In the Air Raid Shelter

Things seem bad for Ukraine, but there are good things on the horizon — they just need time.

I woke up a minute before 3am and thought: I guess it won’t be tonight.

Five hours earlier, I had been sitting at the hotel bar with a new friend. They’re a combat medic, just back from the frontlines, visiting the capital for a few days.

“Either tonight or tomorrow night,” they said. “Weather is perfect for it.”

Since I’ve arrived in the capital, the air raid sirens have been few and far between — and few people have taken them seriously. One went off earlier in the day, as I was interviewing a man whom Putin has tried to kill multiple times. Everyone silenced their phones and we continued talking.

But this one is different. Fleets of Tu-95 long-range bombers and Su-35 fighter jets took to the sky just as I awoke. In waves, they fired barrages of sophisticated missiles, capable of banking, reversing, diving, and dancing through the sky. “Do not be over-confident,” a man’s voice said from my phone’s speaker. “Proceed to an air raid shelter.” People took this one seriously.

So about a dozen of us, a mix of Europeans and Ukrainians, including two weary school children, trudged down the stairs and into a cozy shelter under the hotel.

Here we are. Above us, around the country, Ukrainian air defense systems are gearing up for a night of hard work. But they must know how low their own supplies are running. Thanks to nationalist ideologues halfway around the world, promised shipments of air defense missiles haven’t arrived. So, like those on the frontlines rationing artillery shells, they know they can’t keep doing this forever.

It’s a grim feeling. More than that, it’s enraging to know that political cowardice put Ukraine in this untenable position. As one Ukrainian, who spends a lot of their time talking to Western officials, told me yesterday: “Nobody understands how bad it is.” Without that aid package, “we’re going to lose.”

Even if this aid package passes, there’s good reason to think that it will be the last.

And yet tonight, on a very special Bug-eyed and Shameless, I’m here to tell you that there’s reason to be cautiously optimistic. Ukraine is innovating its way out of this war. Even as they spend their nights in the air raid shelter, like I am now, they are still going to work in the morning to develop new technology that will help them defeat a better-resourced enemy. All they are asking from us is a bit of help.

All we need to do is buy Ukraine some time.

---

---

“You don’t win wars with GDP, or Euros or dollars. You win wars with weapons, and the West doesn’t make enough weapons.”

That bit of wisdom, from Senator J.D. Vance at the Munich Security Conference last month, has become an easy bit of self-justifying logic for a particular class of coward. It goes that America needs to focus on its own domestic security, and that Ukraine’s safety has nothing to do with it. It’s been deployed as a useful shield for those faux-nationalists holding up critical aid for Ukraine.

Vance, a venture capitalist and mediocre memoir writer whose entire life has been about running for office, does not serve on the armed services, foreign relations, or intelligence committees. He refused to meet with Ukrainian President Volodymyr Zelensky in Munich. He gets his news from places like Breitbart and American Greatness — slavishly pro-Trump outlets which peddle the notion that helping Ukraine means antagonizing Russia, and we mustn’t do that.

Had Vance bothered to meet Zelensky, or if he were brave enough to visit Kyiv — or, better yet, the frontlines — he would know how wrong he is.

I wrote earlier this week (Dispatch #92) about the mountain of challenges facing Ukraine. About how people like Vance have forced America to renege on its commitment, leaving Kyiv out to dry at a pivotal moment. And, too, how other NATO countries have vowed to step up, only to muddle through.

It is easy to end the conversation at how dire things are. How few rockets and shells Ukraine has left. How much Russia has expanded its own industrial production. How morally and intellectually bankrupt the American defeatniks are. How difficult of a problem recruitment is, particularly against an enemy willing to sacrifice an unlimited number of its poor.

This isn’t the whole story. And, to give some credit to Vance: He’s right. You win wars with weapons. But he’s wrong about the conclusion. Ukraine is making weapons at a feverish pace, and it’s only moving faster.

Inspired by former command-in-chief Valerii Zaluzhnyi’s essay from last year, wherein he laid out the need for a technological breakthrough to achieve victory, I’ve been talking to industrialists, entrepreneurs, military officials, and technologists — American, Canadian, European, and Ukrainian. Over the last few days, I’ve met defense executives, soldiers, and fixers trying to secure necessary money and technology. I’ve seen some of this tech first-hand, and visit a secret drone manufacturing center.

Not one will deny that this fight is difficult, and that the coming months will be tough. But they are all saying the same thing: Ukraine is making the tools of its own liberation.

This is not rosy optimism. It is a fact. Ukraine has surpassed critical capabilities that not even the Pentagon has figured out. It is scaling up production of critical kit that will degrade Russia’s ability to wage war. It is gilding complex solutions to storm the fortifications which scar Ukraine’s territory in the east.

What it needs now is time, money, and parts.

And not as much as you might think.

---

I sat down yesterday with the head of a Ukrainian defense technology firm, which works in autonomous vehicles. They described to me, from their perspective, the four stages of autonomous weapons.

One is simply an ability to use artificial intelligence to collect a massive amount of information and combine it into a useful product. That is, you can dispatch multiple drones — piloted by human operators — which are collecting images and radar data of the sky ahead and the ground below. AI can help synthesize all that data into a more useful product, in order to inform better decision-making.

Two is “co-piloting.” Rather than just pushing out raw data, AI is analyzing the footage and highlighting the most important features of the battlespace.

Three is planning. Here, AI ingests the data collected by those reconnaissance units and analyzes it to suggest the best possible plan of action. These plans could be pushed down to the units at a fairly complicated level. This would be like ChatGPT calling football plays — only much, much better. The plans may tell a land-based drone, moving in a formation, that if one of its fleet loses connectivity or is destroyed to change course or execute a specific set of actions. We would call this semi-autonomous.

Four is the final step: Combining everything, and letting it self-execute — but with oversight and strict rules. Surveillance drones would collect and analyze intelligence, feed it into plan-making, and push it down to a fleet of units on land in the air, while feeding real-time data to inform tactical maneuvers. Human operators may select certain targets or set certain rules, but they would act mostly as overseers.

There are steps beyond four, they noted, but those steps are scary and full of dangers to humanity. Nobody wants to go there right now.

I nodded along, generally familiar with these concepts. At least as well as a dog recognizing shapes playing on the TV.

So we’re certainly at step two, I said, and we’re experimenting with three. But we’re not yet fully at four.

They grinned and shook their head. “We’re at steps three and four.”

This, to be clear, is far beyond anything that any other country has deployed in combat. America has spent at least a decade trying to deploy networked, autonomous combined arms — but it is still being tested, and there’s not much indication that it’s been successful. Ukraine, in just two years, has made it operational. (With America’s help, it should be noted.)

Ukraine has been loathe to speak too much about its technological advances. The more it brags about its capabilities, the more risks present themselves — that Western countries will fall back to the idea that Ukraine doesn’t need help, and more acutely that Russia will target those facilities. In recent weeks, this executive’s factory was nearly hit by two cruise missiles. Near Odesa, Russia has obsessively tried to destroy stockpiles of Ukraine’s highly-classified, ship-destroying Seababy naval drones.

But they know, too, that showcasing these capabilities has real benefits. They, for one, desperately need critical components — expensive ones, but not that expensive. A $30,000 antenna can mean the difference between whether a drone reaches its intended target, an ammunition stockpile worth millions; or whether its signal will get jammed and it falls dead out of the sky.

Ukraine needs 155mm shells, there’s no doubt. On top of its useful ability to blow things up, artillery is incredibly useful in convincing your adversary to stay put: Artillery fire forces them to stay in the bunker or the trenches, instead of advancing towards you. But drones are, day by day, taking over that critical job of blowing things up. And they’re doing it with a much higher accuracy rate for a fraction of the cost.

It costs somewhere north of $8,000 to produce a single 155mm shell these days. It can often require six, or more, shells to hit its target.¹ Howitzer systems generally reach about 20km to 70km in distance. They deliver roughly 10kg of explosive.

Building a very rudimentary FPV² drone starts around $350, but can be excellently done for around $1,000 and can produce long-range bombers for $15,000. Some FPV drones in service in Ukraine boast an accuracy rate of nearly 70%. At a base level, you’re looking at 5km range all the way up to 80km. They can carry explosive payloads from 1 to 4kg, generally.

What does this tale of the tape tell us? Drones boast higher capabilities as a fraction of the cost. What might cost $40,000 in scarce shells can be done with a single $1,000 drone.

That brings the “cost-per-kill,” as that Ukrainian executive told me, way down.

It’s a grotesque measure, of course. But this is the metric that’s been imposed on Ukraine by both friend and foe: Both Russia and America.

So Ukraine has obliged. It could win this war if it were given enough kit, but NATO has refused. Now it’s being told that the war is taking too long, and NATO citizens are tired — so either go ahead and lose, or stop asking for so much. So that’s exactly what they’re doing.

At the beginning of the war, Ukraine didn’t make drones. Certainly not offensive ones.

I spoke to one drone manufacturer who, in 2022, quickly set up shop and was churning out five to 10 FPV drones per month. These were the drones that provided the early birds-eye-view of the war, and were occasionally jimmyrigged to carry a grenade to an approaching tank.

Today? They produce 5,000 of drones per month — from small 7-inch reconnaissance rotocopters to 12-inch drones capable of delivering enough explosives to destroy a tank.

The other industry executive echoed that sentiment, saying they’ve managed to dispatch more than 200 land-based autonomous vehicles to the field — land based systems are, by far, the hardest to produce and deploy. They are delivering medical equipment, destroying Russian artillery, and even — and this is incredible — evacuating wounded soldiers.

This manufacturer is 3D printing many of these drones and developing the controller system in Ukraine. More technology, such as the on-board logic and the antennas, are manufactured in China. But the manufacturer is already working to on-shore the entire process, meaning Ukraine will produce nearly the entire system, allowing them to break free from the tight supply chains and ramp up production.

All of this will “alleviate the pressure” on these NATO-supplied weapons, they told me.

Companies in this space, however, have to rely on a complicated, inefficient, and circuitous supply chain to get critical components for their vehicles. This is true even when they’re buying from NATO countries. Creating direct and simplified supply routes for antennas, chips, circuit boards and the like — from Los Angeles, say, directly to Kyiv — would be massive. The drone manufacturer, meanwhile, doesn’t necessarily need shipments of off-the-shelf hobby drones from New York or Toronto: They need top-of-the-line antennas, and Raspberry Pis, and good cameras. And money.

They also need orders. The Ukrainian government gives tax breaks to drone companies which reinvest their profits back into research and production — this one drone manufacturer, for example, only charges a 10% mark-up to their clients, mostly the Ukrainian Armed Forces. That money goes right back into building more. More orders means more drones.

Organizations like the Ukrainian World Congress has been relying on donations to directly procure drones, especially the much-needed long-range bombers. Apart from a one-time $10 million donation from the Alberta government, they rarely get government support.

I’m just scratching the surface of this whole system. Ukrainians are, remarkably quickly, figuring out both the economics, the industrial capacity, and the military capabilities needed to overcome Russian aggression.

Certainly, like I keep underlining, Ukraine still needs NATO’s weapons. I’ve heard in recent days that Ukraine desperately needs armored personnel carriers and night-vision equipment in order to more effectively evacuate its wounded from the battlefield. The interceptions happening over my head right are only possible thanks to NATO-made missile systems.

But what Ukrainians have been trying to get across is that they have built remarkable technology and an extraordinary industrial capacity — hampered only by a lack of components, cash, and orders. They have leapfrogged over the Pentagon’s R&D efforts and started producing that technology at scale.

And yet we act like Ukraine is the charity case.

But this won’t work if Kyiv’s air defenses fail, nor if frontline artillery runs dry. If Russia is allowed to raze cities and massacre Ukrainian troops, this was all for naught.

It’s daylight now. The Ukrainian air defenses systems didn’t just work, they worked spectacularly, downing 31 cruise and ballistic missiles. It looks like Russia also fired, to little effect, multiple Kh-47M2 Kinzhals, hypersonic ballistic missiles which cost $10 million per unit.

The consequences were still dire: Homes were ruined, apartment buildings caught fire, cars were destroyed, even a daycare was burnt. Some 10 people suffered injuries, including from shrapnel. This was, considering the circumstances, one of the best possible scenarios.

If America, Canada, Europe, and NATO don’t wake up, we’re all going to see what the worse case scenario looks like.

---

---

That’s it for this special dispatch. I hope it was at least slightly more optimistic than the previous one.

For more on Ukraine’s effort to build an army of drones, head over to The Counteroffensive, an outlet which I cannot recommend enough:

Honey, there’s a Ukrainian drone in the kitchen

For some real fun, head over to WIRED to read how the Ukrainian government helped recreate Europe’s largest salt mine: In Minecraft.

If you’re able, please do subscribe. This dispatch is free, because I want to make this content accessible. But it was not free to produce: In fact, this trip was very expensive!

If you’re already a subscriber, thank you. Consider buying a subscription for a friend:

Until next time!

1. First Person View: This generally refers to aerial drones which require a pilot.

2. This isn’t necessarily true. The U.S. Army says experienced crews should be able to hit their target with just one or two rounds, but absent some advanced U.S. technology it could be five, six, or more. So let’s say six.