Defense Tech: what’s changing now?

Why does defense tech suddenly feel so different?

Defense tech feels different today because militaries are changing what they buy, how fast they buy it and which companies they trust to build it.

Higher military spending is part of the story, but spending alone does not explain the shift. European NATO members and Canada increased defense expenditure by nearly 20% in real terms in 2025. At the latest NATO summit, members announced more than €50 billion of new procurement covering precision strike, air and missile defense, uncrewed systems, intelligence and other advanced capabilities. NATO has also committed more than $40 billion over five years to counter-drone systems.

Look at where that money is going. The U.S. Army has created a dedicated organization for mission autonomy and opened digital marketplaces for drones. The U.S. Air Force is moving autonomous combat aircraft into production programs. Ukraine says 95% of the drones bought through its main defense procurement agency are now domestically made, while its military received roughly 3 million FPV drones during 2025.

Private companies are moving up the food chain at the same time. Anduril raised $5 billion at a roughly $61 billion valuation. Helsing raised $1.8 billion at an $18 billion valuation. Shield AI raised $1.5 billion in common equity at a $12.7 billion post-money valuation.

The change is pretty clear when those developments are viewed together. Defense still has long programs, small production runs and giant contractors, but another model is growing beside them: cheaper autonomous weapons, software that changes quickly, more suppliers and factories expected to produce weapons by the thousands.

If you want more recent data on this point, please see our latest defense tech market report.

What did the war in Ukraine actually change about defense tech?

Ukraine changed defense tech by showing that weapons can be designed, bought, modified and replaced on a completely different clock from traditional military equipment.

The scale is easy to underestimate. Ukraine's armed forces received around 3 million FPV drones in 2025, almost 2.5 times the previous year's volume. Ukrainian officials later put domestic FPV production capacity above 8 million units per year.

Procurement has grown alongside production. During the first half of 2026, Ukraine's Defence Procurement Agency signed drone contracts worth UAH 333.6 billion, twice as much as during the same period a year earlier. FPV drones represented the biggest part of those contracts.

The more interesting change is who gets to choose. Ukraine's DOT-Chain Defence marketplace lets combat units select equipment themselves within allocated budgets. For products already in stock, average delivery has fallen to around nine days. Ukraine's Ministry of Defence says 95% of the UAVs procured through the agency are Ukrainian-made.

That creates an unusually tight feedback loop between the battlefield and the factory. When jamming makes one control system unreliable, manufacturers can change frequencies, communications methods or navigation software. When radio-controlled FPVs become easier to disrupt, fiber-optic drones appear in larger numbers.

Ukraine has shown that battlefield hardware can evolve at software-like speed when users, buyers and manufacturers are connected closely enough.

Ukraine drone indicator	Earlier level	Latest reported level
FPV deliveries	Roughly 1.2 million in 2024	Around 3 million in 2025
Domestic FPV production capacity	Around 4.5 million cited previously	More than 8 million annually
First-half drone contracting	Roughly half the latest level	UAH 333.6 billion
Ukrainian share of procured UAVs	Already dominant	95%

Are cheap drones changing what militaries buy — and creating a huge counter-drone market?

Cheap drones are already changing military buying on both sides of the fight: armies want large numbers of weapons they can afford to lose, while defenders need cheaper ways to stop them.

The economics are particularly obvious in air defense. A force can technically shoot down a cheap one-way drone with an interceptor costing hundreds of thousands or millions of dollars. Doing that repeatedly against hundreds of incoming drones creates an ugly exchange ratio even when every interception works.

NATO's latest spending choices reflect that problem. Members have committed more than $40 billion over five years to counter-drone capabilities and want to train five times as many military drone operators by the end of 2027. NATO is also setting up a marketplace of tested counter-drone products instead of betting everything on one interceptor.

That demand is producing several defensive layers. Electronic warfare can disrupt some drones without firing anything. Guns and interceptor drones can handle other targets. Epirus is developing high-powered microwave systems designed to disable groups of drones electronically. Conventional surface-to-air missiles remain essential against faster or more dangerous threats.

The same cost pressure exists on offense. The U.S. military is backing weapons designed around quantities measured in the thousands. Anduril's Barracuda-500M agreement with the Army, for example, calls for at least 3,000 systems over three years.

No single counter-drone technology solves the defensive side either. Radio jammers struggle against fiber-optic systems. Interceptor drones can produce better economics against certain targets. Directed-energy systems promise a very low cost per engagement but still face limits around range, weather or power. Software increasingly has to decide which defensive layer should engage which incoming threat.

Modern arsenals are therefore developing a much wider price ladder. Militaries still need exquisite missiles and aircraft, but they increasingly want a $20,000 or $200,000 answer when a $2 million answer is unnecessary.

If you want more recent data on this point, please see our latest defense tech market report.

Can defense companies actually build enough weapons?

Production capacity is currently one of the biggest constraints in defense tech, and companies increasingly have to prove they can manufacture at scale rather than merely build impressive prototypes.

Recent wars have made stockpile math impossible to ignore. Drones can be consumed by the thousands. Artillery shells are fired continuously. Air-defense interceptors may be used much faster than factories replace them. Long-range missiles that once looked like weapons for occasional strikes can become recurring consumables.

That pressure is changing how newer defense companies present themselves. Anduril has built Arsenal-1 in Ohio around higher-volume manufacturing and has already started producing its YFQ-44A autonomous combat aircraft there. The company is also designing Barracuda around simpler production and working with Poland's PGZ on localized European manufacturing.

A newer missile company, Covenant, pushes the idea even further. Its Anthem deep-strike weapon is being designed around production sites planned for the United States, Germany and Israel. The company says each site could eventually produce as many as 5,000 missiles per year. Covenant has already conducted more than 200 test flights and secured about €130 million in contracts.

Those targets are still targets; Covenant has to prove it can achieve them. But the fact that a missile startup now talks about factories and annual output as prominently as range or payload shows how much the market has moved.

Today, manufacturability can decide whether a weapon is strategically useful. A slightly less sophisticated system available in huge numbers may create more combat power than the perfect weapon sitting behind a multi-year production queue.

Are autonomous combat aircraft actually becoming real?

Autonomous combat aircraft are moving into real military programs now, although today's systems still operate with far more human control than the phrase “autonomous fighter” sometimes suggests.

The U.S. Air Force's Collaborative Combat Aircraft program is the clearest example. General Atomics' YFQ-42A and Anduril's YFQ-44A have progressed through flight testing, and the Air Force has awarded contracts covering both air vehicles and mission-autonomy software.

The architecture around those aircraft may prove just as important as the airframes. The Air Force has demonstrated its government-owned Autonomy Government Reference Architecture across multiple vendors. Shield AI's autonomy software has been paired with Anduril's YFQ-44, while RTX Collins has worked with General Atomics' YFQ-42.

The Air Force is trying to avoid buying one aircraft whose autonomy remains permanently tied to the original manufacturer. Software from different suppliers can potentially compete and evolve separately from the aircraft itself.

Shield AI is taking a similar approach internationally. The company's Hivemind platform is designed to operate across different aircraft, and its deals increasingly combine physical systems with autonomy software and simulation.

We still do not really know how far these aircraft can be trusted in chaotic combat involving electronic warfare, deception, incomplete information and strict rules of engagement. But the contracts, flight programs and production work already place autonomy inside mainstream force planning.

Why is military software suddenly so important?

Military software matters much more now because modern forces can collect far more information than people can manually turn into useful decisions.

Palantir's Maven Smart System is a good example of the scale involved. The U.S. Army awarded a $480 million Maven prototype agreement in 2024 and followed it with a contract modification worth up to $795 million for additional software licenses. We are talking about more than $1 billion of potential contract value around software used to combine data and support battlefield decisions.

The underlying problem appears everywhere. Satellites produce imagery. Radars track objects. Reconnaissance drones send video. Electronic sensors collect signals. Weapons need target coordinates. A military can own excellent equipment across all of those categories and still perform badly if the information arrives too slowly or cannot move between systems.

Open architecture has become more important for the same reason. The Air Force recently published common technical standards for mission systems and autonomy, allowing software to be separated more cleanly from individual aircraft. Militaries want to replace an autonomy package, sensor or application without redesigning the whole platform every time.

Ukraine has reached a similar conclusion through battlefield pressure rather than formal architecture programs. Drone airframes, radios, navigation systems and software are changing so quickly that tightly closed products can become obsolete within a short procurement cycle.

Military AI fits into this software layer as well. Autonomy gets the attention, but targeting, intelligence analysis, logistics and command software may touch a much larger part of the force.

Can defense startups really challenge Lockheed Martin and RTX?

Defense startups can already take major programs from traditional contractors in specific categories, and several are clearly becoming large defense companies rather than niche suppliers.

Anduril is furthest along. Its latest $5 billion financing valued the company at around $61 billion, roughly twice its valuation less than a year earlier. It now works across autonomous aircraft, missiles, counter-drone systems, command software, underwater vehicles and manufacturing.

Helsing shows that the pattern is spreading outside the United States. The European defense AI company recently raised $1.8 billion at an $18 billion valuation and has expanded beyond software into drones, aircraft and strike systems.

Shield AI has followed another route around autonomy. Its latest financing package included $1.5 billion of Series G equity at a $12.7 billion post-money valuation plus $500 million of preferred equity. It then completed the acquisition of simulation specialist Aechelon, bringing a technology already widely used across the U.S. military into the company.

Those businesses still sit far below Lockheed Martin, RTX, Northrop Grumman, General Dynamics or BAE Systems in revenue, installed equipment and industrial depth. Nuclear systems, submarines, large aircraft programs and integrated missile defense remain brutally difficult businesses for a young entrant to penetrate.

The competitive map has still changed. A startup can now win its own program, build its own factory, raise billions of dollars and sell a complete military system directly to governments.

If you want more recent data on this point, please see our latest defense tech market report.

Why are investors putting so much money into defense tech now?

Investors are pouring money into defense tech because the sector finally combines huge government budgets with believable paths to building very large private companies.

Crunchbase's narrower definition of military, national-security and law-enforcement startups shows how quickly the capital has moved. Funding went from about $1.6 billion in 2020 to $3.9 billion in 2021 and reached a record $9.6 billion for all of 2025. By early June 2026, companies in those categories had already raised more than $14.6 billion.

That is more money in roughly five months than during the entire previous record year.

The financing is heavily concentrated. Anduril alone raised $5 billion. Helsing raised $1.8 billion. Shield AI raised $1.5 billion of common equity as part of a larger financing package. Large rounds are doing much of the work, which suggests investors are increasingly picking companies they think could grow into new prime contractors.

The enthusiasm can still outrun the underlying businesses. Government revenue takes time to recognize, factory construction consumes capital, individual programs can disappear and some startup valuations have risen far faster than disclosed sales.

Even after allowing for that, defense has plainly moved out of venture capital's fringe.

Are militaries really buying new technology faster?

Military procurement is becoming meaningfully faster for drones and other fast-moving technologies, although nobody should confuse that with the disappearance of traditional procurement.

Ukraine provides the most extreme example. Through DOT-Chain Defence, combat units can choose qualified equipment directly, and in-stock orders reach them in an average of about nine days. The system creates continuing competition between suppliers rather than selecting one product and freezing the requirement for years.

The U.S. Army has started borrowing some of that logic. Its drone marketplace first launched as an online storefront for vetted unmanned aircraft. It has since expanded into a wider Department of War Drone Marketplace covering aerial, maritime, ground and counter-drone products. Units can compare products, provide feedback and order approved equipment.

The Army has also reorganized its acquisition structure, including a dedicated Mission Autonomy organization. The Air Force is testing another approach through Collaborative Combat Aircraft, where several companies compete across airframes and software under common architectures.

A submarine or strategic bomber will obviously continue to require long testing and procurement cycles. For drones, autonomy and software, though, a five-year buying process can leave the military purchasing yesterday's product.

Are missiles becoming a startup market?

Missiles are currently becoming a credible startup market because governments want cheaper long-range weapons in quantities that traditional procurement rarely contemplated.

Anduril's Barracuda-500M is one of the clearest examples. The U.S. Army framework calls for a minimum of 3,000 systems over three years. Poland's PGZ is also preparing localized production with Anduril.

The U.S. Air Force is pursuing the same idea through its Family of Affordable Mass Missiles effort, where Anduril, CoAspire and Zone 5 Technologies have all participated. The goal is to create precision-strike weapons cheap enough to buy in much larger numbers.

Covenant has now made the startup angle harder to dismiss. The company emerged publicly with its Anthem deep-strike missile after more than 200 test flights, about €130 million in early contracts and $250 million of financing. Its planned factories are designed around output of as many as 5,000 missiles per site annually by 2028.

Incumbents still dominate the missile industry and have decades of experience in propulsion, seekers, warheads, testing and reliability. Startups face a difficult jump from successful demonstrations to repeatable manufacturing.

Even so, when militaries need thousands of long-range weapons rather than a few hundred premium ones, simpler design and manufacturing cost become much more valuable.

Program	Reported quantity ambition	What it shows
Barracuda-500M	At least 3,000 over three years	Affordable strike is moving into volume procurement
Air Force affordable-missile programs	Thousands envisioned	Multiple suppliers can compete on cost and scale
Covenant Anthem	Up to 5,000 per factory annually	Startups are designing missiles around mass production

If you want more recent data on this point, please see our latest defense tech market report.

Is military space moving toward lots of smaller satellites?

Military space is increasingly moving toward large constellations of smaller satellites because distributed networks are harder to disable and can be refreshed much faster than a few exquisite spacecraft.

The Space Development Agency's Proliferated Warfighter Space Architecture is the clearest example. More than 60 Tranche 1 data-transport satellites are now in orbit, creating a low-Earth-orbit network intended to move military information quickly.

Missile tracking is expanding on top of that network. The Space Development Agency recently awarded approximately $1.75 billion for another 36 missile-warning, tracking and missile-defense satellites. Sierra Space received roughly $798 million for 18 spacecraft, while L3Harris received roughly $955 million for another 18.

That comes on top of earlier Tranche 3 awards worth around $3.5 billion for 72 tracking satellites split among Lockheed Martin, Northrop Grumman, L3Harris and Rocket Lab.

The supplier mix is important. A company such as Rocket Lab can now sit beside some of the largest traditional aerospace contractors on a major military constellation. Sierra Space is doing the same in the newer missile-defense tranche.

Military space is beginning to inherit some economics from commercial constellations: repeated manufacturing, frequent launches, interchangeable batches and several competing suppliers.

Are European defense startups finally catching up with the US?

European defense startups are closing the gap quickly in funding and ambition, although the U.S. still has a much deeper ecosystem.

Helsing gives us the strongest example. The company recently raised $1.8 billion at an $18 billion valuation, one of the largest private defense financings ever completed in Europe. Its product range has also expanded from defense AI software toward autonomous aircraft, drones and strike systems.

The broader European market has changed around companies like Helsing. NATO members are committing much more money to defense, European governments are pushing domestic production, and military customers are increasingly willing to work with companies that did not exist when the previous generation of major weapons programs began.

Cross-border production is becoming more common too. Anduril's partnership with Poland's PGZ to manufacture Barracuda missiles locally shows that the future may involve American technology and European industrial capacity working together rather than a clean split between U.S. and European suppliers.

Ukraine adds another layer. Its domestic drone industry has reached a scale and pace of battlefield iteration that Western European startups rarely experience. The harder task will be turning those wartime manufacturers into companies that can export, certify products and compete internationally once procurement becomes less emergency-driven.

The United States still has advantages that Europe cannot match easily: a single enormous Pentagon customer, larger venture funds, stronger precedents such as Palantir, SpaceX and Anduril, and a more mature network of defense-focused investors.

Europe has nevertheless passed an important threshold. Building a private defense company worth more than $10 billion is no longer a uniquely American phenomenon.

Will drones make tanks, fighter jets and warships obsolete?

Drones are making tanks, fighter jets and warships harder to use safely, but there is currently little evidence that militaries are preparing to abandon those platforms.

Ukraine has shown how badly armored vehicles can suffer when cheap drones provide continuous reconnaissance and attack from above. That has changed how tanks move, where they hide and how much electronic warfare and short-range air defense they need around them.

Warships face a similar pressure from aerial and maritime drones. An inexpensive uncrewed vessel can force a ship worth hundreds of millions or billions of dollars to stay farther away, expend defensive weapons or accept greater risk.

Combat aircraft show the likely direction most clearly. The U.S. Air Force is developing Collaborative Combat Aircraft to fly with crewed fighters. It is adding cheaper autonomous aircraft around the premium platform rather than betting on an immediate replacement.

Ground robotics follows the same practical logic. Ukraine increasingly uses small wheeled or tracked unmanned ground vehicles for resupply, casualty evacuation and combat tasks. Militaries are generally choosing machines built around a specific job rather than humanoid robots designed to imitate a soldier's body.

Future forces could therefore become much more uneven. A relatively small number of expensive aircraft, ships and armored platforms may coordinate much larger fleets of cheaper sensors, decoys, drones and weapons.

If you want more recent data on this point, please see our latest defense tech market report.

What is holding defense tech back right now?

The biggest bottleneck in defense tech today is getting promising technology into large military formations fast enough and in enough quantity.

We already have plenty of useful drones, autonomy software, sensors, counter-drone systems and AI tools. The harder work begins when a military tries to make thousands of them communicate, survive electronic warfare, receive spare parts, fit existing doctrine and remain usable after repeated software updates.

Manufacturing sits right beside integration as a constraint. A company can demonstrate a missile ten times and still fail when asked to make 5,000 of them. Suppliers of motors, propulsion systems, batteries, semiconductors, explosives and specialty materials can become the real limit long before demand disappears.

People create another bottleneck. NATO's plan to train five times as many drone operators by the end of 2027 says something important: buying equipment is quicker than building the organizations needed to use it properly.

Then comes interoperability. The Air Force's new open architecture standards, Ukraine's marketplace model and NATO's push toward compatible counter-drone systems are all attempts to solve versions of the same problem. Militaries want fast innovation without ending up with thousands of disconnected products.

A good prototype can win attention and funding. A company becomes militarily important only when it can manufacture repeatedly, integrate with the rest of the force and keep supporting the product after deployment.

So what’s actually changing in defense tech now?

Defense tech is moving toward a faster, higher-volume and more software-driven model of warfare, while expensive traditional weapons remain central to the force.

The strongest change is the return of quantity. Ukraine procures drones in the millions. The U.S. is pursuing affordable missiles in the thousands. NATO has committed more than $40 billion to counter-drone defenses. Missile startups are designing factories around annual output measured in thousands.

At the same time, software is moving deeper into weapons and command systems. The Air Force wants autonomy that can move between aircraft. Palantir's Maven contracts show that battlefield data software can become a billion-dollar military program. Mission autonomy has become important enough for the Army to build a dedicated acquisition organization around it.

Procurement is starting to adapt as well. Ukrainian combat brigades can select products through a digital marketplace, while the U.S. Army has introduced a similar mechanism for approved unmanned systems.

The industrial landscape is changing with them. Anduril, Helsing and Shield AI can raise billions, own more of the product stack and compete directly for programs that younger companies once entered mainly as subcontractors.

Traditional primes will continue to dominate many of the most complex programs. What has changed is the number of categories where they no longer get a free run.

The industry's center of gravity is moving away from building a small number of perfect systems slowly and toward producing enough capable systems quickly enough to matter in an actual war.

Who is the author of this content?

NEW MARKET PITCH TEAM

We track new markets so founders and investors can move faster

We build living "market pitch" documents for emerging markets: AI, synthetic biology, new proteins, and more. Instead of outdated PDFs or hallucinated LLM answers, our clients get a clean, visual, always-updated view of what's really happening: key players, deals, regulations, and signals that matter. Learn more about us.