Daniel's been reading his own pantry labels again. He wants to know how a tanker that turns its transponder off in the middle of the ocean still gets caught, what the cargo actually is, and whether anyone's out there heaving shipping containers across a gap in the dark.
The container thing is the part everyone pictures, and it's wrong. The real operation is a hose.
A hose.
Two tankers lashed eleven feet apart, black rubber fenders the size of houses between them, and a hose the width of a man's torso pumping crude across. Sometimes more than a million barrels at a go. Takes a day, sometimes several. No cranes, no containers, no cinematic heist. Just a very long pipe and a lot of patience.
So Daniel's mental image of a stolen iPhone in a duffel bag thrown across a bargeway is off by about a million barrels.
Off by roughly the entire economy of a small country. And the cargo isn't stolen phones. It's Iranian crude heading to China, Russian crude heading to India. The incentive is not subtle. A single very large crude carrier evading the sixty dollar price cap when Brent is at seventy eight adds about fourteen million dollars to the sale price. When Brent spiked to ninety four, the incentive was twenty eight million per cargo. Per cargo.
That's not "worth the hassle" money. That's "buy a second ship and paint over the name" money.
Which they do. And the painting is sometimes bad. There's a Wall Street Journal report from May where reporters in a boat off Malaysia watched a sanctioned tanker called the Catalina Seven transfer oil to a ship with its name painted over. But the paint job was incomplete, and a reporter read the old name through binoculars.
From a boat. Not from orbit.
Right, and that's the thing Daniel's actually asking. He's got the right instinct. Satellites don't loiter. A commercial imaging satellite passes over a spot, takes a picture, and it's gone. The window for a full ship-to-ship transfer is six to twenty four hours. The odds that a high-resolution satellite happens to be overhead at the exact right moment are terrible.
So the satellite isn't the detective. The satellite is the evidence locker.
The detective is the transponder gap. Every tanker over three hundred gross tons has to broadcast its position continuously under the Safety of Life at Sea convention. When a shadow fleet tanker goes dark, it doesn't just vanish. It disappears for a very specific window, usually somewhere between six and seventy two hours, and the gap appears right around a known anchorage. Then the ship reappears, broadcasting again, and the voyage looks clean except for that hole in the middle.
So the workflow is: pull ninety days of position data, flag every gap longer than six hours near a known transfer zone, and then look for a second tanker that also went dark in the same five nautical mile box within the same forty eight hour window. Two ships, both silent, same place, same time.
And they're holding station within about five hundred meters of each other. That's the paired blackout. That's the tell. Then you check the ownership chains, the registry, the shell companies. Then you task a satellite to look at the coordinates.
So imagery confirms what the radio silence already told you.
And it can confirm it at surprisingly coarse resolution. A three hundred thirty meter supertanker shows up in free ten meter Sentinel data as about thirty three pixels. That's enough to say a ship is there, how long it is, which way it's pointing. But to actually see that two hulls are rafted together side by side, you need sub-meter imagery. Half a meter to seventy five centimeters, and you can see the deck layout, the two hulls lashed together. At thirty to fifty centimeters, you can see the deck pipework, the manifolds, whether the hoses are connected.
But you can't read the name.
You cannot read a ship's name from orbit at any commercial resolution. Nobody can. What you do is measure the length and the beam, match the deck fittings, and track the same hull between passes. It's fingerprinting, not name reading. A tanker is a three hundred thirty meter rectangle sitting on a flat black ocean. It's one of the easiest things on Earth to spot from space. The hard part is proving which rectangle it is.
So Daniel's premise that analysts are reading individual ship markings from commercial imagery is exactly backwards. They're identifying ships the way you'd identify a person from a grainy security camera. Build, gait, the way they move. Not the name tag.
And the identification is now automated. There's a model from University College London, a YOLO variant, trained on PlanetScope imagery at three meters per pixel. It detects four vessel types and two types of ship-to-ship transfer. Ninety seven percent F one score. It found four hundred and two dark transfers in the Kerch Strait between twenty twenty one and September twenty twenty three.
Four hundred and two events that never appeared in any transponder log.
And that's just the Kerch Strait. Total transfers detected there since twenty twenty one, over twelve thousand, more than half of them since the invasion of Ukraine. The model is fifty two megabytes. It can search a hundred and fifteen square kilometers per second. This is not a supercomputer problem anymore. It's a laptop problem.
So the answer to "how are these actually spotted" is: a gap in the radio record, a query on a database, a satellite tasked to look at a specific coordinate, and increasingly a neural net that just watches the ocean all day. The satellite was never the thing doing the spotting.
And the thing being spotted is oil. Almost exclusively oil. Liquid bulk. Crude, petroleum products, sometimes liquefied petroleum gas. The shadow fleet is about fifteen hundred ships now. The main routes are Iran to China at about one point four million barrels a day, Russia to China at about a million, Russia to India at just under a million, Venezuela to China at about four hundred thousand.
Those are not smuggling volumes. Those are pipeline volumes.
Iran pulled in about thirty one billion dollars in oil revenue from China last year. That's roughly ninety percent of its oil exports and about forty five percent of its government budget. This is not a black market in the sense of some guys in a speedboat. This is the primary export channel of a national economy.
Which answers Daniel's second question. The incentive is that the alternative is not selling the oil at all. When the United States and the European Union sanction your crude, you can't just put it on a normal tanker with a normal manifest and sell it on the open market. So you sell it to a buyer who doesn't recognize the sanctions, and you use a ship that doesn't care.
China's position is explicit. In May, Beijing invoked its blocking rule for the first time, ordering domestic firms not to comply with U.S. sanctions on five Chinese refineries. The Chinese crew member in that Wall Street Journal piece said he saw nothing wrong with the transfers because neither he nor China's government recognizes the validity of U.S. sanctions. From their side, it's not smuggling. It's commerce.
From the Malaysian side, it's a fine.
That's the enforcement problem in one sentence. The tanker Nora was detained in January for an unauthorized transfer. The agent was fined about thirty three thousand dollars. The ship was released. Within weeks it was back doing transfers. In February, Malaysia seized a hundred and thirty million dollars of crude and detained two tankers, arrested both captains. That made headlines. It's also the exception.
Thirty three thousand dollars against a fourteen million dollar margin. That's not a deterrent. That's a rounding error.
And the ship itself doesn't care. It's owned by a shell company registered in a jurisdiction that exists to be a shell company. You fine the agent. The agent disappears. You seize the ship. A new shell company buys a new ship. The shadow fleet is fifteen hundred vessels because every seizure just creates a vacancy.
Now, Daniel's third question was about scale and equipment. He guessed twenty or forty foot containers being hoisted across. You said that's wrong for oil. But there is a crane version.
Dry bulk. Grain. The Wall Street Journal did a piece on stolen Ukrainian grain back in twenty twenty two, and the mechanics are different. You have a feeder ship loading grain in a port where nobody's watching too closely. It sails out to a larger bulk carrier waiting at sea. Then a third vessel arrives. A crane ship.
A dedicated crane vessel.
The example they documented was the Russian flagged feeder M. Andreev coming alongside the Panama flagged bulk carrier Emmakris Two near the Kerch Strait. The crane ship Petra Two joined them, and the three ships sat nestled together for more than seven hours while the grain was loaded. The crane ship is a legitimate ship type. It's not a modification to the cargo ship. There's no equipment to explain.
So if customs boards the bulk carrier, the ship looks completely normal. No cranes on deck, no unexplained gear. The crane was on a different ship that's already gone.
And the grain itself has been mixed with Russian grain to disguise its origin. The quote from the analyst was "It's wheat laundering." You blend stolen Ukrainian wheat with legitimate Russian wheat, and by the time it's in the hold, the origin is unrecoverable.
Wheat laundering. So the container question has a real answer: oil moves by hose, grain moves by crane ship, and standard ISO containers being hoisted across at sea is not actually a documented sanctions evasion method. It's too slow, too visible, and too easy to trace.
A container has a number on the side. You can scan it, track it, match it to a manifest. Bulk liquid has no serial number. Once it's in the receiving tanker, it's just crude oil. You relabel it as Malaysian blend or Kazakh CPC or just Middle East crude, and the bill of lading says whatever you want it to say.
That's the actual mechanism Daniel was asking about. The relabeling. The receiving vessel adds the cargo to its manifest stating it came from somewhere else. That's not a lie anyone can easily check, because the cargo is a liquid with no provenance.
And the scale is enormous. The MT Tifani was carrying one point nine million barrels of Iranian crude when it was seized. That's not a shipment. That's a strategic reserve.
So let me make sure I've got the whole picture. The detection is a layered workflow. Transponder gaps first, then ownership analysis, then satellite confirmation. The cargo is sanctioned oil, mostly Iranian and Russian, moving at pipeline scale to China and India. The transfer is a hose between two tankers, or a crane ship for grain. And the incentive is tens of millions of dollars per voyage against fines that are a rounding error.
And the satellites are getting better at the confirmation step. There are commercial providers now doing image analytics that resolve every detection in a frame to a vessel identity. One of them has filtered more than two point two million false ship-to-ship meeting records since late February. The problem isn't seeing the ships. The problem is knowing which of the thousands of meetings are real transfers and which are just two tankers passing close.
Two point two million false positives. That's the scale of the haystack.
And the needle is a ship that went dark for eight hours, reappeared with a different name painted on the stern, and is now carrying a million barrels of oil that officially came from a country that doesn't export oil.
There's one more angle I want to sit with. The United States is now using this same technique.
When the Strait of Hormuz closed in late February, the U.S. military had a problem. It needed to keep Gulf oil moving out, but the normal tanker routes were not viable. So the U.S. oversaw what Reuters described as scores of secretive ship-to-ship oil transfers, using drones and helicopters to shuttle oil out of the Gulf. The same shuttling technique Iran pioneered.
So the tactic is now legitimate when the United States does it. Which is not a criticism. It's an observation about how the line between smuggling and logistics depends entirely on who's writing the manifest.
The technique is just a technique. It's a hose and a pump. The legality is in the paperwork. And the paperwork is whatever the buyer and seller agree it is.
Which is why the enforcement is so hard. You can't ban the hose. You can't ban the pump. You can only ban the specific cargo, and the cargo is unlabeled.
And the ships are getting better at hiding. During the Hormuz crisis, on May fifth, there were a hundred and sixty seven commercial vessels in the strait. A hundred and forty six of them were running dark. Nearly four hundred seventy vessels hit by GPS jamming in a single twenty four hour window. The jamming is a separate layer. It doesn't hide the ship. It just makes the ship's own position reports unreliable.
So the transponder says one thing, the GPS says another, and the satellite sees a third. And somewhere in the middle is the truth.
And the truth is that dark doesn't mean invisible. A very large crude carrier is three hundred thirty meters long and sixty meters wide, sitting on a flat, dark, essentially featureless background. From a satellite's point of view, there are few things on Earth easier to spot. The darkness is about denying the paper trail, not the visual.
So the whole operation is a paperwork problem with a plumbing solution.
That's the cleanest summary I've heard. The physical act is trivial. Two ships, a hose, a pump. The hard part is making the cargo's history disappear. And that's done with registries and shell companies and bills of lading, not with stealth technology.
Daniel asked how the vessels explain having the equipment if they're caught. The answer is they don't have to explain anything. The equipment is standard. Every tanker has manifolds and pumps for loading and unloading. The only question is whether the ship it's connected to is on the manifest.
And if you're caught mid-transfer, the penalty depends entirely on where you are. In Malaysia, it's a fine and maybe an arrest. In the European Union, the fourteenth sanctions package explicitly bans ship-to-ship transfers of Russian crude in EU waters. But the transfers don't happen in EU waters. They happen in the exclusive economic zone of Malaysia, or off the coast of Oman, or in the middle of the East China Sea.
The middle of the East China Sea is where North Korea gets its oil. The UN capped North Korean crude imports at five hundred twenty five thousand tons a year and petroleum products at five hundred thousand barrels a year. The UN Panel of Experts documented a Dominica flagged tanker called the Yuk Tung transferring petroleum to the North Korean tanker Rye Song Gang One in the East China Sea. That was back in twenty nineteen.
North Korea is still doing it. The caps are still in place. The transfers still happen. The UN documents them. The ships change names and flags. The pattern is identical.
The same mechanism serves a pariah state with a nuclear program and a supermarket shelf in Jerusalem. The difference is scale and cargo. North Korea needs oil to keep the lights on. Iran needs oil revenue to keep the government running. And the Tunisian olive oil on Hannah's shelf just needed a label change in a warehouse.
The olive oil probably never went through a ship-to-ship transfer. That was a paperwork relabeling. The ship-to-ship transfer is for when the cargo is too big, too valuable, or too radioactive politically to touch a port.
The port is the thing. A sanctioned tanker can't call at a normal port. The port authority would flag it, the customs would impound it, the insurer would drop it. So the transfer happens at sea, outside territorial waters, where the only authority is the flag state of the ship itself.
The flag state is often a country that has no interest in enforcing anyone's sanctions. The ship flies the flag of Cameroon or Palau or the Cook Islands. The registry is a post office box. The owner is a shell company in the Marshall Islands. The insurer is a front in Dubai. By the time anyone traces the chain, the ship has changed hands three times.
When Daniel asks how they get away with it, the answer is that the entire system is designed to make the question unaskable. There's no one to ask. There's no jurisdiction that wants to answer.
The buyers are not hiding. The teapot refineries in Shandong and Liaoning are private Chinese companies that use Chinese banks and never touch the U.S. financial system. They're beyond the reach of U.S. sanctions by design. The oil arrives, they refine it, they sell the products domestically. The dollar never enters the transaction.
The sanctions are a tax on the transaction, not a prohibition. The oil still flows. It just flows through a more expensive, more complicated, more opaque channel. And the cost of that opacity is borne by the shadow fleet and the intermediaries.
Which is why the shadow fleet keeps growing. Fifteen hundred ships. Four hundred tankers sanctioned by the U.S., the EU, or the UK for Iran-related activity alone. And the transfers in the Eastern OPEC Plus League, the EOPL, more than doubled from two hundred eighty in twenty twenty three to six hundred seventy nine in twenty twenty five.
The EOPL. That's the stretch of water off Malaysia where all this happens.
The Eastern OPEC Plus League. It's a critical logistics node. China cannot get its oil from Iran without going through it. One analyst called it a little epicenter of maritime lawlessness. A smorgasbord of deceptive shipping tactics.
A smorgasbord. That's the polite term for a floating black market.
The U.S. response has been to sanction the ships, the captains, the agents, the insurers. But the ships keep sailing. The quote from the former naval officer was "Sanctions alone don't stop the ships."
They just change the name on the stern.
Sometimes not even that. The paint job on the Nora was so bad you could read the old name through binoculars.
Which is either incompetence or confidence. If you know the penalty is a thirty three thousand dollar fine, why bother with a good paint job?
The calculus is exactly that. The expected value of getting caught is a fine that's a rounding error on the profit. So the rational move is to spend as little as possible on evasion and just pay the occasional fine.
That's not a black market. That's a toll road.
The toll is set by the Malaysian Maritime Enforcement Agency, which seized a hundred and thirty million dollars of crude in February. That got attention. But the same agency had fined the Nora's agent thirty three thousand dollars a month earlier. The enforcement is real but inconsistent.
The answer to Daniel's question about how they explain the equipment is: they don't. The equipment is standard. The ship is standard. The transfer is standard. The only thing that's not standard is the paperwork, and the paperwork is handled by people who never set foot on a ship.
The scale question. Daniel guessed twenty or forty foot containers. The reality is a million barrels of oil pumped through a hose over a day or two. The hose itself is a piece of equipment that costs more than most houses. It's not a garden hose. It's a floating pipeline with its own flotation devices and quick connect couplings.
For grain, it's a crane ship that can lift bulk cargo from one hold to another. The crane ship is the key. It's a third vessel that makes the whole operation look like a legitimate lightering operation.
Lightering is the legitimate version of this. A big tanker can't get into a shallow port, so it transfers cargo to a smaller tanker offshore. That's been standard practice for decades. The illegal version just does it in the dark, with the transponder off, and lies about where the oil came from.
The line between legitimate and illegitimate is literally just the transponder and the manifest.
The transponder is the thing that gets them caught. The gap is the tell. A ship that goes dark for eight hours near a known transfer zone and comes back with a different cargo level is not subtle. The analysts have been doing this long enough that the pattern recognition is almost automatic.
Now the pattern recognition is literally automatic. The YOLO model is doing it in real time. Fifty two megabytes. A hundred and fifteen square kilometers per second. That's not a research project. That's a product.
The next step is integrating the satellite detections with the transponder data in real time. Right now, the workflow is still mostly manual. An analyst pulls the AIS data, flags the gaps, tasks the satellite, waits for the image, and then writes up the finding. The model can do the image part in milliseconds. The bottleneck is the human in the loop.
Which is where the whole thing is heading. The same way Daniel's automation work is about removing the human from the loop, the OSINT community is automating the detection of sanctions evasion. The ships can't hide from a model that watches every pixel of every pass.
The ships are not getting smaller. A very large crude carrier is the size of a small city block. It cannot hide. It can only lie. And the lie is getting easier to catch.
To answer Daniel's three questions in order. One, the imagery is good enough to confirm but not to name. The detection starts with the transponder gap, not the satellite. Two, the incentive is tens of millions of dollars per voyage for sanctioned oil, mostly Iranian and Russian, moving at pipeline scale. Three, the scale is a million barrels through a hose, not containers across a gap. And the equipment is standard, so there's nothing to explain.
The one correction I'd make to Daniel's framing is that the receiving vessel doesn't relabel the cargo after the transfer. The relabeling happens in the paperwork before the transfer even starts. The bill of lading is written to describe the cargo as something it never was, and the transfer just makes the lie physically true.
The lie becomes the cargo.
The lie becomes the cargo. That's the whole operation in five words.
Hilbert: The hose is forty six centimeters inside diameter. Not the width of a man's torso. Forty six centimeters. I know because I spent a summer on a lightering crew in the Gulf of Mexico, and the hose we used was forty six centimeters. It takes two men to handle the coupling, and if it lets go while it's pumping, you don't want to be standing anywhere near it.
That's a very specific correction.
Hilbert: The width of a man's torso would be a small hose. The big ones are wider than a dinner plate. And they don't get thrown across. They get lifted by a crane on the tanker itself, with a line attached to the flange. The receiving ship hauls it over, and then the two crews bolt the flanges together. It's not complicated. It's just heavy.
The tanker has its own crane for the hose. That's the equipment Daniel was asking about. It's standard. Every tanker has one.
Hilbert: Every tanker has a small crane on the deck, usually amidships. It's for the hoses, the stores, the spare parts. Nobody looks twice at it. If you board a tanker and see a deck crane, that's not evidence of anything. It's just a deck crane.
The fenders. The black rubber things between the ships.
Hilbert: Yokohama fenders. They're about the size of a small car, and they're filled with air. When two tankers come alongside, you hang them over the side so the hulls don't grind against each other. They squeak. The whole time. It's a terrible noise.
You spent a summer doing this.
Hilbert: Eleven years. I was on a lightering crew for eleven years. The Gulf of Mexico, the Caribbean, a few runs off West Africa. I've done more ship-to-ship transfers than I can count. All of them legal, all of them in daylight, all of them with the transponder on. The only difference between what I did and what the shadow fleet does is the paperwork and the radio.
The technique is just a technique.
Hilbert: The technique is just two ships and a hose and a pump. The first time you do it, you think it's going to be dramatic. It's not. It's mostly waiting. You wait for the ships to get alongside, you wait for the hoses to be connected, you wait for the pumps to start, you wait for the pumps to stop. The whole thing is waiting.
The cargo. You were moving crude.
Hilbert: Crude, fuel oil, sometimes condensate. The condensate is the worst. It's lighter than crude, so it sprays. You get it on your boots and it never comes off. I still have a pair of boots in a box somewhere with condensate stains.
The scale. A million barrels.
Hilbert: A million barrels takes about thirty six hours with a good pump. The pump on a very large crude carrier moves about ten thousand cubic meters an hour. You do the math. It's not fast. It's just continuous. The hose never stops.
When the shadow fleet does it in six hours, they're either moving less cargo or they're running multiple hoses.
Hilbert: They're moving less. The six hour transfers are usually Aframax to Aframax, maybe seven hundred thousand barrels. The big ones take a day or more. The six hour window is for the smaller ships.
The darkness. You did it in daylight.
Hilbert: Daylight, floodlights, the whole thing lit up like a stadium. Because if something goes wrong at night, you can't see the hose, you can't see the fenders, you can't see the gap between the ships. The shadow fleet does it in the dark with the lights off. That's the part that would scare me.
The risk of a spill.
Hilbert: The risk of a spill, the risk of a fire, the risk of a hose letting go and whipping across the deck. A forty six centimeter hose under pressure is a lethal weapon. If it lets go, it doesn't just spray. It flails. I saw a coupling let go once on a fuel oil transfer. The hose took out a section of railing. Nobody was standing there. If they had been, they wouldn't have walked away.
The shadow fleet is doing this in the dark, with the lights off, and the transponder off. The risk profile is not the same as a legal lightering operation.
Hilbert: The risk profile is the same. The risk tolerance is different. They're getting paid more than I ever was, and they don't have an inspector looking over their shoulder. If something goes wrong, they just cut the hose and leave. The oil floats. The ship sails. Nobody files a report.
The oil floats. That's the environmental cost that doesn't show up in any of the sanctions analysis.
Hilbert: The oil floats and it washes up somewhere. Or it doesn't, and it just spreads out and kills everything under it. I've seen the sheen from a small spill. A big one from a broken hose in the dark would be a disaster. But nobody's measuring that.
Because the whole point is that nobody's measuring anything. That's the entire business model.
Hilbert: The business model is that the oil gets there and the paperwork says it came from somewhere else. The rest of it, the risk, the spill, the fines, that's just overhead.
The overhead is priced in. The thirty three thousand dollar fine is just a line item.
Hilbert: The fine is a joke. When I was doing it, the fine for an unauthorized transfer was more than my annual salary. Now it's less than a month's profit on one voyage. The economics have changed. The physics haven't.
The physics is still a hose and a pump.
Hilbert: The physics is still a hose and a pump. And the hose is still forty six centimeters.
The next time Daniel pictures a ship-to-ship transfer, he should picture a very long, very heavy hose, two crews bolting flanges together in the dark, and a pump that runs for thirty six hours. No containers. No cranes. Just plumbing.
A satellite that saw the whole thing, but only because someone already knew where to look.
The open question is whether the automated detection gets fast enough to matter in real time. Right now, the analysis catches up weeks or months later. The ship has already delivered, the money has already moved, the cargo is already refined and sold.
If the model can flag a dark transfer as it's happening, and the enforcement can respond in hours instead of months, the economics change. The question is whether anyone with the authority to act is watching the feed.
That's the thing to watch. The satellites already work. It's the response time that's the bottleneck.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop.
If you enjoyed this, leave us a review wherever you listen. Or email us at show at my weird prompts dot com.
We'll be back soon.
