February 2022. In a quiet room in Central Europe, a man sits before a screen and begins to count. His name is Jakub Janovsky, and he is one of the analysts behind Oryx—an open-source intelligence project that does something almost no government does publicly: it counts destroyed military vehicles. Not claims, not estimates, but photographs. Every entry must be visually confirmed, cross-referenced, and geolocated where the terrain allows, so that nothing is double-counted or taken on anyone’s word.

Within months, he is looking at thousands of them—burned-out husks in sunflower fields, vehicles between tree lines with snow still packed against their road wheels, the same shapes over and over in mud that changes color with the seasons. And somewhere in that catalog, he begins to notice a pattern that has nothing to do with who won the engagement.

Some of these tanks are burned and intact. Charred, empty inside, finished—but clearly recognizable as tanks. Others are not. Others are two separate objects, a hull and a turret lying twenty or thirty meters apart in the grass, flipped upside down like something a child threw.

He tries the obvious explanations. Bigger warheads, better weapons, luckier hits. None of them hold. Because the split does not follow the weapon that struck the tank. It follows the factory that built it.

Pause on that sentence for a moment. This is not a story about which tank is better. This is a story about the value each army decided to place on its crew. The ammunition is the argument.

Part One. To understand the Russian tank in this argument, you have to stop thinking about tanks for a moment and start thinking about railways. The Soviet Union was a continental power with a continental problem. Its army had to move west across thousands of kilometers of its own territory before it could fight anyone. And it had to move by rail. That meant every tank had to fit inside a fixed loading gauge, cross bridges rated by a peacetime ministry, and be light enough that a division could get somewhere with all its vehicles still working. Weight was not a design preference. Weight was a hard ceiling imposed by the transport network.

So Soviet designers spent forty years solving the same equation. How do you put the thickest possible armor and the biggest possible gun into the smallest possible box? The answer they arrived at was to remove a man. Take out the human loader and you take out a man-sized volume from the turret. A smaller turret has less surface area to armor. Less surface area means less weight, which means you can make what remains thicker. It also means a lower profile, which is harder to spot and hit from two thousand meters in bad light.

Then there are the calculations nobody puts on a recruiting poster. A crew of three instead of four means twenty-five percent fewer trained crewmen per tank across a fleet of thousands. At army scale, that is not a rounding error. That is an entire training institution you do not have to build.

The machine that replaced the human loader was a carousel autoloader. A horizontal ring of ready rounds laid flat in the hull floor directly beneath the turret and beneath the crew. That placement was not random. It was a calculated bet. The front of the hull holds the thickest armor on the vehicle. During an advance or from a hull-down firing position, the lower hull is hidden by the terrain itself. And in frontal engagements between formations, the statistics said the majority of hits concentrate on the turret—the part that sticks up above the terrain line. By putting the ammunition low and forward, you put it behind the strongest armor and below ground level.

They were not wrong about the railways. They were not wrong about the low profile. They were not wrong about the trained-personnel math. They were not even wrong in 1975 about the expected source of hits.

The T-90M Proryv is the final shape of that logic. Unveiled in 2017 and in service by 2020, it replaced the old cast turret with a welded one, added Relikt explosive reactive armor to the front and sides, and mounted a 125 mm smoothbore 2A46M-5. It has modern thermal sights and a digital fire-control system. Through the same barrel, it can fire 9M119-series guided missiles beyond five thousand meters—farther than most tanks can accurately hit with conventional rounds. It is fast, low, and genuinely accurate. It is also cheap by the standards of the tanks it was built to fight. Russian figures put the T-90M at about 4.5 million dollars per unit. The Challenger 2, by current comparative estimates, costs around 8.5 million. Two tanks for one from the Nizhny Tagil factory that never stopped producing.

In June 2023, Vladimir Putin described it in four words as the best tank in the world. That is a government claim carried by a state news agency, and it should be treated as such. But it was not an absurd claim. On paper, measured against the Russian armored doctrine that had been optimizing since 1945, it was close to the truth.

A U.S. officer later summarized that entire design philosophy in six words. I will tell you the six words later, because they only land properly after you have seen what happened.

There were only about 67 T-90Ms in actual service when the full-scale invasion began in February 2022. The model was logical, the tank was good, and the factory was ready to make more. The model worked—until it stopped working.

Part Two. This is the machine that broke it, and the reason is not what anyone expected.

In January 2023, the United Kingdom pledged fourteen Challenger 2 main battle tanks to Ukraine. It was the first Western commitment of a modern main battle tank in this war, and it broke a political barrier after which Leopards and Abrams flowed. The tanks arrived in the country by late March. On 28 March 2023, Ukrainian Defense Minister Oleksii Reznikov confirmed their arrival and said a Rolls-Royce driver would not feel the same level of comfort as the crew of this tank. On 11 May, it received official backing in London when the British Defense Secretary told the House of Commons that all fourteen tanks had been delivered. The tanks went to the 82nd Air Assault Brigade.

Challenger 2’s armor is a composite called Dorchester, a second-generation development of Chobham, and its exact composition is classified and will remain so. What was never classified was its record. Over more than two decades of service, in Bosnia, Kosovo, and Iraq, no Challenger 2 was destroyed by an enemy. The only catastrophic loss, in March 2003, was caused by another Challenger 2. In 2006, near Al Amarah, one was hit by an RPG-29, a weapon designed specifically to defeat that kind of armor. It was damaged. But it returned to base. And it was repaired.

Its gun is the strangest thing about it. The L30A1 120 mm rifled gun makes it the only rifled main gun on any NATO main battle tank. The rest of the world switched to smoothbore decades ago. The British kept the rifled barrel for one reason: HESH rounds, high-explosive squash head, which compress against the target before detonating and need the spin that only a rifled barrel provides. Range reaches about eight kilometers. Against armor, the tank fires L27A1, a depleted-uranium dart known as CHARM 3.

But the detail that settles this argument is not the gun. It is the ammunition that goes into it. Challenger 2 uses two-part ammunition. The projectile is one body. The propellant charge is a separate body inside a bag. They are stored apart from each other, and the charges are placed in armored bins below the turret ring, surrounded by liquid. There is no carousel autoloader. There is no ring of complete rounds under the crew’s feet. And in that turret, there is a fourth man, a human loader, whose entire job is to bring those two things together—and who can inspect the charge and decide not to take it out.

Nobody at Vickers designed this arrangement because it was elegant. It is heavier. It is slower. It costs you a man and costs you interior space. It exists because British armored doctrine had made a decision about priorities, and the decision was that the tank is the cheap part.

Then Ukraine tested it. In early September 2023, near the village of Robotyne in Zaporizhzhia, a Challenger 2 was lost in combat for the first time in the history of the type. The account given by Ukrainian and British officials was that it struck a mine, which immobilized it. The crew dismounted. Then a loitering munition hit the stationary vehicle, and it burned. There were four men in that tank. All four men walked away from it. And when you examine the footage, the thing that is missing matters as much as the thing that is present. The turret was still in place on top of the hull. Twenty tons of steel that had every reason to be somewhere else was still exactly where it belonged. Whatever burned inside that tank, it burned. It did not explode.

On 2 September 2023, a tanker from the 82nd Brigade told the Ukrainian General Staff that its fire accuracy was comparable to a sniper rifle. That is not a technical specification. That is a man describing what he can hit.

On the other side, the reaction shifted quickly from contempt to something else. Russian military channels, including Ribar, began announcing the destruction of Challengers as a recurring news item. By 2025, the claimed total had reached four tanks. The visually confirmed total never exceeded three. Two destroyed and one damaged out of fourteen delivered. When a claim is made with that much repetition and supported by that little evidence, the repetition tells you about the anxiety, not the tank.

What that brigade did with fourteen tanks near Robotyne and later inside Kursk changed how many armies now think about survivability. But before we get to why, if this deep dive into British engineering is interesting to you, press the subscribe button. It costs nothing and helps the channel grow.

And now, the record.

Part Three. To understand why the British tank was built this way, you have to go back not to the Cold War but to Normandy. British armored regiments fighting across France in 1944 learned a specific and expensive lesson, and they learned it inside a machine famous for burning. A tank can be replaced in weeks. A tank crew cannot. A commander who has fought for a year, a loader who knows his gun, a driver who reads terrain—these are men who took years to make, and if they survive losing a tank, they climb into another on Tuesday and are immediately useful.

An army that loses tanks but keeps its crews rebuilds itself. An army that loses both starts from zero. That insight became institutional, and by the 1970s it took geometric form. British and American designers reached the same principle from different directions: put the ammunition away from the crew, and give the fire somewhere to go. In the Abrams and Leopard 2, the solution was a turret bustle behind an armored door with blowout panels in the roof. In Challenger 2, the solution was different in execution and identical in purpose. Separate the charges, put them in protected bins, and accept the weight penalty of both.

The Russian solution trusted the armor and stored the rounds in the floor. Ukraine has now settled the argument between these two solutions at a scale no proving ground could afford.

The number is this: 159. 159 T-90M Proryv tanks visually documented by Oryx as destroyed, damaged, abandoned, or captured. Not claims—photographed. And because each one requires a picture, this is not an estimate of losses. It is the actual floor of them. Read that number again, then compare it to the fleet it came from. There were 67 in service the day the invasion began. A factory that open-source analysts, Janovsky among them, assess can produce no more than 200 new T-90Ms a year under sanctions. So Russia has lost, in wreckage alone, nearly a full year of its maximum national production of its newest tank.

The aggregate is worse. By January 1, 2026, Oryx had confirmed 4,308 Russian tanks of all types lost, 3,222 of them destroyed outright. That is more tanks than the Russian army had in active service on the day it invaded. And the thing causing most of it is not a tank at all. A T-90M costing about 4.5 million dollars has been repeatedly killed by an FPV drone assembled from commercial parts for a few hundred dollars. That is the exchange rate governing armored war now, and no amount of frontal protection answers it, because the drone does not come from the front. It comes from above, where there is no reactive armor, targeting the roof, the engine deck, and the thin plate over the carousel.

So why couldn’t the losing side simply copy the solution? It was never about the armor. It was never about the gun. It was about the floor. Moving ammunition out of the hull is not a modification. It is a different tank. A rear bustle big enough for twenty rounds needs a bigger turret. A bigger turret needs more armor to cover it, adding weight, pushing you past the bridge and rail limits the entire hull was designed to meet. Venting a fire upward means redesigning the roof. Loading from a rear bustle means a fourth crewman, which means a fourth seat, a larger fighting compartment, more weight, and twenty-five percent more tank crews to find, train, and pay across the entire force. The carousel is not a component of the T-90M. It is the foundational assumption everything else is built on. That is what makes this a story about doctrine rather than procurement. The ammunition is the argument. And by the time the argument was settled, one side had already built four thousand tanks around the wrong answer.

If you served in the British Armoured Corps, or if you have been following the equipment side of this war closely, I would genuinely like to read what you know in the comments. The people who have actually sat inside these turrets consistently know things that data sheets do not print.

Part Four. Here is what the other side did once the pattern became undeniable. And the most revealing response was not a statement. It was a weld. Long before any institution admitted anything, Russian crews began fixing their own tanks in the field. Metal cages over the turret roof, cut from whatever steel was available, welded without authorization, mocked by the press as cope cages. Then rubber mats. Then by 2024, full metal shells encasing entire vehicles—so-called turtle tanks, crude and half-blind, but genuinely effective against drones during the ninety seconds that matter. Nobody in Moscow designed those. Anonymous crews at the front did, because they understood what was underneath them and wanted the drone to explode somewhere else.

Official acknowledgment came through steel too, which is easier to deny. Newer T-90Ms feature a small storage compartment at the rear of the turret with venting panels. That is the Western principle applied precisely to reserve ammunition. The autoloader stayed where it was, which tells you the engineers understood the problem completely but could only solve the part that did not require an entirely new vehicle. Some tanks also entered service without the Shtora defensive system, an admission that the answer to a laser-guided missile is not an effective answer to a man flying a camera drone.

None of it changed the outcome on the battlefield. In mid-January 2024 near Stepove outside Avdiivka, a T-90M met two M2 Bradley infantry fighting vehicles of Ukraine’s 47th Mechanized Brigade. A Bradley is not a tank and cannot penetrate one. What it can do is put its 25 mm cannon precisely into the turret. That is what it did, methodically, until the Russian tank was blind. A drone finished the job. Ukrainian Ground Forces described it simply afterward: most of the work was done by the two Bradleys. American technology won. Add the intelligence loss. In September 2022, near Kharkiv, Ukrainian forces captured an intact, abandoned T-90M. The newest Russian serial-production tank, with its armor array and fire-control system, placed in Western analysts’ hands without a single shot fired at it.

By 2025 and beyond, reports and open-source analysis indicate the type has been largely withdrawn from frontline assault missions and is being preserved. I would note that this is a report rather than a confirmation, because the reasons given are analysts’ inferences. But the observable fact is not in dispute. The tank that was supposed to lead the breakthrough—whose name literally means breakthrough—is being kept away from where the breakthroughs happen.

And now, the six words. In U.S. reporting on why Russian turrets keep flying through fields, Major Mike Liescano Jr. of the U.S. Army summarized sixty years of design divergence in one sentence. The Russians prioritize lethality over survivability. They were right that a smaller tank is harder to hit. They were right that an autoloader never tires. They were right that mass and numbers have a quality of their own. They were right that the front is the strongest part of the vehicle. But they were wrong about the direction. All of those judgments assumed the threat comes horizontally, from another tank across open ground at the gunner’s line of sight. It did not. It came from above, cheaply and in numbers, flown by a man in a treeline with a screen strapped to his chest.

There was British integrity in this period too, and it deserves mention. On 6 September 2023, Defense Secretary Grant Shapps stood in the House of Commons and confirmed the loss of a Challenger. He said it may be the first loss as far as we know. He did not dispute the footage, did not offer excuses against it, and made clear Britain would not send a replacement. That is the other half of an honest account. The tank was not invincible. The government said so within forty-eight hours, and fourteen became thirteen.

Part Five. So back to the man counting wreckage. What was that pattern actually recording in Jakub Janovsky’s ledger? Not accuracy, not courage—both sides supplied those in quantities neither deserves to be lectured about. What those photographs were recording was a decision made in a design office decades before the war about where to put the most dangerous thing inside a tank, and therefore about what happens in the four seconds after the armor is penetrated. Their metric was rounds fired per ton of steel. The British metric was men walking out of the hatch.

And now the honest half, because a verdict that condemns only one side is not a verdict. Fourteen Challenger 2s are not a fleet. They are a company, and they never carried real weight in a war of thousands of vehicles. Two were destroyed and one was damaged. The rifled gun Ukrainian gunners admired fires ammunition that fits nothing else in the entire Ukrainian arsenal. So every round had to arrive on its own supply line from Britain. That barrel erodes several times faster than a smoothbore, a real problem for a tank used as long-range self-propelled artillery. At sixty-four tons and up, with a 1,200-horsepower engine, it bogged down in Ukrainian ground that lighter tanks crossed. David Axe argued in Forbes, plainly, that it was the wrong tank for Ukraine. On the metrics of quantity, logistics, and mobility, he had a case.

And the two never fought each other. There is no confirmed engagement between a Challenger 2 and a T-90M anywhere in this war, and anyone who tells you otherwise is selling a fantasy. Mines, artillery, missiles, and drones destroyed far more of both than either destroyed of the other. The word against in this story is not a battle. It is an audit, and the war was simply the auditor.

But on the one question this entire comparison was designed to ask, the answer is not ambiguous. Challenger 2 lost tanks but kept its crews. T-90Ms were lost with their crews inside them, most of the time. The British machine was fewer in number, badly supplied, wrong for the ground, too heavy for the mud—and it brought its men home. The Russian machine was modern, accurate, well-armored against the threats it was designed for, and built on an assumption that stopped being true at a treeline in Ukraine in 2022.

So the tribute here is not to a defeated enemy, because that enemy is still fighting. It is to the engineers at Vickers and Chobham who accepted the weight, the cost, and the complexity in exchange for a fourth man and a separate magazine, and who did not live to see the argument settled. It is to the Ukrainian mechanics keeping thirteen orphaned British tanks running through a supply chain stretching to another continent. It is to the crews of the 82nd Air Assault Brigade, who took a machine nobody expected to see in this war, used it carefully, lost some of it, and brought it back. The ammunition is the argument. Every army designing a tank now has read that verdict, and what they are building because of it is the next chapter.

If this was worth your time, a like helps the story of these fourteen tanks reach more people who will never read a defense procurement sheet. It costs nothing. But it means more than it should.

So here is the question I’ll leave you with: if you had to choose today—knowing what the treelines of Ukraine have taught us—would you rather ride into battle in a machine that never tires, or one that makes sure you walk out of the hatch? Tell me in the comments.