survival is the first layer of value.
So here's the framework I've used since, and it applies to tankers as much as to rollups:
A chokepoint is anything that, once it becomes risky, forces everyone else to pay to route around it. The real measure of decentralization isn't the number of nodes. It's how cheaply you can leave.
Iraq can't leave cheaply. That's why the discount is twenty dollars and not two. Give Iraq a pipeline, or a ship-to-ship capability it should have built a decade ago, or revenue that isn't 90% oil โ and the discount becomes a rounding error. Instead it's a quarter of the barrel.
Compare that to a well-built protocol. If one RPC endpoint misbehaves, a good wallet swaps to another in a single config change. If one bridge turns risky, liquidity moves. The cost of exit is low, so the chokepoint has no pricing power. The moment exit gets expensive โ proprietary APIs, locked liquidity, a sequencer you can't bypass โ you've rebuilt Hormuz inside your own stack.
And notice who pays. In the Hormuz story, it's Iraq โ a majority-Shia nation, aligned with Iran on paper, and still one of the biggest losers in Iran's own coercive play. Alignment doesn't protect you from a chokepoint. Only an alternative route does.
That's a hard truth for a lot of crypto communities. Being aligned with a foundation, a chain, a narrative means nothing if the underlying infrastructure has a single throat to choke. The ideology is the guest list. The architecture is the venue. If the venue has one exit and it's on fire, the guest list is irrelevant.
Here's where I push back on my own side, because the crypto answer to all of this is usually wrong.
The reflex is to say: decentralize everything, problem solved. But the soft blockade teaches the opposite. Iraq's failure wasn't a lack of decentralization. Iraq's failure was that it never built a bypass while the strait was calm. It waited for the crisis to build the alternative, and by then the alternative cost $20 a barrel.
We do this constantly. We ship the centralized version because it's fast, cheap, and it works. We promise to decentralize later. And later arrives the moment the chokepoint gets weaponized โ which is precisely when decentralization is most expensive and least feasible.
So the pragmatism test isn't "is this decentralized?" It's "if this specific thing gets captured, how fast and how cheaply can users leave?" Most of our decentralized L2s fail that test. So does a lot of our infrastructure. So, honestly, does a lot of DeFi โ a liquidity pool is only as decentralized as the oracle it trusts. Chaos isn't a bug; it's the protocol. The honest answer isn't maximalism. It's redundancy, paid for in advance, during the calm, at a discount.
Iraq paid $20 a barrel to route around a chokepoint it should have bypassed a decade ago. Web3 is making the same bet, on the same schedule, inside our own stack.
We didn't dodge the chaos; we danced through it. But the smart ones build a second door before the fire โ not after. Survival is the first layer of value, and it is always cheaper to build it while the strait is still calm.