What lessons from the Ukraine and Gaza conflicts are reshaping military doctrine and procurement worldwide
Wars Are Changing How Militaries Buy Weapons and Fight — Here's What the Data Shows
Based on analysis of a 138-node, 460-edge knowledge graph mapping lessons from the Ukraine and Gaza conflicts across military doctrine, procurement systems, and international law…
What this is actually about
Two ongoing wars — Russia’s invasion of Ukraine and Israel’s campaign in Gaza — have become the biggest real-world military experiments in a generation. Militaries around the world are watching closely, taking notes, and changing what they buy and how they plan to fight.
A knowledge graph maps out all the ideas, events, technologies, and rules that connect to each other in this process. Think of it like a very complicated web of “A caused B, which led to C, which changed D.” This document explains what the shape of that web reveals — not just what’s in it, but how it’s connected.
The most important discovery: drones made the battlefield see-through
The single biggest idea in this entire graph is something called “drone-transparent battlefield doctrine.” Here’s what that means in plain English: for most of military history, you could hide. Troops, tanks, supply trucks — if you stayed off roads, used darkness, and kept quiet, the enemy might not know exactly where you were.
Cheap drones — the kind that cost a few hundred dollars — changed that. Combined with commercial satellites that anyone can pay to use and apps like Telegram where soldiers post photos and videos, the modern battlefield has almost no hiding places. If you are above ground, you can probably be found. And if you can be found, you can be targeted.
This one shift connects to almost everything else in the graph. It explains why old-style tank formations have become dangerous. It explains new attention to tunnels and underground warfare. It explains why soldiers operating drones from kilometers away are themselves now considered high-value targets. Almost every other change the graph tracks flows through this central idea.
Ukraine as the world’s biggest weapons testing lab
The graph has a special node that represents what you might call the “Ukraine laboratory effect.” Ukraine, by necessity, became the fastest-moving weapons development and testing environment on earth. When a drone design stopped working because the enemy found a way to jam it, Ukraine needed a new design in weeks — not the years that normal military procurement takes.
This created a kind of live feedback loop: test something on Tuesday, find out it doesn’t work by Thursday, build something different by the following Monday. Ukraine developed a procurement system called BRAVE1 specifically to support this speed. Instead of a slow government approval process, it acted more like a tech startup accelerator for weapons.
Here’s why this matters beyond Ukraine: the knowledge graph shows this node acting as a bridge between two halves of the map. On one side are all the battlefield lessons — what works, what doesn’t, how drones fight, how tunnels complicate things. On the other side are institutions — NATO, the European Union, Poland’s military buildup, the American defense procurement system. Ukraine’s laboratory effect is almost the only connection between those two halves. If you removed it, the lessons and the institutions would no longer be talking to each other.
The giant gap between how fast weapons evolve and how fast bureaucracies move
The graph identifies something called the “3-6 month obsolescence cycle.” A drone design that was effective in January may be useless by April because the other side found a way to jam it, blind it, or shoot it down cheaply.
This creates a problem for countries that buy weapons through normal government processes. The United States spent years and considerable resources on a program called the Replicator Initiative — an attempt to mass-produce small drones. The graph shows it placed in direct contrast to Ukraine’s BRAVE1 model, described as a failure. The reason, according to the graph’s structure, is not money. NATO countries are committing to spend more. The problem is time. Bureaucratic procurement systems designed for buying fighter jets over ten-year cycles do not work for a technology that changes every few months. The graph predicts this problem persists regardless of budget increases, because the binding constraint is institutional speed, not funding.
The cost trap that keeps going in circles
Here is a feedback loop that the graph makes visible: cheap drones are very hard to defend against economically. Shooting down a drone that costs $500 with a missile that costs $100,000 is not a sustainable strategy. This is called a cost-exchange ratio problem — the attacker spends almost nothing, the defender spends a fortune.
So defenders developed cheaper interceptors and layered defense systems. But then attackers responded by launching drones in large swarms, because if any individual drone is likely to be shot down, you send hundreds. Which means defenders need to shoot down hundreds. Which brings back the original cost problem.
The graph shows this as a loop: the cost problem drives the search for a solution, and the solution drives a counter that restores the problem. What keeps the loop spinning is outside pressure — specifically, the continued development and deployment of cheaper attack systems that re-assert the original imbalance.
Rules about war are falling further and further behind the technology
The graph has a consistent structural feature that is easy to miss but important: every path from “AI systems making targeting decisions” to “legal frameworks governing those decisions” runs in one direction only. Operational practice shapes legal reality; legal frameworks do not shape operational practice back.
Think of it like water flowing downhill. The laws, treaties, and accountability frameworks sit at the bottom. The technology and battlefield decisions sit at the top. Water flows one way.
There is one exception: a rule called the Laws of Armed Conflict creates a very weak constraint on the AI arms race — but the graph marks it as the lowest-weight constraining edge in the entire system. It is technically there, but it is not structurally doing much.
The graph also shows a non-obvious pathway: drone operators under sustained high-tempo combat develop psychological strain. That strain creates operational pressure to reduce how often a human actually reviews a targeting decision before the weapon fires. That pressure gets embedded into how AI targeting systems are designed. So battlefield mental health becomes, through a chain of structural connections, an international law problem. This is not a connection that comes up in most policy discussions.
The tunnels problem nobody is solving (except adversaries)
Gaza’s tunnel network predates modern drone surveillance by decades. But the graph reveals something structurally interesting: tunnel warfare is the most effective known counter to drone-transparent battlefield doctrine, and almost no Western military is adopting it.
The doctrine — moving forces underground to avoid drone surveillance — is structurally confined in the graph to adversary actors: Hamas, and to some extent Chinese military planners thinking about Taiwan. It connects back to the drone-transparency cluster (it counters it), but it does not connect forward to any Western procurement programs, NATO planning documents, or institutional adoption chains.
This is an asymmetry the graph makes visible: the dominant new doctrine (transparent battlefield, drone saturation) is spreading widely. Its primary known tactical counter is spreading only among the parties the dominant doctrine is designed against. Why this is the case is not encoded in the graph — but the structural gap is there.
The military now depends on things it doesn’t control
Three nodes in the graph sit inside military critical pathways but are civilian or commercial: Starlink (a commercial satellite internet service), commercial satellite imagery companies, and Telegram (a messaging app). The graph shows Starlink connected to five separate military systems including targeting speed and autonomous drone navigation.
This is not an editorial observation — it is a structural one. The graph encodes a situation where a kill chain (the sequence of steps from detecting a target to firing a weapon) runs through infrastructure that militaries do not own, cannot fully secure, and cannot replace quickly. This is described in the graph as a dependency, not an asset.
The one threat the graph identifies but doesn’t resolve
The most uncomfortable structural feature in the graph is what happens around Taiwan and semiconductors. The advanced chips that power AI targeting systems, autonomous navigation, and drone guidance are manufactured in a very small number of places. The graph shows this dependency receiving inputs from multiple threat vectors — Chinese hypersonic missiles that could strike chip factories, China’s own AI military development, drone obsolescence cycles amplifying the demand for new chips. But the graph shows no outgoing connections. No resolution. No alternative supply chain. No mitigation pathway.
The graph identifies the vulnerability clearly and then, structurally, goes silent.
Bottom line: what the shape of this graph actually tells us
Five structural insights stand out:
The most-connected nodes are not the most-important nodes. The high-weight nodes (the ones representing well-established, empirically grounded developments) are the sources. The highly-connected, low-weight nodes are the transmission mechanisms — theoretical concepts that the graph uses to connect everything. This means the graph’s architecture distinguishes between “things we know happened” and “how we think they connect.”
Ukraine’s battlefield is the only current bridge between operational lessons and institutional change. If that bridge degrades — if Ukraine’s conditions become unique rather than generalizable — the operational and institutional halves of the map decouple.
Accountability for AI targeting is structurally downstream of AI targeting. The laws follow the technology; they do not currently guide it. The graph does not predict whether this changes.
The primary tactical counter to the dominant new warfare doctrine is spreading in the wrong direction — toward adversaries, not toward the militaries promoting the doctrine.
Commercial infrastructure is inside the kill chain. This is not metaphorical. The graph encodes it as a literal structural dependency at the operational level.
The graph does not tell you what to do about any of this. It tells you how the pieces connect. What you do with the map is a different question.