Is it already too late for defense tech?

What does “too late for defense tech” really mean?

Defense tech is still early as an industrial cycle, although the easy entry window has already closed for generic drone, autonomy and “AI for defense” startups.

The question depends on who is asking. A founder wants to know whether the best problems have already been claimed. An investor wants to know whether valuations have moved too far ahead of revenue. A government supplier wants to know whether new companies can still break into programs dominated by established contractors.

We use a practical definition. It would be “too late” if military demand were close to peaking, the leading startups had locked up most new programs, or the remaining opportunities required so much capital and political access that a newcomer could barely compete.

None of those conditions fully applies today. Demand is still climbing, procurement is slowly opening and several major capability gaps remain unresolved. Yet the standard for entering has changed. Spotting that militaries need more drones or better software no longer counts as an insight.

The market now rewards companies that can solve one hard operational problem, survive a long procurement process and manufacture at a useful scale.

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

Why does defense tech feel so crowded now?

Defense tech feels crowded because investor money and company valuations have risen much faster than the number of startups reaching large-scale military production.

According to CB Insights data reported by Defense News, defense-tech equity funding climbed from $7.3 billion in 2024 to $17.9 billion in 2025. Overall startup funding also recovered, but defense tech grew more than three times faster.

The pace has stayed intense this year. The Financial Times counted $12.3 billion of defense venture investment during the first half alone, nearly twice the comparable period. A broader Wall Street Journal tally reached $16.8 billion. The difference comes from how each database defines defense and dual-use technology, yet both point in the same direction.

The largest rounds show how concentrated the excitement has become. Anduril recently raised $5 billion at a $61 billion valuation. Helsing followed with $1.8 billion at $18 billion. Shield AI assembled $2 billion at $12.7 billion, Saronic raised $1.75 billion at $9.25 billion, and Quantum Systems raised $1.2 billion at roughly $8 billion.

We count $11.75 billion raised by those five companies within a few months. Their combined valuation is about $109 billion. Investors are clearly backing a small group to become a new generation of prime contractors rather than ordinary equipment suppliers.

That concentration makes the sector look more mature than it really is. Most defense startups remain small, pre-production and dependent on pilots or research contracts. A handful of enormous private companies sit above a much less proven market.

Recent company	New capital	Latest valuation
Anduril	$5.0B	$61.0B
Helsing	$1.8B	$18.0B
Shield AI	$2.0B	$12.7B
Saronic	$1.75B	$9.25B
Quantum Systems	$1.2B	About $8.0B
Combined	$11.75B	About $109B

Has the defense spending boom already peaked?

The defense spending boom has slowed from its first surge, but the underlying rearmament cycle is still expanding today.

SIPRI's latest estimate places worldwide military expenditure at $2.887 trillion in 2025. That was the eleventh consecutive annual increase and 41% above the level recorded ten years earlier.

Real growth slowed to 2.9%, compared with 9.7% in 2024. At first glance, that looks like a peak. The regional data says otherwise. European spending rose by 14%, while Asia and Oceania increased expenditure by 8.1%. The global slowdown largely came from reduced US military-aid spending rather than a widespread reversal.

Long-term political commitments also keep moving higher. NATO members have agreed to work toward spending 5% of GDP by 2035, including 3.5% for core defense and 1.5% for wider security and resilience.

Europe has already moved beyond plans. The European Defence Agency's newest figures show that the EU's 27 members spent €418 billion in 2025, 20% more than in 2024. The agency currently expects the total to reach €454 billion in 2026. Equipment procurement alone reached €115 billion last year.

Governments can still delay programs, cut individual weapons or miss their NATO targets. Even so, a broad collapse in military demand would now require several regions to reverse course simultaneously. The evidence points toward years of uneven expansion.

Is the $2.9 trillion defense budget really a startup market?

No. Defense startups can realistically compete for only a narrow slice of the world's $2.9 trillion military budget.

The headline figure includes salaries, pensions, bases, training, fuel, maintenance, nuclear forces, military aid and decades-old weapons programs. Most of that money cannot suddenly move into autonomous systems or startup software.

Europe gives us a clearer sense of the available pool. Its members spent €418 billion in 2025, but equipment procurement represented €115 billion. Defense investment, including research and equipment, accounted for just over 32% of total expenditure.

The startup-accessible amount is smaller again. Established contractors control most fighter, submarine, missile, vehicle and sustainment programs. National governments also protect domestic suppliers and often require years of certification before approving a new vendor.

The same gap appears in the United States. The Wall Street Journal recently found that Pentagon spending with the 15 largest defense-tech startups had tripled since 2022. Those companies still collected less than 1% of total defense contract spending.

That tiny share leaves plenty of room for growth, but it also exposes inflated market-size claims. A startup cannot value itself by taking one percentage point of the whole Pentagon budget unless it can identify the programs, production quantities and purchasing authority behind that percentage.

The useful market measure is funded procurement that a new supplier can genuinely win. Anything broader quickly becomes fantasy arithmetic.

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

Is private money getting ahead of military orders?

Private capital is now financing defense companies faster than governments are turning their products into dependable, recurring orders.

The Defense Innovation Unit offers one of the clearest long-term comparisons. Since 2016, it has awarded 450 prototype contracts and directly helped 62 reach follow-on contracts. Those transitions carried potential contract ceilings of $5.5 billion, supported by $19.1 billion in private capital.

Those ceilings are maximum values, rather than guaranteed revenue. The comparison also excludes contracts won through other channels. Still, it shows why startup financing matters so much in defense: private investors often fund the years between a promising prototype and a serious production order.

Those years remain difficult. The US Government Accountability Office says major Pentagon programs now take more than 12 years on average to provide an initial capability. Faster pathways exist, but GAO continues to find immature technologies, delayed decisions and programs that finish without fielding something useful.

Hardware companies face the sharpest version of this problem. They need test facilities, specialized employees, restricted components and production equipment well before the customer commits to buying thousands of units.

These days, investors will fund that gap for companies with strong teams and credible military demand. Anduril, Helsing, Saronic and Quantum Systems are using fresh capital to build factories, shipyards and multinational production networks.

A newcomer cannot assume the same generosity. Its first factory plan must connect to a real purchasing authority, an operational user and a believable production schedule. Otherwise, venture capital merely postpones the procurement problem.

Have Anduril, Helsing and Shield AI already locked up the market?

Anduril, Helsing, Shield AI and the other leaders have built a formidable head start, but the defense-tech market remains too broad and politically fragmented for them to control it all.

Their advantage now extends far beyond technology. They have government relationships, security-cleared teams, operational data, factories, export experience and enough capital to survive procurement delays.

Anduril reported $2.2 billion of revenue in 2025, more than twice its previous year, and is building a five-million-square-foot manufacturing campus in Ohio. Its portfolio stretches across command software, autonomous aircraft, missiles, underwater vehicles, air defense and rocket motors.

Helsing has moved from defense AI into strike drones, electronic warfare, underwater systems, robotics and tactical space. It operates manufacturing facilities in several European countries and recently selected West Virginia for its first US factory.

Shield AI combines autonomy software, simulation and its own aircraft. Saronic is moving from small autonomous boats into large vessels and an automated shipyard. Quantum Systems is building a multinational production network across aerial, ground and maritime autonomy.

A founder entering one of those companies' core categories now faces more than a product competitor. The rival can offer integration, financing, production and political reassurance in one package.

Yet governments rarely want one private supplier controlling an entire class of military capability. They need backup vendors, competing architectures, local manufacturing and bargaining power.

The US Air Force recently demonstrated this preference by selecting Anduril's aircraft while separately awarding the autonomy software to Shield AI and other suppliers. The aircraft and its intelligence can evolve independently.

That kind of modular procurement keeps parts of the market open. The best opening for a newcomer often sits inside a larger system where governments actively want more than one supplier.

Are defense startups finally winning real production contracts?

Yes. Several defense startups are now winning production work that would have been almost unimaginable for young companies a decade ago.

The clearest example is the US Air Force's Collaborative Combat Aircraft program. Anduril's FQ-44 moved from a prototype award in 2024 to a production decision just over two years later. The company describes it as the first semi-autonomous fighter aircraft to enter serial production.

Shield AI received a separate production award for the Hivemind mission-autonomy software flying on the aircraft. That makes the software a purchased military capability in its own right, rather than a demonstration attached to an airframe.

Missiles are following a similar path. A recent agreement between Anduril and the US Defense Department is intended to support the delivery of thousands of air-launched Barracuda-500 systems each year for seven years, beginning in 2027.

Ukraine is generating a different kind of order at even larger unit volumes. Quantum Systems has been tasked with supplying 15,000 interceptor drones to the Ukrainian National Guard. Helsing received an order for another 6,000 HX-2 strike drones and says its first German factory can produce more than 1,000 units monthly.

Saronic has meanwhile moved beyond prototype boats. Its Corsair recently completed what the company describes as the first known rescue of downed aircrew by an autonomous vessel during combat operations. The company is now preparing a multibillion-dollar automated shipyard in Texas.

These contracts show that the startup model works across aircraft, software, missiles, drones and maritime systems. The winners, though, keep coming from the same well-funded group.

Production has arrived, but selectively.

Is the Pentagon genuinely open to newcomers now?

The Pentagon is currently more open to commercial defense companies, although its buying system remains slow enough to destroy an undercapitalized startup.

One visible change is the US Army's new Unmanned Aircraft Systems Marketplace. Authorized users can browse, compare and procure vetted drones through a centralized online storefront.

The Army and allied partners are extending the same idea to counter-drone technology. The planned marketplace covers interceptors, radar, sensors, electronic warfare and passive defenses. Buyers that once spent years running separate evaluations can now draw from a shared catalogue of tested systems.

Broader acquisition reforms point in the same direction. The Pentagon has instructed buyers to favor commercial solutions, modular architectures, direct supplier relationships and faster software pathways. It has also made the Defense Innovation Unit its main interface for adopting commercial technology.

The system has changed less than the announcements suggest. Congressional funding cycles, testing rules, security approvals and program offices still slow purchases. GAO's latest assessment found that the Pentagon continues to struggle with reforms already available to it.

For founders, the practical improvement is real but narrower than the rhetoric. It has become easier to obtain a test, prototype award or first military user. Converting that opening into predictable production remains the harder part.

A good company should know who can fund the next contract before celebrating the current one.

Is Europe now the easier defense-tech market?

Europe offers more defense-tech opportunities than it did a few years ago, but selling across Europe remains harder than the headline spending boom suggests.

EU defense expenditure reached €418 billion in 2025 and is expected to rise again. The SAFE program can provide up to €150 billion in loans for common procurement, while the wider Readiness 2030 plan aims to unlock as much as €800 billion of additional spending capacity.

Startup capital is also arriving. Dealroom and the NATO Innovation Fund counted a record $8.7 billion invested in European defense, security and resilience startups during 2025.

The difficult part is converting continental demand into one repeatable sales process. France, Germany, Poland, Italy and the Nordic countries buy differently, favor different national suppliers and often want production or intellectual property located inside their borders.

Only 24% of EU equipment procurement was collaborative in 2025. That figure has improved, but it still means that most purchases are being made through national systems.

European industrial rules add another filter. Major EU programs increasingly require local facilities, European design control or a high proportion of components sourced from the EU and associated countries.

This favors startups that embed themselves locally. Helsing has factories in Germany and the United Kingdom. Quantum Systems produces across several allied countries and has established German-Ukrainian joint ventures. Anduril is working with Rheinmetall and Poland's PGZ on European production.

Europe is currently an attractive market for companies willing to become European suppliers. A startup hoping to export one unchanged American product into 27 countries will find the continent far less welcoming.

Did Ukraine permanently change what militaries buy?

Ukraine has permanently changed how militaries judge drones, autonomy, electronic warfare and production speed.

The most important change is the feedback loop. Ukrainian units can identify a battlefield problem, test several products, report what failed and push an improved version back to the front within weeks.

Brave1's latest operational data shows the scale of that process. Russia launched more than 10,000 attack drones during one recent month, while the number destroyed rose by nearly 50%. Brave1 says interceptor drones played a decisive role.

One Ukrainian system has already automated 95% of the interception process, covering the flight from launch to the final attack. Ground robots have also completed more than 14,000 frontline logistics and evacuation missions.

These are large operational samples, rather than carefully prepared demonstrations. They show how products behave under jamming, bad weather, exhausted crews, equipment losses and constant enemy adaptation.

The industrial response is now spreading beyond Ukraine. Quantum Systems and its Ukrainian partners are building drones in Germany, with capacity planned for 10,000 units within a year. Helsing is producing Ukrainian-bound strike drones through distributed factories. Other European governments are creating programs that combine Ukrainian designs with Western capital and manufacturing.

Militaries have learned that battlefield performance depends on more than range, speed or payload. The product must be easy to update, cheap enough to lose and simple enough to manufacture repeatedly.

Peacetime buyers will still require certification, cybersecurity and interoperability. Those requirements remain unavoidable. Ukraine has nevertheless raised the burden of proof for anyone selling an expensive system that cannot adapt quickly.

Are military drones already too crowded?

The basic military-drone market is crowded now, especially for companies offering familiar airframes with minor improvements.

A small drone can be assembled from widely available motors, batteries, cameras and flight-control components. That has allowed hundreds of teams to produce working prototypes and convincing videos.

Military customers have become harder to impress. Ukraine has exposed how quickly drones lose effectiveness once an enemy learns their frequencies, identifies their navigation method or develops a cheap countermeasure.

The US Army's decision to use a marketplace also tells us where the category is going. Buyers expect to compare multiple models, switch suppliers and refresh their fleets frequently. That weakens the long-term advantage of owning one conventional airframe.

The valuable layer is moving toward what lets the drone complete its mission: navigation without GPS, autonomy under communications jamming, target recognition, coordinated flight, propulsion, sensors, secure networks and inexpensive interception.

Manufacturing also offers room for differentiation. A company capable of producing 10,000 dependable units with traceable allied components may beat a technically better design that can only be built by hand.

Founders should therefore start with the military problem. Beginning with “we are building a drone” places the company inside one of defense tech's most crowded queues.

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

Where is defense tech still wide open?

Defense tech remains wide open in areas where militaries have urgent shortages and today's suppliers cannot provide enough affordable, deployable equipment.

Air defense is the most obvious case. Militaries are trying to stop cheap drones and cruise missiles without spending millions of dollars on every interception. Ukraine's order for 15,000 interceptor drones and the recent emergence of new low-cost air-defense companies show that the market can support several approaches.

Electronic warfare and resilient navigation are another deep gap. Drones, vehicles and soldiers need to operate when GPS, communications and cloud connections disappear. Technologies that work only inside a clean demonstration environment will struggle here.

Maritime autonomy is earlier than aerial drones. Navies want more distributed surveillance, mine warfare, logistics and strike capacity, while conventional shipyards remain slow and overloaded.

Software still matters, particularly when it controls a physical outcome. Mission autonomy, sensor fusion, targeting, logistics and command systems can become large businesses when they fit directly into a funded military workflow. A generic AI dashboard has a much weaker position.

Finally, factories have become part of the opportunity. Militaries need suppliers of propulsion, energetics, seekers, sensors, testing equipment, robotics, secure components and maintenance systems. Those companies may attract less public attention than a new aircraft, but every large production plan depends on them.

Open area	Why buyers still have a problem	Strong entry point
Counter-drone and affordable air defense	Existing interceptors are too expensive for mass drone attacks	Cheap detection-to-intercept systems
Resilient navigation and communications	GPS and radio links fail under electronic attack	Jam-resistant hardware, networks and autonomy
Maritime and undersea autonomy	Navies lack enough ships, sensors and crews	Autonomous vessels, payloads and navigation
Affordable missiles and propulsion	Inventories are difficult and expensive to replenish	Motors, seekers, guidance and scalable weapons
Mission software and sensor fusion	Data exists, but operators struggle to act on it quickly	Software tied to a funded operational decision
Manufacturing and sustainment	Prototypes reach production bottlenecks	Robotics, testing, components and maintenance

Does geography matter almost as much as the technology?

Geography now has a major influence on who can sell defense technology, where the product can be built and which government will fund it.

Defense buyers care about what happens after a supplier's home country changes policy, blocks an export or stops providing spare parts. Local production gives them more control over inventory, upgrades and wartime replenishment.

European programs increasingly formalize that preference. Under the European Defence Industry Programme, at least 65% of component costs must originate in the EU or associated countries. Eligible projects also need European facilities and, in many cases, European control over the design.

The United States has its own restrictions. Export-controlled technology, classified data and sensitive components cannot freely move between engineering teams or customers.

Leading startups are adapting their company structures around these rules. Helsing, a European company, is building a US factory. Anduril is partnering with national industrial groups in Germany and Poland. Quantum Systems combines local subsidiaries with joint production. Saronic has facilities in the United States, United Kingdom and Australia.

For a new defense company, the choice of headquarters can shape the market more strongly than taxes or talent costs. It affects security clearances, export rights, investor eligibility, manufacturing subsidies and the nationality of customers who will trust the company.

Defense tech may use global science, but the final product is usually sold through national politics.

Can traditional defense companies simply copy or buy the winners?

Traditional defense contractors can copy features and acquire startups, but rebuilding the whole startup operating model inside a large incumbent is much harder.

The major contractors still hold most of the money. They operate classified programs, maintain global supply chains and support equipment for decades. A newcomer rarely matches that institutional weight.

Startups have developed a different advantage. They use private money to build before the government fully commits, hire heavily from commercial technology companies and update products faster than traditional programs allow.

The two groups are increasingly working together. Saab has invested in Helsing. Airbus participated in Quantum Systems' latest round. L3Harris and Hanwha have invested in Shield AI. Rheinmetall and Anduril are collaborating on autonomous aircraft and weapons for Europe.

These partnerships allow established companies to gain technology while startups gain production access and political credibility. In some cases, an acquisition will be the logical outcome.

The strongest startups are harder to absorb. Anduril's $61 billion valuation already exceeds the market value of several public defense contractors. Helsing and Shield AI are also becoming expensive enough that only a few buyers could realistically acquire them.

We expect a mixed market. Some startups will become component suppliers, others will be bought, and a small group will grow into new prime contractors.

The dangerous position is the middle: too narrow to win major programs alone, yet too expensive for an incumbent to acquire comfortably.

Is defense tech already a bubble?

Defense tech contains a genuine valuation bubble, even though the underlying demand and technological shift are real.

The strongest evidence is the price investors are paying for future revenue. The Financial Times recently estimated that Helsing's $18 billion valuation equals about 32 times its projected 2026 revenue. The same analysis placed Shield AI near 21 times, Anduril around 13 times and Quantum Systems at roughly 8.5 times.

Traditional defense contractors commonly trade much closer to two or three times annual sales. Startups deserve higher multiples when they grow faster, but today's gap assumes that several private companies will simultaneously become enormous global suppliers.

Investor behavior also looks crowded. Defense venture funding has jumped sharply, mega-rounds are appearing within weeks of one another and new companies are clustering around drones, autonomy and AI-enabled weapons.

The fundamentals stop us from dismissing the whole sector as speculative fiction. Military expenditure has risen for 11 consecutive years. Europe is expanding equipment procurement. NATO has adopted higher long-term targets. Startups have won production contracts for combat aircraft, autonomy software, missiles and thousands of drones.

The sector looks closer to the internet during a capital boom than to a market built on nothing. The technological change will survive. A meaningful number of companies and valuations probably will not.

Bubble evidence	Evidence of a real market
Private valuations reach roughly 8.5–32 times projected sales	Global military spending has risen for 11 consecutive years
Mega-rounds are concentrated in a few fashionable categories	Startups have entered real aircraft, missile and drone production
Private capital is growing faster than startup procurement	NATO and European spending commitments extend into the next decade
Many companies offer similar drone and autonomy stories	Air defense, electronic warfare and production capacity remain undersupplied
Investors price future prime contractors before most revenue arrives	Military users are deploying startup products in active operations

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

Who is already too late for defense tech?

Founders offering a generic drone, a broad military AI platform or a defense version of an existing commercial tool are already late.

Investors and military buyers have seen those pitches repeatedly. New entrants now need a precise answer to four questions: which operational problem they solve, why existing suppliers fail, who can buy the product and how it reaches meaningful production.

Founders without access to military users face another problem. A technically strong team can spend years optimizing the wrong requirement because battlefield constraints rarely appear in a procurement document or investor presentation.

Companies dependent on one government innovation office are also vulnerable. A prototype program can provide money and credibility, but the next contract often belongs to a different budget holder with a different timeline.

Late-stage investors face their own version of “too late.” Anduril, Helsing, Shield AI, Saronic and Quantum Systems may become very large companies, but their present valuations already assume major success. The potential return is less forgiving than it was for investors who entered several rounds earlier.

The market remains attractive for teams with unusual technical knowledge, privileged operational access, a difficult component or a credible way to manufacture at scale.

These founders are entering a tougher market, but they are not arriving after the opportunity.

So, is it already too late for defense tech?

No. It is not too late for defense tech, although the period when a fashionable story and a polished prototype could carry a company has largely ended.

The demand side remains powerful. Global military spending is at a record level, European equipment procurement is increasing, NATO countries have adopted higher targets and militaries are trying to rebuild depleted industrial capacity.

The technology side has also moved beyond presentations. Autonomous fighter aircraft are entering production, AI pilots are receiving separate contracts, low-cost missiles are being ordered in planned volumes of thousands and interceptor drones are becoming a major part of air defense.

At the same time, competition has become much harsher. The leading private companies can spend billions, build factories and operate across several countries. Drone airframes are easier to copy. Governments want local production. Procurement still takes too long, and startup valuations already price in years of future growth.

Defense tech remains early in deployment but has become late in attention.

The first wave rewarded founders and investors who understood that software, autonomy and commercial engineering could reshape military systems. That observation is common knowledge now.

The next wave will be won by companies that solve a narrow but urgent capability gap, prove the product under difficult conditions and then build it in serious quantities.

For anyone arriving with only a defense-tech narrative, the opportunity has passed. For teams that can actually deliver missing military capability, the market still has years to run.

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

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.