Chapter 2 / Analysis
Chapter 2 № 02 · 2025

Drones, AI, and Middle Power Defense

Cost asymmetry restructures security itself

$4 Million vs. Tens of Thousands

One Patriot missile: $4 million. The unit cost of the drone interceptors Ukraine uses: tens of thousands of dollars. A hundredfold cost gap.

This is not a story about cost savings. It means the structure of military power itself is changing.

The old security model:
Great-power military strength expensive weapon systems
enormous defense budgets
only great powers can afford it
The new security model:
Drones + AI low-cost means of defense
middle powers can afford it
the military superiority of great powers shrinks

The 2026 Iran War — The Structure a Superpower Cannot Beat

In March 2026, the United States launched "Operation Epic Fury" against Iran. Five weeks of daily airstrikes by the world's strongest military. More than 12,300 targets struck.

The result?

US intelligence assessment (as of April 2026):
Missile launchers about 50% still intact
One-way attack drones thousands remain (about 50% of capability)
Coastal-defense cruise missiles mostly intact
IRGC navy hundreds to thousands of small boats and unmanned surface vessels remain
Strait of Hormuz cannot be reopened; the US itself admits it

President Trump said it would be "over in two or three weeks." A US intelligence official replied: "Anyone who thinks that is out of their mind."

Why can't a superpower win? Look at the structure and it is obvious.

Underground tunnels + mobile launchers = the asymmetric shield Iran spent decades building a vast network of tunnels and caves. Launchers hide underground, come out, fire, and go back under. The US bombs tunnel entrances — even bulldozers and heavy equipment — but it cannot physically destroy everything underground. The same structure as Yemen's Houthis.

The reality of cost asymmetry One US precision-guided bomb = hundreds of thousands to millions of dollars. One Iranian drone = tens of thousands. Strike more than 12,300 targets, and you still cut only half the opponent's capability. The attacker's cost exceeds the value of the defender's assets. This is the essence of asymmetric war.

The Strait of Hormuz — the wall of coastal missiles Most of Iran's coastal-defense cruise missiles remain, because the US campaign focused on inland missile facilities. Hundreds to thousands of IRGC small boats and unmanned surface vessels remain, and keep threatening shipping in the strait. The world's strongest navy cannot open a single strait.

The Ukraine War — The Other Asymmetric War

Ukraine was the first in the world to prove that drone-plus-AI defense can stand against a great power's conventional forces. Not a technology demonstration — proof in combat.

Three structural changes deserve attention.

1. The inversion of attack cost and defense cost The old rule said attack requires three times the force of defense. With drones, the attacker's cost is extremely low. A tank formation Russia deploys for billions of dollars can be destroyed by FPV drones costing tens of thousands. When the defender uses drones, the attacker's expensive weapons are neutralized at low cost.

2. "More vs. sufficient," demonstrated Great powers pursue "more" — more weapons, higher performance. Drones need only "sufficient" — enough units at sufficient accuracy. No need for one precision missile that must hit. Fly ten cheap drones; one hit is enough. The cost structure is fundamentally different.

3. From security "recipient" to security "provider" Ukraine signed ten-year defense agreements with three Gulf states. A country that once depended on others for its security became a security provider, on the strength of its drone technology and operational know-how. There is no precedent for this reversal in history.

The Decisive Element AI Adds

Drones alone are not enough. When AI joins the drone, the change becomes structural.

Capability Drones only Drones + AI
Piloting human operator required autonomous flight; resistant to jamming
Target search humans watch camera feeds automatic detection and classification
Swarm control one operator per drone one person runs dozens simultaneously
Link loss return or crash autonomous judgment, mission continues
Adaptation by-the-manual operation learns enemy countermeasures, changes tactics
Night / bad weather difficult sensor fusion, 24-hour operation

Swarm control matters most. Without AI, one drone needs one human operator. With AI, one person commands dozens. Large effect from few personnel — decisive for a middle power.

Middle-Power Defense Changes at the Root

Traditionally, a middle power had two options for its security.

Option A: alliance with a great power the cost of political subordination
Option B: military buildup economically impossible
Option C (new): drone + AI defense low-cost, self-reliant deterrence

Drones + AI give middle powers a third option. No need to buy fighter jets costing tens of billions. Drones costing tens of thousands, plus AI systems, build a defense sufficient for deterrence.

Why it works even on 1% of GDP: One fighter jet = ¥10 billion. One drone = a few hundred thousand yen. For the same budget, deploy tens of thousands of drones instead of one jet. With AI swarm control, a few hundred operators run tens of thousands of drones. A middle power's limited budget and personnel can make invasion expensive.

The Structural Change in Deterrence

The essence of deterrence: make the cost of attack higher than the benefit of invasion.

Nuclear weapons achieved this in the ultimate form. But middle powers effectively cannot have them (the NPT regime, cost, technical barriers).

Drones + AI achieve the same effect at a different level. Continuously attrit the invading force at low cost, until "attacking this country doesn't pay."

The three-layer structure of deterrence:
Layer 1: reconnaissance drones + AI 24-hour surveillance; approach detected instantly
Layer 2: interceptor drone swarms destroy the attacker's weapons at low cost
Layer 3: long-range strike drones retaliation; threaten the invader's rear

This three-layer structure corresponds to what used to be an air force (reconnaissance aircraft + fighters + bombers). But the cost is under one hundredth. Personnel, under one tenth. A middle power can build it.

The New Zealand Example

New Zealand: population five million. Defense budget around 1.5% of GDP. Geographically isolated; security traditionally dependent on alliance with Australia and the US.

Under the drone + AI defense model:

Maritime surveillance — long-range drones + AI image recognition keep the entire exclusive economic zone (EEZ) under constant watch. Cheaper than patrol vessels, wider coverage

Coastal defense — anti-ship drone swarms make amphibious operations extremely expensive. Saturation attack on an invasion fleet with cheap drones

Personnel efficiency — AI swarm control lets a small number of operators run large numbers of drones. Sufficient defense even for a nation of five million

This is not theory. It is the model Ukraine has already proven, applied to an island nation.

Ukraine's Use of AI — The State of the Art in Combat (2026)

Ukraine fields drones + AI at the largest scale and fastest pace in the world. Since the full-scale invasion of 2022, Ukraine has shed the rigid Soviet-style, state-led R&D model and become a "Silicon Valley of defense technology" — integrating private-sector capability with real-time feedback from the front.

Astronomical industrial growth: 2022: domestic defense-industry production capacity about $1B 2025: estimated $35B (35x) 33–34% of weapons used on the battlefield are domestically produced (exceeding the European share) 2025: plans to manufacture up to 4.5 million military drones, budgeted at over 110 billion UAH (~$2.6B) 100% of drones striking inside Russian territory are domestic

But the biggest reason Ukraine's weapons evolve so fast is not technical superiority. It is innovation in the "feedback loop" and the "procurement process" that adapt technology to the battlefield. That is the most important insight of this section.

The inversion of battle ratios In early 2024, about 10% of battlefield strikes were drone-delivered. By 2026, about 70%. Artillery strikes fell to 10–15%. In just two years, the lead role passed from shells to drones. In March 2026, 96% of Russian combat losses (35,551 men) were attributed to drones. [needs verification: CEPA, Kyiv Independent, CSIS and other reporting]

AI quadrupled the hit rate Manually piloted FPV drones hit their targets around 10–20% of the time. Add an AI autonomous-navigation and terminal-guidance module (about $50 per drone), and the hit rate jumps to 70–80%. A $50 AI module multiplies one drone's results about fourfold — making a weapon already costing only tens of thousands another order of magnitude more efficient. [needs verification: IEEE Spectrum, Ukraine's Arms Monitor, etc.]

The Victory of Institutional Architecture — Brave1 / Delta / Army+ / E-Points

Ukraine evolves faster than Russia not because it has Patriots, and not because its AI is better. It invented four institutions that connect battlefield needs directly to private development lines.

Brave1 — the civil-military innovation hub

In April 2023, Ukraine established Brave1, a defense-technology platform spanning the Ministry of Defense, the Ministry of Digital Transformation, the Ministry of Economy, and the National Security and Defense Council (NSDC). It connects the military's operational needs directly with private technology companies and investors.

Brave1's scale (as of 2025): Companies integrated: over 1,500 Grant projects: over 540 Total grants: 2.2 billion UAH Additional 2025 budget: 2.9 billion UAH Strategic priorities: missiles, precision-guided weapons, lasers, drone swarms, anti-jamming (CRPA) antennas, anti-Shahed interception, unmanned surface vessels (USVs)

Most notable is the "Brave1 Dataroom" — a secure environment that opens the military's classified combat data (visual data by weather, time of day, sensor type) to defense companies for AI model training. Sharing the most valuable asset — real combat data — with the private sector has sharply raised the accuracy and development speed of AI weapons.

Delta — battlefield SaaS, built bottom-up

The situational-awareness system Delta fuses drones, satellite imagery, human intelligence, and sensors in the cloud, visualizing enemy movements on a real-time digital map.

Note the development path. Delta was originally built bottom-up by Aerorozvidka (aerial reconnaissance), a volunteer organization of drone operators and programmers — in contrast to CJADC2, the top-down all-domain command-and-control program of the US Department of Defense. Delta's essence is agile software culture brought directly into military command and control.

Army+ — direct feedback from soldier to maker

The Army+ app, introduced by the Ministry of Defense in 2024, built a digital feedback loop directly linking front-line soldiers to weapons manufacturers.

The old multi-layer defect-reporting chain — command hierarchy → ministry → manufacturer — was short-circuited to soldier → manufacturer. This is the institutional basis of the development cycle shrinking "from years to weeks."

E-Points / Brave1 Market — "points for kills" and digital Darwinism

The most radical mechanism is the E-Points initiative and the Brave1 Market (dubbed the "military Amazon").

Points-for-kills price list: One enemy soldier eliminated → 6 points Tank damaged → 20 points Tank destroyed → 40 points Mobile rocket system destroyed → up to 50 points

Units upload proof-of-strike video to Delta for verification and earn points according to the value of the destroyed target. Points are spent in the Brave1 Market on the drones, AI tools, and EW systems the unit itself needs.

The system functions as a powerful evolutionary selection pressure on the defense industry (digital Darwinism). Front-line units selectively buy the platforms that punch through enemy electronic warfare and reliably score — so superior technology spreads instantly, and companies making obsolete weapons are culled from the market. Battlefield evaluation becomes market selection, directly — this is the real reason corporate iteration accelerates to the limit.

The Defense-Tech Startup Ecosystem

With the shift from state-led to private-led development, defense-tech startups have multiplied explosively. Drone makers, numbering a few dozen before the invasion, now exceed 200 including AI-equipped systems, with over 1,500 companies involved in the defense industry.

Major VC networks: MITS Capital, D3 Venture Capital, Ukrainian Startup Fund (USF), Green Flag Ventures, Nezlamni, Angel One Fund, and dozens of other funds and syndicates finance domestic defense startups.

Company Core technology
Swarmer AI swarm control in GPS-denied environments; has supported over 100,000 missions
Buntar Aerospace AI reconnaissance UAV (Buntar-3), mission-prediction copilot ($10.4M raised)
Mantis Analytics real-time information-environment monitoring and disinformation-detection AI ($240K raised)
HIMERA EW-resistant secure communications and tactical radios
Odd Systems FPV drones and thermal imaging, iterated rapidly on battlefield feedback
DROPLA TECH AI for demining and ambush detection
TELETACTICA jamming-resistant battlefield communication networks

These companies carry battle-tested tech and have international markets in view.

The State of the Art by Weapon Category

Fiber-optic FPV drones — physically neutralizing electronic warfare

With Russia's powerful electronic warfare routinely jamming radio links and cutting video feeds, Ukrainian engineers adopted an approach that severs dependence on radio signals entirely.

Fully autonomous AI drones and long-range strike aircraft

Even with GPS totally denied, next-generation drones recognize terrain with optical sensors (optical flow and terrain matching) and dive precisely into preset targets. One operation reportedly involved a coordinated swarm attack by 800 autonomous drones.

Aircraft Performance
Batyar (DeepStrikeTech, May 2025) optical terrain matching, range 800 km, warhead 18 kg
Artemis ALM-20 (Auterion + Ukrainian airframe) GPS-free AI navigation, warhead 45 kg, range 1,600 km

Interceptor drones — replacing expensive missiles

Against the Shahed/Geran-2 and reconnaissance drones Russia deploys en masse, Ukraine replaces the mission of anti-air missiles costing hundreds of thousands to millions of dollars with drones costing thousands.

System Performance and features
Octopus 100 UK co-production. >300 km/h, 4,500 m. Autonomous terminal guidance by image recognition; very high hit rate
Sting small interceptor quadcopter the size of a large thermos (30–45 cm)
P1-Sun (Skyfall) >300 km/h, heavy use of 3D-printed parts; cheap and fast
Bagnet 250 km/h; attacks loitering munitions and reconnaissance aircraft
VB140 Flamingo operational range up to 50 km, altitude 4,500 m
Merops (US-developed) AI guidance, GPS-jamming resistance, about $15,000 per unit

This approach has been deployed at US bases in Saudi Arabia together with Ukraine's command-and-control platform Sky Mapbeginning to influence air-defense doctrine worldwide.

Strategic strike — missiles and rocket drones, produced at home

Western-supplied ATACMS and Storm Shadow come with political restrictions on strikes inside Russia. To route around them, Ukraine gives top priority to developing and producing its own long-range strike weapons.

Palianytsia — a rocket drone (in substance, a low-cost cruise missile)

Length / wingspan: 3.5 m / 1.7 m Total weight / payload: 320 kg / up to 100 kg Maximum range: 650 km Top speed: 900 km/h (comparable to the Kh-101 cruise missile) Propulsion: solid-fuel booster launch → turbojet Guidance: GPS + inertial navigation (IMU)

Hrim-2 (domestic name Sapsan) — a short-range ballistic missile

With these, Ukraine acquired an autonomous capability to destroy strategic targets deep inside Russia without Western approval.

Asymmetric naval war — unmanned surface vessels (USVs)

Despite having no large surface fleet, Ukraine destroyed or drove back most of Russia's Black Sea Fleet with home-built USVs.

USV Key specifications
Magura V5 length 5.5 m, payload 300 kg, range 800 km, 78 km/h, unit cost $273K, GNSS + inertial + visual guidance
Magura W6 V5 derivative carrying two R-73 (AA-11 Archer) air-to-air missiles — "Sea Dragon" air defense: a sea drone that eliminates air threats
Sea Baby 400 hp, 400 kg+ warhead, 500 nautical miles, Starlink + Kymeta, handles 1.5 m waves
Stalker 5.0 reconnaissance + kamikaze capability

The Magura W6 in particular marks the evolution of the sea drone from prey into a platform that removes threats from the air. These USV technologies are entering the stage of being provided to Gulf states in a multinational initiative for the security of the Strait of Hormuz.

Electronic warfare (EW) — domestic production and international influence

As drone weapons multiply, the fight in the electromagnetic spectrum (electronic warfare) matters as much as physical combat or more. Since January 2024 alone, Ukraine's Ministry of Defense has signed procurement contracts for over 11,000 EW systems (about 6 billion UAH), most of them domestic.

International Spillover and Strategic Joint Ventures

Rather than relying on finished weapons from the West, Ukraine has pivoted to building weapons production and repair hubs inside the country (localization).

Financial backing: a $50B loan facility drawing on G7-frozen Russian assets (including a £2.26B loan from the UK).

Ukraine is building an entirely new model of security provision: refine technology in combat, then export it to allies. NATO is turning into the side that learns from Ukraine. This is a historic reversal.

Strategic Technology Cooperation with Japan — Rakuten, Tohoku University, AirKamuy

Japan, which does not export lethal weapons, also works closely with Ukraine's defense-technology ecosystem in the context of dual-use technology and humanitarian support.

Rakuten Group × Brave1

With the backing of Rakuten founder Hiroshi Mikitani, the company's messaging app Viber enjoys dominant penetration in Ukraine and functions as infrastructure for information sharing and military networks.

Tohoku University × the ALIS mine detector

Ukraine's farmland faces mine and unexploded-ordnance contamination on the largest scale in the world. At the Ukraine Mine Action Conference (UMAC 2025) held in Tokyo in October 2025, JICA and UNDP demonstrated the latest demining technology. At its core is ALIS, the ground-penetrating- radar mine detector developed by professor emeritus Motoyuki Sato of Tohoku University:

AirKamuy 150 — "cheap drones at scale" reaches Japan

The AirKamuy 150, developed by the Japanese company AirKamuy, is a low-cost, disposable cardboard drone.

Japan's Ministry of Defense itself is drawing lessons from the war in Ukraine and exploring cheap swarm-control drones. Even under Japan's constraint of holding no lethal export weapons, the doctrine of cost asymmetry holds.

Conclusion — AI Does Not "Make" the Judgment; Institutions Drive the Evolution

The rapid advance of weapons in Ukraine did not come from the invention of a single breakthrough weapon. In a total war for national survival, Ukraine

It is a victory of institutional architecture.

And AI's role is not to replace military judgment.

Drone video and sensor fusion AI classifies and nominates targets
operator approves mission executes
Satellite imagery, OSINT, signals intercepts AI detects threats on Delta
commander decides units move

By raising the speed and accuracy of information processing by orders of magnitude, AI lets a middle power inferior in population and budget make high-quality decisions, with a small staff, in real time. This is the other axis of cost asymmetry.

A drone costing tens of thousands × a $50 AI module × cloud battlefield SaaS × digital Darwinism — this is the new equation of middle-power defense that Ukraine has shown the world.

[Sources: survey of Ukrainian weapons development (Gemini Deep Research); Reuters, CEPA, Kyiv Independent, CSIS, Atlantic Council, IEEE Spectrum, Roland Berger, company press releases, etc.]

The Structural Connection to Natural Farming

At first glance, drone defense and natural farming look unrelated. Structurally, the same principle is at work.

The common structure: the "small vs. large" inversion
Military: a middle power's drone swarms stand against a great power's regular forces
Agriculture: small-scale natural farming avoids the fragility of industrial agriculture
AI: an individual + Claude analytical capability on par with large organizations

Large systems are efficient but fragile. Destroy one refinery and chemical fertilizer stops. Sink one carrier and air power is lost.

Distributed systems look inefficient but are resilient. Shoot down 50 of 100 drones, and 50 keep functioning. Natural farming is unaffected when chemical fertilizer stops.

The paradox of Fukuoka's "do-nothing farming" and drone defense: Masanobu Fukuoka drew out the power of the ecosystem by "doing nothing." Drones + AI make middle-power defense viable by "not owning big weapons." Both reject "more" and pursue "sufficient." The structure is the same.

Why Superpowers Stopped Winning

Vietnam, Afghanistan, Iraq, Yemen's Houthis, and now Iran — the US keeps repeating a history of losing asymmetric wars. Why?

Structural weaknesses of the superpower military model:
Centralized, expensive weapons losing one unit means enormous loss
Complex supply chains logistics lines thousands of kilometers from home
Political constraints bound to domestic opinion and election cycles
Time constraints results demanded "in two or three weeks"
Structural strengths of the asymmetric side:
Distributed, low-cost weapons lose half, and half remains
Home terrain and underground defense networks built over decades
The will to survive retreat is not an option
Time is an ally endure, and win

America Has No History — The Structural Blind Spot

The United States was founded in 1776. 250 years. Iran's civilization spans 2,500 years, since the Persian Empire. China, 4,000. Japan, over 2,000. The land of Ukraine carries over 1,000 years of history.

Having no history creates a decisive blind spot in military strategy.

1. Unable to grasp what "defending the land" means A nation with 2,500 years of history spends decades digging tunnels. That is not a rational military calculation; it is a civilizational will. "They'll negotiate and it will end" is the idea of a country that has never had to defend its own soil. Iran and Ukraine both fight not to negotiate but to survive.

2. A fundamentally different sense of time The US says "two or three weeks." Iran thinks "we have endured for 2,500 years." Afghanistan's Taliban said: "The Americans have the watches, but we have the time." A country without history cannot understand what a long war means.

3. Overconfidence in technology Precision-bomb more than 12,300 targets, and half the capability remains. The belief that technology solves everything underestimates the power of terrain, time, and will. Countries with history know from experience that technology is not all-powerful.

What Comes Next

The Iran war and the Ukraine war prove the same thing simultaneously: the age of superpowers is ending structurally.

Drone + AI technology evolves exponentially. In five years, what Ukraine uses today will be obsolete. But the structure — cost asymmetry, distributed defense, underground protection, AI swarm control — will not change.

Middle powers should start building drone + AI defense now. Expensive weapons are the archetypal "measure you keep buying" in defense: after the airframe come the maintenance, parts, training, and upgrades — the dependence lasts the life of the platform. Instead of buying expensive weapons from great powers, develop drones at home, integrate AI, and build a self-reliant defense. As Iran proved, the combination of underground infrastructure and low-cost weapons cannot be destroyed even by the world's strongest air force.

The age of relying on a great power's protection is ending.
Drones and AI hand middle powers the tools of self-reliance.
Iran goes underground on 2,500 years of history; Ukraine strikes back with drones.
A 250-year superpower can defeat neither.
History and will beat technology.