The way militaries fight wars is fundamentally changing. For centuries, defense meant tanks, aircraft, and artillery built by a handful of prime contractors. Today, software is becoming equally central, with soldiers increasingly fighting through screens rather than boots on the ground. The military AI industry now spans everything from cloud computing to autonomous drones and decision-making software, reshaping what was once called the military-industrial complex into a military-tech one.

This shift traces back to the US drone program of the 2000s and 2010s, which normalized remote, screen-based warfare. But the pace of transformation is accelerating. Global military expenditure climbed to $2.88 trillion in 2025, with governments pouring record sums into defense budgets. Billions of dollars are now flowing into a handful of companies leading this AI transformation, and the landscape is far more fragmented than the old defense industry ever was.

Who Are the Four Types of Defense AI Companies?

Researchers at the Stockholm International Peace Research Institute (SIPRI) mapped the military AI industry and identified four distinct categories based on business models and roles. Understanding these categories helps explain why the defense tech sector looks so different from traditional defense contracting.

• Defense Primes: Long-established arms manufacturers like Lockheed Martin, RTX, Northrop Grumman, BAE Systems, and General Dynamics that have supplied militaries with hardware for decades and are now layering AI onto existing systems.
• Neoprimes or Defense Startups: Newer, software-first firms like Palantir and Anduril that specialize in data products and AI platforms for military applications, competing directly with traditional contractors.
• Big Tech Companies: Giants like Google, Meta, and Amazon Web Services (AWS) that control roughly 67 percent of the global cloud infrastructure market and are increasingly providing computing power for military AI applications.
• Foundational Model Providers: Frontier AI labs like OpenAI and Anthropic in the US and Mistral AI in Europe that build the underlying AI models other companies integrate into military tools.

How Are Traditional Defense Contractors Adapting to AI?

The old guard arms companies are not disappearing; they are evolving. In February, Lockheed Martin announced it was flight-testing an AI tool for its F-35 fighter jets to help pilots identify hostile aircraft faster. L3Harris, another American defense firm, built AI directly into its camera systems for real-time targeting and integrated platforms like Shield AI's Tracker, which helps detect hostile drones. These companies hit all-time highs after the onset of the US-Israel war on Iran, reflecting how deeply AI is now being woven into their product lines.

The integration is not superficial. Traditional contractors are embedding AI into weapons systems, surveillance platforms, and command-and-control networks. This approach allows them to leverage decades of military relationships and hardware expertise while adding software capabilities that were previously the domain of startups.

What Makes Neoprimes Like Palantir and Anduril Different?

The newer category of defense startups operates on a fundamentally different model. Palantir is perhaps the best-known name, largely through its role in Project Maven, the Pentagon's flagship military AI program designed to speed up decision-making in battle. Maven relies on AI and machine learning to automatically sift through satellite imagery, drone videos, and signals intelligence to identify and prioritize targets. The system has been deployed in various conflicts, including the Ukraine war and the US-Israel war on Iran, where Maven identified 1,000 targets in the first 24 hours of the latter conflict.

Anduril's Lattice platform is a direct competitor. Unlike Maven, which focuses on data integration and analysis, Lattice ties its own hardware, drones, sensor towers, and autonomous vehicles into a single command network. This hardware-software integration approach has made it heavily utilized by the US Army. Both companies represent a shift toward integrated platforms rather than point solutions.

Why Are Big Tech Companies Entering the Military Market?

In May, the US Department of Defense announced a groundbreaking agreement with eight tech and AI companies to use their technologies for classified networks. The companies included SpaceX, OpenAI, Google, NVIDIA, Reflection, Microsoft, Oracle, and AWS. This deal was brokered to "accelerate the transformation towards establishing the United States military as an AI-first fighting force". The agreement means these companies have been cleared for IL6 and IL7 networks, the most sensitive and classified computing environments the Department of Defense operates.

This represents a major shift. Tech giants that once avoided military work are now deeply embedded in defense AI infrastructure. Their massive cloud and computing capacity makes them essential partners for processing the enormous volumes of data modern warfare generates. Companies like AWS control roughly 67 percent of the global cloud infrastructure market, giving them outsized influence over military AI capabilities.

What Role Do AI Model Makers Play in Military Technology?

The final category, foundational model providers, builds the underlying AI systems that other companies integrate into military tools. OpenAI and Anthropic in the US and Mistral AI in Europe are key players. Notably, many of these frontier AI labs were initially against their products being used in militaries. However, in 2024, OpenAI removed its military-use ban, and Palantir and Anthropic partnered in what has become a relationship fraught with lawsuits over how Claude was being used by the US military.

China is also developing military AI models. Chinese researchers have been building AI models for military use since the early 2020s, including using Meta's Llama model as the base for a People's Liberation Army (PLA) tool called ChatBIT, designed to gather intelligence and support decision-making in the field. The Pentagon estimated Chinese military AI spending at roughly $1.5 to $2 billion annually, comparable to American figures. This year, the Pentagon added Alibaba, Baidu, BYD, NIO, RoboSense, and Unitree to its list of Chinese military companies believed to be helping the military.

What Recent Contracts Signal About the Industry's Direction?

Recent major contracts reveal how quickly the military AI market is consolidating and expanding. On July 10, the UK Ministry of Defence signed a 2 billion pounds ($2.7 billion), 15-year contract with Omnia Training, a consortium led by Raytheon UK, to train 60,000 soldiers a year on using AI to make a "combat-ready" British military. Days earlier, NATO awarded contracts to run the alliance's Enhanced Air Command and Control Data Platform to Anduril for its Lattice AI software alongside Palantir and the French military-tech joint venture Athea SAS.

These deals signal that military AI is moving from experimental programs to core operational capability. The Pentagon's 2014 "Third Offset Strategy" bet explicitly on AI defining the next generation of warfare, but the pace of implementation has accelerated dramatically as global conflicts create urgent demand. With wars raging from Ukraine to the Middle East, deals like these are multiplying fast, and the industry itself is evolving just as quickly as traditional arms firms and new entrants jostle for market share.

The military-tech complex is no longer a hypothetical future. It is here, fragmented across four distinct business models, each competing for billions in contracts and shaping how the world's militaries will fight wars for decades to come.