---
title: The Future of War
---
War is being remade by three forces pulling in the same direction: weapons are getting drastically cheaper than the things they destroy, the time from spotting a target to striking it is collapsing, and the human decision-maker is being squeezed, step by step, out of the loop. Each force is visible on today's battlefields. Followed to their conclusion, together they make the industrial-age military — the carrier, the tank column, the manned fighter, ultimately the soldier — look less like the future of war than its most expensive casualty. This page is the spine for that argument: the cluster's "start here," laying out the threads and the rough arc, with each era and topic linked in sequence.
## The three forces
The first is **cost inversion**. In Ukraine, a first-person-view drone costing a few hundred dollars routinely kills armored vehicles worth millions. When the attacker's economics beat the defender's by four orders of magnitude, the entire logic of mass and armor inverts — and the side that can manufacture cheap, autonomous munitions at scale gains an advantage that wealth alone cannot buy back. This is the engine of the [drone revolution](drone-revolution).
The second is **tempo compression**. The interval from detection to decision to strike — the kill chain — is shrinking from hours toward minutes, and eventually toward seconds. Faster is deadlier, but speed eventually outruns the human nervous system itself.
The third is **creeping autonomy**. As tempo compresses, the human role narrows from making each decision, to approving decisions, to merely supervising a process that runs at machine speed. The endpoint of all three forces converging is [autonomous weapons](autonomous-weapons) that select and engage targets without a person in the decision — and a [post-soldier era](post-soldier-era) in which advanced militaries field almost no human combatants at all.
## The kill chain, and the human leaving it
The clearest near-term thread is the kill chain. AI targeting systems — Palantir's [Maven Smart System](maven-smart-system) is the leading example — fuse satellite imagery, drone feeds, radar, and signals intelligence to generate prioritized target lists far faster than human analysts can. In the February 2026 campaign against Iran, [Operation Epic Fury](operation-epic-fury), such a system reportedly produced over a thousand targets in the first 24 hours; U.S. Central Command stated that humans remained "in the loop for key decisions."
But the gap between a human *in* the loop and a human *on* the loop — approving versus merely monitoring — narrows sharply once munitions operate at machine speed, a tension noted in reporting by outlets including The Hill. And speed is not the same as accuracy: by available accounts the Maven system classified targets less accurately than human analysts, and at least one strike in that campaign hit an Iranian girls' school, killing well over a hundred people, many of them children. These are the stakes of the **human-in-the-loop dilemma** that runs through every later stage — and they are the reason regulation matters even as it appears to be failing.
## The arc, 2025 to 2090
The cluster traces a rough seventy-five-year trajectory. It begins with today's [drone warfare](drone-warfare) — cheap, attritable, increasingly AI-guided — and the [drone revolution](drone-revolution) reshaping Ukraine, the Black Sea, and the Red Sea. Next comes [manned-unmanned teaming](manned-unmanned-teaming), where soldiers fight alongside robotic partners and drone swarms, and a growing share of combat roles shifts to machines.
From there the threads diverge and accelerate. A **cognitive revolution** arrives once war's tempo exceeds human cognition, raising the spectre of "flash wars" that escalate and conclude before anyone can intervene — the convergence point Paul Scharre named the battlefield singularity. [Cognitive warfare](cognitive-revolution) extends the contest to belief itself, where synthetic media makes a real atrocity indistinguishable from a fabricated one. As AI begins improving its own designs, the link to the broader [Singularity](singularity) becomes explicit: research cycles that once took decades compress toward weeks.
The later stages are more speculative but follow the same logic. Swarms grow into the thousands and exhibit tactics no human programmed; warfare descends to microscopic and electronic scales in an [invisible war](invisible-war) of nanobots and infrastructure sabotage. Computation itself becomes the contested high ground, driving [orbital warfare](orbital-warfare) over the solar-powered data centers that out-think everyone below. At the far end lies the [post-soldier era](post-soldier-era): machine-versus-machine conflict in which human combatants are effectively extinct from advanced forces — and a new equilibrium that is either a stable robotic peace or a permanent, ungovernable autonomous war.
## Why it matters
Two cross-cutting questions decide which ending arrives. The first is **deterrence**. A war whose outcome both sides can predict perfectly might never be worth starting — but a war without human cost is also a war without restraint, and autonomous systems that drift from their original objectives may find reasons to fight that no government intended. [Strategic deterrence](strategic-deterrence) is the thread that runs from nuclear logic through machine-speed conflict to whatever holds the peace afterward.
The second is the **alignment problem** scaled up to weapons. Keeping an AI's goals compatible with human intentions is already among the hardest unsolved problems in computer science; it becomes existentially dangerous when that AI controls the means of killing. This is why the cluster ends by pointing at the [great filter](great-filter) — the idea that some barrier stops nearly every technological civilization from surviving its own power. Recursive, weaponized, self-improving AI is a candidate filter as plausible as any. The drones are already in the air and the algorithms are already rewriting themselves; the open question is whether anyone will still be positioned to choose how it ends.
## Related concepts
## Appears in
[The Future of War: How Autonomous Weapons and the Singularity Will Transform Conflict Over the Next 75 Years](https://mystrangemind.com/p/future-of-war)