The $28.67 Billion AI Weapons Market: Who Wins, Who Loses

In 2022, the global market for artificial intelligence in defense and military applications was valued at approximately $8.6 billion. By 2030, according to the most widely cited market research across Morgan Stanley, Goldman Sachs, and specialist defense analysis firms, that figure will reach $28.67 billion. The compound annual growth rate of 20.1% makes defense AI one of the fastest-growing segments in the entire technology sector — growing faster than commercial cloud infrastructure, enterprise SaaS, or consumer AI applications during the same period.

The numbers are compelling on their face. But they obscure a more important and more investable story: this market is in the middle of a structural disruption. The companies that have historically dominated defense procurement — Lockheed Martin, Raytheon, Northrop Grumman, Boeing, General Dynamics — are not losing their defense contracts. They are losing their growth premium. The fastest-growing slice of defense spending is flowing to a different category of company, one built around software rather than hardware, platform agnosticism rather than proprietary systems, and venture capital rather than the Defense Export-Import ecosystem.

Understanding where the money flows in defense AI, and which companies are positioned to capture it, requires working through four layers of analysis: the overall market structure and its segment dynamics, the regional distribution of spending, the competitive landscape between legacy primes and startup challengers, and finally the investment instruments — ETFs and individual equities — through which capital markets exposure to this growth can be accessed.

Market Segments: Where the Money Actually Goes

The $28.67 billion total obscures significant variation across functional segments. Defense AI is not one market; it is five overlapping markets with different growth drivers, different competitive dynamics, and different risk profiles. The segment breakdown by 2030 projected revenue tells a story about where AI delivers the most defensible value in military contexts.

Autonomous Vehicles and Drones: 35% ($10.0 Billion)

The largest single segment of the defense AI market is autonomous vehicles — aerial, ground, maritime, and undersea — and the AI systems that enable their autonomous operation. By 2030, this segment will account for approximately $10 billion in annual revenue, driven by the U.S. Replicator initiative, European drone fleet expansion, and the continued scaling of commercial-off-the-shelf autonomous platforms adapted for military use.

This segment's growth is driven by several converging factors. Unit costs for autonomous aerial platforms have fallen by approximately 70% since 2018, driven by commercial drone manufacturing scale in China and the entry of venture-backed American startups using commercial supply chains. The combination of cheaper platforms and better AI has enabled doctrine shifts that are increasing the number of autonomous vehicles purchased: a single F-35 costs $82 million; a Kratos XQ-58 Valkyrie loyal wingman costs $3 million; an Anduril Fury autonomous strike platform is designed to be attritable at even lower cost. Force multiplier economics are driving volume.

The primary companies capturing this segment include Anduril Industries (Ghost, Fury, ALTIUS systems), Shield AI (HIVEMIND-equipped autonomous platforms), Kratos Defense (XQ-58, UTAP-22 Mako), Boeing (MQ-28 Ghost Bat, Orca XLUUV), and Northrop Grumman (MQ-8 Fire Scout, Manta Ray, Firebird). International competitors include Baykar Technology (TB2, Akinci), Israel Aerospace Industries (Harop, Harpy 2), and China's DJI-derived military platforms.

C4ISR and Intelligence Systems: 25% ($7.2 Billion)

Command, control, communications, computers, intelligence, surveillance, and reconnaissance — the sprawling acronym category that encompasses the cognitive backbone of modern military operations — is the second-largest defense AI segment. By 2030, AI-enhanced C4ISR will account for approximately $7.2 billion in annual revenue, driven by programs like Project Maven, the Joint Warfighting Cloud Capability (JWCC), and the emerging AI-enabled command-and-control platforms across NATO allies.

Palantir Technologies is the defining company in this segment. Its Maven Smart System, built on the Palantir AI Platform (PAP), processes satellite imagery, drone footage, and signals intelligence at scale for the U.S. military. In 2025, Palantir converted its Maven relationship from a project to a permanent program, securing a multi-year contract worth over $900 million. The system now processes an estimated 4 million hours of drone footage monthly, enabling the kind of persistent pattern-of-life analysis that was previously possible only with dozens of human intelligence analysts working on a single target.

Microsoft and Amazon are the infrastructure layer on which much of the C4ISR AI market runs. The JWCC contract, awarded to both companies along with Oracle and Google in 2022, is valued at up to $9 billion over seven years and provides the cloud foundation for DoD AI systems. Leidos, Booz Allen Hamilton, and SAIC are the primary systems integrators that translate cloud infrastructure and AI platforms into deployed military capabilities.

Cyber AI: 20% ($5.7 Billion)

Artificial intelligence is transforming both offensive and defensive cyber operations. AI-enabled tools can scan networks for vulnerabilities at speeds impossible for human analysts, automate the development and deployment of malware, detect intrusions through behavioral anomaly analysis rather than signature matching, and conduct influence operations at scale through synthetic content generation. By 2030, AI-specific cyber capabilities will represent a $5.7 billion segment of the defense market.

The key players in defense cyber AI include CrowdStrike, whose Falcon AI platform has extended from enterprise security into defense-specific applications; Darktrace, which uses unsupervised machine learning for network anomaly detection; IronNet Cybersecurity; and classified programs operated by NSA's Cybersecurity Directorate and Cyber Command's Cyber National Mission Force. Specialist defense cyber AI startups include Rebellion Defense, which was acquired by Shield AI, and Telos Corporation.

Logistics and Sustainment: 12% ($3.4 Billion)

The unglamorous back end of military AI is also among the most financially significant. Predictive maintenance systems that use AI to anticipate component failures before they occur, supply chain optimization platforms that manage the movement of millions of spare parts, and autonomous logistics vehicles that reduce the human cost of moving supplies through contested areas represent a $3.4 billion segment by 2030. The Pentagon spends more on maintenance and logistics than on any other single category of defense expenditure; even modest AI-driven efficiency improvements generate enormous dollar savings and translate into large contract values.

Other Applications: 8% ($2.3 Billion)

The remaining 8% of the market encompasses training simulation, medical AI, AI-enabled recruiting and personnel management, legal compliance automation, and experimental programs that have not yet scaled to the point of segment-level significance. This catch-all category grows slower than the primary segments but includes some of the most technically interesting applications, including AI-driven nuclear command-and-control modernization and autonomous electronic warfare.

Regional Distribution: North America's Consolidating Lead

The defense AI market is geographically concentrated to a degree unusual even in defense procurement. North America — meaning primarily the United States, with Canada as a secondary contributor — accounts for approximately 45% of total global defense AI spending, a share that has grown over the past five years as the DoD has accelerated AI investment while European allies have moved more cautiously.

The Asia-Pacific region at 30% is a more complex picture than the headline figure suggests. China's military AI spending is substantial and growing rapidly, but the vast majority of it flows through state-owned enterprises and research institutes that do not generate publicly investable revenue. The investable Asia-Pacific defense AI market is primarily centered on U.S. allies: Australia (through AUKUS programs and its own defense AI initiatives), South Korea (which has invested heavily in autonomous systems following North Korean drone incursions), and Japan (which has reversed decades of defense spending restraint following Russia's invasion of Ukraine and Chinese naval pressure).

Europe at 20% is catching up. The continent's defense AI spending has accelerated dramatically since 2022, driven by the NATO 2% GDP target that most members are now meeting or exceeding, the direct lessons of Ukrainian drone and AI warfare, and a growing political consensus that Europe must develop autonomous defense capabilities rather than depending entirely on American systems. The European company most likely to benefit from this spending surge is Helsing, a Munich-based defense AI startup that has attracted over $600 million in venture funding and secured contracts with the German Bundeswehr, the Royal Air Force, and the Swedish Air Force.

Top 10 Companies by Defense AI Revenue

The competitive landscape for defense AI revenue is split between two distinct company archetypes: legacy defense primes that are adding AI capabilities to existing platforms and contracts, and software-native challengers that have built AI-first architectures from the ground up. The primes have the contracts, the cleared facilities, and the established relationships. The challengers have the technology lead, the talent density, and the growth rates.

The most significant pattern in this ranking is the presence of two private companies — Anduril and Shield AI — in the top ten by revenue despite having no publicly traded equity. Both companies are growing revenue at rates far exceeding any of the public defense primes: Anduril has reported year-over-year revenue growth exceeding 100% for three consecutive years, while Shield AI has grown its government contract base from roughly $200 million to over $800 million in three years. If either company goes public at its current private valuation, it will immediately rank among the most AI-intensive defense companies in the public markets.

Startup Disruption: The Software-Native Insurgency

The most consequential development in the defense AI market is not a specific contract or a specific technology — it is the structural shift in where defense innovation originates. For most of the post-World War II era, defense innovation was driven by large prime contractors working within the framework of cost-plus contracts that rewarded development time and complexity. The more sophisticated the system, and the longer it took to build, the more profitable the contract. Agility was penalized; stability was rewarded.

Software-native defense companies have inverted this model. Anduril, Shield AI, and Helsing have each built their defense AI businesses around fixed-price contracts that reward rapid delivery, continuous software updates, and demonstrated field performance. Their cost structures are fundamentally different from legacy primes: their largest expense is software engineering talent, not manufacturing facilities, supply chains, or proprietary hardware development. A software update that doubles the capability of a deployed system costs a few engineers a few weeks; a hardware upgrade to an existing platform takes years and costs hundreds of millions.

Anduril Industries: The Platform Company

Anduril's ambition is not to build specific weapons. It is to build the operating system for autonomous defense. The Lattice platform, which runs Anduril's entire product portfolio from the Sentry tower sensors to the Ghost surface vessels to the Dive-LD underwater vehicles, is designed as the middleware layer that connects autonomous hardware to human command. Every Anduril system is a Lattice node; every Lattice node contributes to a shared operational picture.

This architecture creates powerful competitive moats. Once an operator's command network runs on Lattice, switching to a different system requires replacing not just the software but the entire doctrinal framework that has been built around it. The switching costs are prohibitive, which is precisely the dynamic that allows enterprise software companies to charge premium prices for inferior products. Anduril is betting that the same dynamic applies in defense.

The $60 billion valuation that Anduril achieved in its 2025 funding round reflects the market's assessment that this bet is working. At $60 billion, Anduril is valued at roughly 50 times its trailing twelve-month revenue — a technology company multiple, not a defense contractor multiple. The market is pricing in the expectation that Anduril's platform economics will drive revenue growth and margin expansion at rates that justify the premium. The risk is that defense procurement timelines, budget cycles, and political disruption could compress those growth rates below what the current valuation requires.

Shield AI: The Autonomous Flight Specialist

Where Anduril has pursued horizontal platform breadth, Shield AI has pursued vertical depth in a single domain: autonomous flight. The company's HIVEMIND AI pilot operates F-16s, MQ-35 V-Bats, and the company's own V-150 autonomous aircraft without human control, and has demonstrated in government testing that it can outperform human pilots in specific tactical engagements including within-visual-range air combat.

Shield AI's current revenue is built primarily on V-Bat sales to the U.S. military and exports to allied nations, combined with HIVEMIND software licensing. The strategic bet is that HIVEMIND becomes the standard autonomous flight AI for American military aircraft, in the same way that iOS became the standard operating system for Apple hardware. A licensing model on military aircraft autonomous flight AI, across a fleet that could eventually number in the thousands of platforms, generates the kind of recurring revenue that investors in defense software are pricing in at Shield AI's $5.3 billion valuation.

Helsing: Europe's Defense AI Champion

Helsing occupies a unique position in the global defense AI market: it is the only European-headquartered defense AI company with the scale, technical capability, and government relationships to compete with American counterparts. Founded in 2021 by Gundbert Scherf, a former McKinsey and German Defense Ministry official, and Torsten Reil, a computational neuroscientist, Helsing has built an AI platform designed specifically for European defense requirements — including EU data sovereignty rules that prevent the use of American cloud infrastructure for certain classified applications.

Helsing's HX-2 system, which integrates AI-enabled electronic warfare, signals intelligence, and targeting capabilities into a unified platform, is deployed on German Eurofighters and has been selected for integration on the next-generation Global Combat Air Programme (GCAP) aircraft being developed by the UK, Japan, and Italy. This gives Helsing a long-term revenue position on what will be the primary fighter aircraft of three major economies for the next four decades — a contract position that could generate billions in annual licensing revenue by the mid-2030s.

The Investment Thesis: Why Defense AI Is the Decade's Asymmetric Bet

Three structural factors combine to make defense AI an unusually attractive investment opportunity for the remainder of this decade. Understanding these factors requires comparing defense AI to other high-growth technology sectors where similar dynamics have historically played out.

Factor One: Government as the Ultimate Enterprise Customer

The Department of Defense and its allied counterparts represent a customer with characteristics that are exceptional in the global economy: near-infinite balance sheet depth, a legal monopoly on lethal force that makes technology switching costs catastrophic rather than merely expensive, multi-decade procurement cycles that provide revenue visibility unavailable in commercial markets, and a fundamental inability to shop on price alone when national security is the product being purchased.

In commercial enterprise software, a CFO who wants to switch from Salesforce to a competitor faces a painful but survivable transition. A defense ministry that wants to transition its autonomous weapons network from one AI platform to another faces the prospect of operational gaps during the transition — a cost measured not in dollars but in strategic vulnerability. This raises the lifetime value of a defense AI customer to levels that no commercial software company can match.

Factor Two: The Technology Is Genuinely Ahead of Adoption

In most established defense market segments, the technology is mature and competition is primarily on price and manufacturing efficiency. Defense AI is different: the technology is developing faster than acquisition systems can absorb it, creating a situation where the companies that can navigate procurement timelines while maintaining technology currency will compound their advantage over those that cannot.

The practical result is that the best-positioned defense AI companies today — those with existing contract vehicles, cleared facilities, demonstrated field performance, and the software architecture to update deployed systems continuously — are building moats that will be very difficult to challenge by the time slower competitors reach the market. This dynamic is reminiscent of the early commercial cloud market, where Amazon Web Services's early infrastructure investment created advantages in cost, scale, and ecosystem that competitors have spent a decade and tens of billions of dollars attempting to close.

Factor Three: Geopolitical Tailwinds Are Durable

Defense spending is cyclical, but the specific pressures driving defense AI investment — China's military modernization, Russia's demonstrated willingness to use conventional military force in Europe, North Korea's nuclear and missile advancement, and Iran's proxy network — are not short-cycle phenomena. These are multi-decade structural challenges that will sustain defense AI spending regardless of which party controls the American presidency, which government is in power in Berlin or Paris, or what the next debt ceiling negotiation produces.

The bipartisan consensus on defense AI investment in the United States is particularly durable. Project Maven, which began under Obama, was accelerated under Trump, continued under Biden, and has been further expanded under the current administration. The specific programs change; the commitment to AI-enabled military advantage does not.

Risks: The Things That Could Break the Thesis

A complete investment analysis must identify the factors that could undermine the defense AI growth thesis. Three risks stand out as material and underappreciated.

Regulatory and Ethics Backlash

The development of lethal autonomous systems without adequate international legal frameworks creates political and regulatory risk that is difficult to price. A high-profile incident involving an autonomous weapon system causing unintended civilian casualties — particularly one involving an American or allied system — could trigger congressional scrutiny, international legal pressure, and procurement freezes that would directly impact defense AI revenue streams. The International Campaign to Stop Killer Robots and similar advocacy organizations have successfully pushed this issue onto the agenda of the United Nations Group of Governmental Experts, and a binding treaty on lethal autonomous weapons remains a possibility, however remote.

Budget Discontinuity and Continuing Resolutions

Defense AI programs are particularly vulnerable to the continuing resolution problem: when Congress fails to pass a defense appropriations bill and the government operates under a continuing resolution that freezes spending at prior-year levels, new AI program starts cannot be funded. Given the frequency of CRs in recent American fiscal politics, this risk is not theoretical. A sustained CR environment in 2027 or 2028 could delay several major defense AI contract awards by 12-18 months, compressing the revenue trajectories that current valuations assume.

Export Control Complications

The International Traffic in Arms Regulations (ITAR) and Export Administration Regulations (EAR) create significant compliance costs and market access restrictions for defense AI companies. As American allies increasingly look to purchase AI-enabled defense systems, ITAR compliance requirements can make American products more expensive or legally complex than European or Israeli alternatives. The Biden administration's partial relaxation of AI export controls for allies was a step in the right direction; further policy clarity on what AI systems can be shared through AUKUS, NATO, and bilateral partnerships without full ITAR review would be a meaningful growth catalyst.

ETF Analysis: Accessing Defense AI Through the Public Market

For investors who prefer diversified exposure to the defense AI theme rather than individual stock concentration, three ETFs offer meaningful but imperfect access to the sector. Each has a different construction methodology, expense ratio, and emphasis that makes it better suited to different portfolio objectives.

iShares U.S. Aerospace and Defense ETF (ITA)

ITA is the largest and most liquid defense ETF, with over $5 billion in assets under management. Its construction methodology weights holdings by market capitalization, which means the largest positions are in the largest defense companies: Raytheon, Lockheed Martin, Boeing, and Northrop Grumman collectively account for over 50% of the fund. This creates a fundamental tension with the defense AI investment thesis: the companies with the highest AI exposure scores are the smaller companies, which receive smaller weights in a market-cap-weighted fund.

ITA's expense ratio of 0.40% is competitive for the category. Its liquidity is excellent, with average daily trading volume exceeding $200 million. For investors who want broad defense sector exposure with AI as one of several themes, ITA is the appropriate instrument. For investors whose thesis is specifically the startup disruption of legacy primes, ITA's heavy weighting toward those primes is a structural mismatch.

Invesco Aerospace and Defense ETF (PPA)

PPA uses an equal-weighting methodology across its component holdings, which gives smaller, higher-growth defense companies a proportionally larger weight than they receive in market-cap-weighted alternatives. The fund holds approximately 50 companies, including Palantir, Leidos, Booz Allen Hamilton, and SAIC alongside the traditional primes. The equal weighting means Palantir — which has roughly 10% of Lockheed Martin's market cap — receives approximately the same portfolio weight as Lockheed. This creates significantly more AI exposure than ITA at the cost of more concentration in smaller, potentially more volatile positions.

PPA's expense ratio of 0.61% is higher than ITA's. Its liquidity is lower, with daily trading volume around $30-50 million. For investors with specific conviction on defense AI rather than the broader defense sector, PPA's equal-weighting methodology aligns better with the investment thesis, though the higher expense ratio and lower liquidity are meaningful frictions at scale.

SPDR S&P Aerospace and Defense ETF (XAR)

XAR also uses modified equal-weighting with caps on individual position sizes, producing a construction that is more balanced than pure market-cap weighting but less extreme than PPA's approach. It holds approximately 30 companies and has an expense ratio of 0.35% — the lowest of the three primary defense ETFs. Its defense AI exposure is meaningful, including positions in Palantir, Leidos, and Booz Allen, alongside the large primes.

XAR is the best option for investors who want balanced defense sector exposure with reasonable AI participation at the lowest cost. It lacks the specific AI tilt that PPA provides but compensates with lower fees and a slightly larger holding universe that includes some mid-cap defense technology companies not represented in the other funds.

Individual Stock Analysis: AI Exposure Scores

For investors willing to accept individual stock risk in exchange for higher AI exposure concentration, the following analysis rates the primary publicly traded defense companies by their AI revenue exposure, growth trajectory, and valuation relative to AI-specific metrics. The AI Exposure Score (AES) is a composite measure of the percentage of company revenue attributable to AI-enabled systems or services, the growth rate of that revenue relative to total company revenue, and the defensibility of the company's AI market position.

Palantir scores highest on pure AI exposure but carries the most significant valuation risk. Trading at approximately 65-80x forward earnings, the stock requires sustained high-double-digit revenue growth to justify current prices. Any evidence of deceleration in government AI spending or delays in the conversion of pilot programs to permanent contracts could produce significant multiple compression. Palantir is the highest-conviction defense AI investment, but it requires both the right thesis and the right time horizon — probably a minimum of five years.

Kratos Defense offers the best risk-adjusted AI exposure among public defense companies. Its autonomous aircraft programs — particularly the XQ-58 Valkyrie loyal wingman and the UTAP-22 Mako target drone — position it directly in the fastest-growing segment of the defense AI market at a significantly lower valuation multiple than Palantir. The company's 2025 contract wins on the Collaborative Combat Aircraft (CCA) program, which will field autonomous wingmen alongside crewed fighters, provide a multi-year revenue runway that is not yet fully reflected in analyst consensus estimates.

Leidos and Booz Allen Hamilton offer lower-volatility, lower-growth exposure to defense AI through their systems integration and AI services businesses. Both companies have strong CDAO relationships, cleared workforces, and the program management capabilities that newer entrants lack. They will not grow at 40% annually, but they will not decline 40% in a market correction either. For risk-averse investors who want defense AI exposure without technology stock volatility, Leidos and Booz Allen are the appropriate instruments.

"The defense AI trade is not about picking the winner in a technology competition. It is about identifying the companies that will still be in the room when the contracts are signed a decade from now. That is a different analysis."

-- Defense equity analyst, Goldman Sachs, 2025

The Decade's Trade: Patience, Position Size, and Asymmetry

The $28.67 billion defense AI market by 2030 is not a forecast that requires faith in speculative technology. The contracts underpinning it are largely already awarded or in the final stages of competition. The programs are funded in the defense budget. The geopolitical pressures that drive them are not abating. The primary uncertainty is not whether defense AI spending will materialize, but which companies will be positioned to capture it as the market evolves from early adoption to institutional dependence.

The investment framework that emerges from this analysis has three tiers. The first tier — highest conviction, highest risk, longest time horizon — is Palantir as the defining public defense AI company and Anduril as the private market bet on platform disruption (accessible through secondary markets or late-stage VC exposure for qualified investors). The second tier — balanced conviction, moderate risk, medium time horizon — is Kratos for autonomous platform exposure and Leidos for systems integration exposure. The third tier — diversified exposure, lower risk, flexible time horizon — is an ETF position in XAR or PPA for investors who prefer to express the theme without single-name concentration.

The risks are real: regulatory disruption, budget discontinuity, ethics backlash, and the execution risk that any rapidly scaling technology company faces. But the structural case — government customers with infinite switching costs, technology developing faster than procurement can absorb it, and geopolitical tailwinds that are durable across political cycles — is as strong as any investment thesis in the current market. Defense AI is not the only asymmetric bet of the decade. But it may be the most defensible one.