Which defense tech startups have the largest order books?
In our defense tech market deck, you will find everything you need to understand the market
SUMMARY
Anduril has the largest defense tech startup order book overall, with Helsing, Stark, Quantum Systems and Castelion forming the clearest second tier once we separate real orders from maximum contract ceilings.
The biggest number is not necessarily the biggest backlog. Anduril's $20 billion U.S. enterprise contract is an IDIQ ceiling, Shield AI can compete for work under an $800 million Navy vehicle, and Helsing and Stark each have roughly €1 billion of German options that have not yet become firm orders.
Anduril still comes out on top without counting that $20 billion ceiling. Its roughly $2.2 billion of 2025 revenue, wide mix of active programs and minimum 3,000-missile Barracuda production commitment point to a much deeper contracted business than any other private defense tech company we reviewed.
Helsing probably has Europe's deepest visible physical weapons pipeline. Its German starting order is worth roughly €268 million, while previously announced Ukraine programs cover about 10,000 HF-1 and HX-2 strike drones whose contract values have not been disclosed.
Stark is unusually close to Helsing inside Germany despite being much younger. Both received initial procurements of roughly €268 million, but Helsing still appears to have the larger overall pipeline because of its much bigger disclosed production commitments outside the Bundeswehr.
Quantum Systems has one of the easiest large startup order books to measure. Roughly €210 million of German procurement plus €30.7 million from Romania gives it more than €240 million of identifiable recent orders without relying on speculative unit pricing or giant contract ceilings.
Castelion has made one of the fastest jumps from development work into real weapons production. Three recent Blackbeard Navy awards total about $218 million, while a separate framework points toward a minimum production rate of 500 missiles a year after testing and validation.
Shield AI may be understated by any ranking based only on disclosed dollars. V-BAT has attracted repeat purchases across the United States and several allied countries, but many international quantities and prices remain private, making the breadth of demand easier to see than the size of the firm backlog.
Saronic shows why order book and industrial ambition need to be kept separate. It is building enormous autonomous-vessel manufacturing capacity, but the publicly visible production orders still lag the scale already reached by the drone and missile companies above it.
The most important pattern is that defense startups are crossing from prototype economics into factory economics. Governments are ordering hundreds or thousands of systems, and the next jump in these order books can come from exercising existing options and frameworks rather than announcing entirely new flagship programs.
So the hierarchy is reasonably clear even though exact private-company backlogs are not. Anduril is the strongest overall leader, Helsing has the biggest visible European weapons pipeline, Quantum Systems has unusually clean measurable firm orders, Stark has reached the top group remarkably fast, and Castelion is turning Blackbeard into a serious production program.
This market map, featured in our defense tech market deck, highlights top companies and startups in the defense tech market
What actually counts as a defense tech startup order book?
For defense tech startups, the only useful order-book ranking is one that separates work customers have actually ordered from options and contract ceilings that may never turn into revenue.
That sounds obvious until we look at how defense contracts are announced. The U.S. Army currently has a ten-year Anduril enterprise contract with a maximum value of $20 billion. SAM.gov describes it as an indefinite-delivery, indefinite-quantity agreement covering Anduril software, hardware, data, compute infrastructure and support. Federal agencies can place orders through it, but the government did not buy $20 billion of Anduril products when the agreement was signed.
Germany gives us an even cleaner example. The Bundestag approved initial contracts of roughly €268 million each for Helsing and Stark to supply HX-2 and Virtus loitering munitions. Each contract also carries roughly €1 billion of possible additional purchases if the systems prove themselves and lawmakers approve more spending.
Those are two completely different numbers. For this article, we count signed production orders and funded work as the strongest evidence. Options and IDIQ ceilings tell us how large the relationship could become, so we show them separately rather than pretending they are backlog.
There is another unavoidable problem: private defense companies rarely publish a consolidated backlog. We can reconstruct the visible pieces, but we cannot see every classified order, foreign contract, delivery already completed or dollar of remaining performance obligation. The ranking is strongest when several independent pieces of evidence point in the same direction.
| Company | Publicly visible firm or strongly committed work | Extra capacity we should not count as firm backlog |
|---|---|---|
| Anduril | Multiple large programs, including at least 3,000 Barracuda missiles; full backlog undisclosed | $20B U.S. enterprise IDIQ |
| Helsing | ~€268M German initial order plus large Ukraine drone orders with undisclosed values | ~€1B of additional German options |
| Stark | ~€268M German initial order plus other NATO business | ~€1B of additional German options |
| Quantum Systems | Roughly €240M+ across identifiable recent German and Romanian orders | Options for hundreds of additional German systems |
| Castelion | More than $200M of recent Blackbeard Navy awards, plus earlier development work | 500 missiles a year framework after validation |
| Shield AI | Multiple V-BAT purchases and active U.S. task orders; total unknown | $198.1M Coast Guard IDIQ and access to an $800M Navy pool |
Why are defense tech startup order books getting so big now?
Defense tech startups are moving from prototype contracts into production orders for hundreds or thousands of weapons, which is why order books have become far more meaningful than they were a few years ago.
The change shows up across very different companies. Germany has committed hundreds of millions of euros to young drone makers Helsing and Stark. Quantum Systems received a German program worth roughly €210 million for 520 Falke reconnaissance systems, with options for another 500. Castelion has gone from developing Blackbeard to receiving multiple Navy integration and production awards in the space of months.
Anduril is operating at another scale. The company says 2025 revenue reached $2.2 billion, more than double the previous year, and the Financial Times has reported that management expects revenue to approach $4.3 billion this year. Anduril has also signed a framework covering a minimum of 3,000 surface-launched Barracuda-500M cruise missiles over three years.
Europe is changing particularly fast. Helsing was founded only in 2021, while Stark is even younger, yet Germany is already preparing to buy their weapons by the thousands if initial qualification goes well. That would have been a very unusual procurement path for a European startup a few years ago.
Funding tells us investors expect this shift to continue. Helsing recently raised $1.8 billion at an $18 billion valuation, while Quantum Systems raised about $1.2 billion at an $8 billion valuation. The more interesting development is happening on the customer side: governments are giving these companies orders large enough to require factories rather than larger engineering teams.
As this chart shows, and as featured in our defense tech market deck, search interest in defense tech has risen sharply
Does Anduril have the biggest defense tech startup order book today?
Anduril is the best answer: no other private defense tech company shows the same combination of multi-billion-dollar annual revenue, dozens of government programs and newly committed high-volume weapons production.
We cannot put an exact backlog number on Anduril because the company does not publish one. Even so, its operating scale gives us a much stronger clue than the headline value of any single contract. Anduril generated about $2.2 billion of revenue last year and says its business more than doubled. A company already delivering that much hardware, software and services needs a very large flow of contracted work simply to sustain that revenue base.
The recent Barracuda agreement makes the next step easier to see. According to Anduril's announcement in May, the Department of War framework covers at least 3,000 surface-launched Barracuda-500M systems over three years, with the possibility of buying more. That is a genuine production commitment measured in thousands of missiles rather than a small development award.
Anduril also sells across very different programs: counter-drone systems, Ghost reconnaissance aircraft, Roadrunner interceptors, autonomous underwater vehicles, command-and-control software and collaborative aircraft. Its new federal enterprise contract replaced a procurement landscape that had involved more than 120 separate Anduril purchasing actions.
That breadth is more convincing than one spectacular contract number. Helsing can have a huge drone program and Castelion can have a rapidly growing missile program, but Anduril already has several businesses contributing meaningful government revenue at the same time.
The exact backlog remains private. Still, there is enough evidence to be quite confident about the ranking: Anduril is the most likely number one.
If you want more recent data on this point, please see our latest defense tech market report.
Is Anduril's $20 billion Army deal really a $20 billion order book?
Anduril does not have a proven $20 billion order book from its Army enterprise contract; $20 billion is the maximum amount federal agencies are allowed to order through the vehicle.
SAM.gov describes the agreement as a ten-year, firm-fixed-price IDIQ covering Anduril's full commercial offering. IDIQ stands for indefinite delivery, indefinite quantity. In plain English, the customer creates a purchasing route and then issues individual orders as needs arise.
That makes the agreement commercially powerful. Agencies no longer need to build a new contract every time they want more Anduril software, hardware or support, and the vehicle is available beyond the Army to other authorized federal buyers.
But there is no guarantee that agencies will spend the full ceiling. That's the catch.
We see the same thing with Shield AI. The Coast Guard gave Shield AI a five-year V-BAT services contract with a maximum potential value of $198.1 million. The Navy has also selected Shield AI to compete alongside other suppliers for up to $800 million of ISR task orders. Neither number can be treated as money Shield AI has already booked.
Germany's Helsing and Stark contracts show the difference from the other direction. Lawmakers approved roughly €268 million for each company first, while roughly €1 billion of extra capacity remains conditional. The initial amount belongs much closer to an order-book calculation.
| Contract structure | How we treat it |
|---|---|
| Firm production order | Count it |
| Funded task order | Count it |
| Exercised option | Count it |
| Unexercised option | Keep separate |
| IDIQ maximum | Keep separate |
This chart, included in our defense tech market deck, shows annual VC investment in defense tech startups
Does Helsing have one of Europe's biggest defense startup order books?
Helsing has one of the largest defense startup order books in Europe, and the €268 million German HX-2 contract only captures part of it.
Germany approved roughly €268 million for Helsing's initial HX-2 loitering-munition procurement. Defense News reported that another roughly €1 billion can be added later if the technology meets the required maturity and parliament approves subsequent orders.
Helsing had already built a much larger physical production pipeline before that German award. The company announced 4,000 HF-1 strike drones for Ukraine, followed by another 6,000 HX-2 systems. We therefore know about 10,000 Ukraine-bound strike drones across those two batches, although the financial value was not published.
That undisclosed pricing is important. We would make the ranking less credible if we multiplied 10,000 drones by a guessed unit price and called the result backlog. Military systems include different payloads, software, support, training and manufacturing arrangements, so apparent unit prices can be misleading.
What we can say confidently is that Helsing combines a quarter-billion-euro German starting order with production commitments measured in many thousands of weapons elsewhere. Few European startups show both.
The recent $18 billion valuation makes more sense in that context. Helsing is still priced very aggressively for a young defense company, but its commercial story has moved well beyond pilot contracts.
If you want more recent data on this point, please see our latest defense tech market report.
Is Stark already as big a weapons supplier as Helsing?
Stark has caught Helsing surprisingly quickly in German procurement, but Helsing still appears to have the larger overall order pipeline.
Germany approved roughly the same initial amount for the two companies: about €268 million for Stark's Virtus loitering munition and about €268 million for Helsing's HX-2. Stark says the framework covers several thousand Virtus systems, with deliveries planned within roughly two years after qualification.
Stark has also disclosed another loitering-munition contract with an unnamed NATO member. The company did not publish the customer, quantity or contract value, which prevents us from adding it cleanly to the German figure.
Helsing has much more visible production outside the Bundeswehr. Its previously announced Ukraine orders cover 10,000 strike drones across HF-1 and HX-2. We do not know what those contracts are worth, but the quantity gap is too large to ignore.
Germany itself gives us something close to a draw. The broader order pipeline gives Helsing the edge for now.
That could change quickly. Stark's German contract carries a very large expansion option, and converting even part of it would push a company founded only recently into a production scale normally associated with established missile manufacturers.
This chart, included in our defense tech market deck, shows why Anduril is winning in defense tech
Has Quantum Systems quietly built one of the largest firm drone order books?
Quantum Systems has one of the cleanest large order books we can reconstruct from public data, with more than €240 million of identifiable recent German and Romanian procurement.
The biggest piece is Germany. Procurement documents reviewed by Bloomberg put the Bundeswehr contract at roughly €210 million for 520 Falke reconnaissance systems. The same arrangement includes options for another 500 systems between 2027 and 2032.
The €210 million initial order is unusually large relative to Quantum Systems itself. At the time the contract emerged, Bloomberg calculated that it was equivalent to roughly three-quarters of the company's projected 2025 revenue. For a startup, one customer ordering work worth about nine months of annual sales creates a very substantial production book.
Romania adds another €30.7 million. The Romanian defense minister said the contract covers 34 Vector systems and 15 Scorpion conversion kits, with delivery scheduled for 2027. Quantum Systems already had an earlier Romanian procurement, so this is repeat buying rather than a first trial.
That repeat behavior makes the order quality stronger. Germany is buying hundreds of systems, Romania has come back for more, and Quantum Systems has supplied Ukraine and other military customers as well.
Unlike some companies in this ranking, Quantum Systems does not need a giant IDIQ ceiling to look large. Its strongest publicly visible numbers already represent hardware customers expect to receive.
Is Castelion becoming a serious missile supplier faster than anyone expected?
Castelion is making the fastest jump from defense startup to meaningful U.S. missile producer, with more than $200 million of recent Navy awards around Blackbeard and a much larger production framework sitting behind them.
The sequence matters. In April, Castelion announced a $105 million Navy contract to integrate Blackbeard onto the F/A-18 Super Hornet and push the system toward early operational capability. In June, the Navy placed a $23.4 million firm-fixed-price order covering 50 Blackbeard pre-production weapons and 50 storage and shipping containers.
Then came the latest step. In August, the Navy awarded Castelion another $89,997,162 firm-fixed-price order. The award supports 50 early-operational weapons, manufacturing and logistics work, exportability modifications and the transition toward a formal Navy program of record. Work runs through 2028.
Those three recent awards alone total about $218 million. We should be careful about adding every Blackbeard announcement ever made because some contracts support successive stages of one program and work is already being performed. Still, the trajectory is unusually clear: development, aircraft integration, pre-production and early operational production are happening one after another.
Castelion's May framework is potentially much bigger. The company says the Department of War has committed to a guaranteed minimum production rate of 500 Blackbeard missiles per year once testing and validation are complete, with a path toward thousands more.
We cannot attach a dollar backlog to that production commitment yet because no total price has been disclosed. But 500 missiles per year would put Castelion in a very different category from a startup making a few dozen prototypes.
| Recent Castelion Blackbeard award | Value | What the Navy is buying |
|---|---|---|
| F/A-18 integration | $105M | Testing, certification and integration |
| Pre-production order | $23.4M | 50 weapons plus containers |
| Early operational order | ~$90M | 50 weapons plus transition work |
| Production framework | Undisclosed | Minimum 500 missiles/year after validation |
If you want more recent data on this point, please see our latest defense tech market report.
This chart, included in our defense tech market deck, shows annual funding in defense tech startups
Does Shield AI have a bigger order book than the public numbers show?
Shield AI probably has a much larger real order book than we can calculate publicly because V-BAT keeps winning repeat international customers while many contract values remain undisclosed.
The easy mistake is to start with the biggest U.S. numbers. The Coast Guard's $198.1 million V-BAT agreement is an IDIQ running through 2029, so actual services are ordered progressively. The Navy's newer $800 million ISR vehicle is even less useful as backlog because Shield AI must compete with other selected companies for task orders underneath it.
The international customer list tells us more about the health of the business. The Netherlands originally procured V-BATs and now has 12 being acquired for Royal Netherlands Navy operations. Greece announced another V-BAT purchase this year after already operating the system. Poland's Armament Agency has signed a contract for V-BAT aircraft for Polish Navy maritime surveillance. Japan has selected the platform as well.
Greece's latest agreement also includes multi-year service, training and maintenance work, according to Janes. That makes the revenue stream more valuable than a one-off drone shipment because the customer remains tied to the platform after delivery.
The frustrating part is that Greece and Poland did not disclose quantities or contract values, while several other foreign awards are similarly opaque. Shield AI could therefore have a larger firm book than Quantum Systems on an economic basis and still look smaller in a public-data ranking.
For that reason, we put Shield AI in the top group for order-book quality but avoid giving it a fake aggregate number.
Is Saronic's huge valuation backed by a huge Navy order book yet?
Saronic still has far more manufacturing ambition than publicly disclosed production backlog, so it does not belong near the top of this ranking yet.
That distinction is useful because Saronic is one of the most valuable private defense companies and is investing heavily in autonomous ship manufacturing. Port Alpha is designed to give the company the ability to produce vessels at a scale far beyond traditional prototype work.
The publicly visible Navy business remains much smaller. Federal procurement records have shown development and testing contracts in the tens of millions rather than the hundreds of millions or billions we can associate with the leaders above.
There is nothing inherently wrong with building capacity ahead of demand. In fact, defense customers increasingly complain that traditional suppliers cannot expand production quickly enough. Saronic may be trying to solve that problem before the Navy places its biggest autonomous-vessel orders.
But an order-book ranking has to reward signed customer demand rather than industrial potential. Helsing has thousands of drones to deliver, Quantum Systems has hundreds of reconnaissance systems under major contracts, and Castelion has started receiving firm missile-production orders.
Saronic could jump several places with one large fleet procurement. For now, it remains a capacity story more than a backlog story.
If you want more recent data on this point, please see our latest defense tech market report.
This chart, included in our defense tech market deck, compares the main business model options for defense AI contractors
Which defense tech startups have the safest order books?
Anduril, Quantum Systems and Shield AI have the strongest order-book quality because their demand is spread across multiple customers or programs rather than depending on one enormous procurement decision.
Anduril is the clearest example. The company sells autonomous aircraft, counter-drone systems, missiles, underwater vehicles and command software. The U.S. government is the dominant customer ecosystem, but individual programs sit across different military branches and missions, while Anduril also sells internationally.
Quantum Systems has a narrower product family but increasingly broad customer exposure. Germany's large order sits alongside repeat Romanian procurement, Ukrainian deployments and other international programs. If one customer slows down, the whole company does not suddenly lose its only path to production.
Shield AI is even more geographically diversified. V-BAT is used or being bought by a growing list of U.S. and allied military customers, including the Coast Guard, the Netherlands, Greece and Poland. Its public backlog is harder to reconstruct, but repeat procurement across several countries lowers concentration risk.
Helsing and Stark carry more program concentration. Helsing's scale is impressive, although a large share of its visible hardware pipeline comes from Ukraine and Germany. Stark is even more dependent on successful Virtus qualification and German follow-on buying.
Castelion has the most obvious product concentration. Blackbeard could turn the company into a major missile supplier, but one weapon family still drives most of the visible commercial story.
A large order can make a startup grow very quickly. The safer order book comes from customers returning for more across several programs.
Could the biggest defense startup order books double without any new flagship contract?
Yes. Existing options and framework agreements are already large enough to add billions of dollars of new orders if current systems perform well.
Helsing and Stark each have around €1 billion of German options beyond their initial contracts. Quantum Systems can supply another 500 Falke systems under the optional part of its German program. Castelion's production framework can move from the current early-operational quantities to at least 500 Blackbeard missiles a year after successful testing and validation.
Anduril has even more room to expand because federal agencies can order its full product suite through the enterprise IDIQ. As pointed out above, the $20 billion ceiling itself is not backlog, but it gives existing customers an unusually large and simple route for placing more orders.
Shield AI has the same mechanism on a smaller scale. The Coast Guard can issue additional orders under its existing V-BAT vehicle, while Navy ISR task orders can flow through the newer competitive contract pool.
This is why the next stage of the ranking depends heavily on execution. Helsing and Stark need successful qualification. Castelion needs Blackbeard to clear validation and operational testing. Quantum Systems has to turn a 520-system program into repeat German purchases. Shield AI needs international selections to keep converting into larger fleets.
The companies that do that can see their order books grow sharply without another dramatic procurement announcement.
This chart, featured in our defense tech market deck, shows the share of revenue generated by each customer segment in the defense tech market
Which defense tech startups really have the largest order books today?
Anduril has the largest defense tech startup order book overall, while Helsing, Stark, Quantum Systems and Castelion form the clearest second tier once we focus on actual orders instead of maximum contract values.
Anduril comes first because its scale is already visible in the business. It produced about $2.2 billion of revenue last year, operates across many major defense programs and now has a minimum 3,000-missile Barracuda production commitment layered onto that existing base. Its exact backlog remains private, but no peer gives us enough evidence to justify putting another startup ahead.
Helsing probably comes next if we care about physical weapons volume. Germany has approved an initial order worth roughly €268 million, while its previously announced Ukraine programs cover thousands more strike drones with undisclosed contract values.
Stark has reached almost the same German starting point astonishingly quickly. Its roughly €268 million Virtus procurement gives it one of Europe's largest clearly disclosed startup weapons orders, although its broader book appears smaller than Helsing's for now.
Quantum Systems may deserve the highest confidence score of the European group if we insist on measurable firm orders. Roughly €210 million from Germany plus €30.7 million from Romania gives us more than €240 million of recent procurement that is relatively easy to identify, with physical deliveries attached.
Castelion is the company moving fastest up the U.S. ranking. Its three recent Blackbeard Navy awards total about $218 million, and the latest $90 million order takes the missile into early operational production. The 500-missile-per-year framework could eventually push Castelion far above that disclosed amount.
Shield AI belongs nearby, but public disclosure works against it. Customers keep buying V-BAT, yet the company rarely gives us enough pricing information to reconstruct the combined firm value. We therefore know more about the breadth of Shield AI's demand than its exact size.
Saronic sits lower because the factory ambition has arrived before the large public production orders.
The ranking gets much clearer once we stop treating every billion-dollar procurement ceiling as booked business. Anduril is the strongest overall leader. Helsing probably has Europe's deepest physical weapons pipeline. Quantum Systems has one of the cleanest large firm order books we can actually measure. Stark has reached that group at remarkable speed, and Castelion is turning hypersonic-missile development into real production orders.
Those five companies are the names to watch if the question is who already has serious defense work to deliver, rather than who merely has access to the biggest possible contract.
If you want more recent data on this point, please see our latest defense tech market report.
OUR METHODOLOGY
The size of a private defense company's order book is rarely available as one clean number. We therefore compared several signals together: identifiable firm awards, committed production quantities, how far programs have progressed from development into delivery, repeat customer demand, customer breadth, and additional procurement already embedded in existing frameworks.
We treated funded awards, signed production orders and exercised work as the strongest evidence. Unexercised options, IDIQ ceilings and competitive contract pools were kept separate because they show how large a relationship could become without proving that the work has actually been ordered. When governments disclosed quantities but not prices, we used the quantities as evidence of production scale rather than inventing a contract value from an assumed unit price.
The final ranking comes from aggregating those signals rather than relying on a single headline contract. That is important because private defense companies disclose very different pieces of the picture: one may disclose precise dollar awards, another may have thousands of systems in production without publishing the contract value, while another may have access to a huge procurement vehicle from which only a small part has actually been ordered.
We prioritized government procurement records, parliamentary approvals, official customer announcements, company contract disclosures and high-quality financial or defense reporting. Key sources include the U.S. Army on Anduril's enterprise procurement contract, Anduril's Barracuda-500M production announcement, the Financial Times on Anduril's operating scale, Defense News on the German Helsing and Stark procurements, Helsing on its Ukraine strike-drone production, and STARK on the Bundeswehr Virtus agreement.
Other important sources include Romania's Ministry of National Defence on the Quantum Systems procurement, Quantum Systems on its latest financing and commercial expansion, Castelion on the Blackbeard production framework, Castelion on its $90 million Navy award, the U.S. Coast Guard on the V-BAT IDIQ, and Shield AI on the Navy ISR contract vehicle.
For Shield AI's international demand, we also used the company's disclosures covering the Royal Netherlands Navy, Greece's repeat V-BAT procurement, and Polish naval procurement. For industrial-capacity context we used Saronic's disclosures on its autonomous-vessel manufacturing plans, while Helsing's financing announcement provides context for the scale investors are assigning to its expanding defense business.
This chart, included in our defense tech market deck, shows how tactical networking platform technology has evolved over time
Related blog posts
- Which defense startup has the largest contract backlog?
- Which defense tech startups have won the most contracts?
- Which defense tech startups generate the most revenue today?
