Software Becomes Central to Warfare
For most of modern history, military power was measured in tanks, aircraft, and artillery — hardware built by a small number of large defense contractors under long-term government agreements. That model is not obsolete, but it is changing rapidly. Software, data, and artificial intelligence are now integral to how militaries operate, and an increasing number of soldiers find themselves fighting wars through computer screens rather than on open battlefields.
This transformation has deep roots. The expansion of the US drone programme during the 2000s and 2010s introduced remote, screen-based warfare as a standard feature of modern conflict. Since then, military AI has expanded to include cloud computing infrastructure, autonomous drone systems, real-time surveillance tools, and AI-driven decision-making software. The traditional military-industrial complex is quietly reshaping itself into a military-technology complex, and enormous sums of money are following that shift.
A Growing Market Fueled by Record Defense Spending
Global military expenditure reached $2.88 trillion in 2025, with governments channeling unprecedented sums into defense budgets. Against this backdrop, AI procurement contracts are multiplying. In July 2026, the UK Ministry of Defence signed a £2 billion ($2.7 billion), 15-year deal with the Omnia Training consortium — led by Raytheon UK — to train 60,000 soldiers annually using AI tools aimed at building a combat-ready force.
Around the same time, NATO awarded contracts to US firms Anduril and Palantir, as well as the French joint venture Athea SAS, to manage the alliance's Enhanced Air Command and Control Data Platform. Anduril's Lattice AI software, which uses artificial intelligence and computer vision to consolidate data from multiple sources into a single operational picture, is central to that agreement.
The Pentagon has also struck deals with eight major technology companies — including Google, Microsoft, Amazon Web Services, OpenAI, SpaceX, NVIDIA, Oracle, and Reflection — to deploy AI across its most sensitive classified networks. The stated goal is to establish the United States military as an "AI-first fighting force."
Four Categories of AI Military Companies
Researchers at the Stockholm International Peace Research Institute (SIPRI) have mapped the AI military industry into four distinct categories, offering a clearer picture of who is involved and what role each type of company plays.
Defense Primes and Traditional Contractors
These are the long-established arms manufacturers — companies like Lockheed Martin, RTX, Northrop Grumman, BAE Systems, and General Dynamics — that have supplied militaries with hardware for decades and are now integrating AI into their existing product lines. Lockheed Martin, for instance, announced in early 2026 that it was flight-testing an AI-powered tool for its F-35 fighter jets designed to help pilots identify hostile aircraft more quickly. Meanwhile, L3Harris has embedded AI directly into its camera systems for real-time targeting capabilities.
Defense Startups and Neoprimes
A newer category of software-first companies, sometimes called "neoprimes," has emerged to sell specialized data and AI products to defense markets, as well as to policing and border agencies. Palantir is among the most well-known, largely due to its involvement in Project Maven — the Pentagon's flagship military AI programme. Maven uses machine learning to automatically analyze satellite imagery, drone footage, and signals intelligence in order to identify and prioritize targets. The system has reportedly been deployed in both the Ukraine conflict and other active war zones.
Anduril is a direct competitor, with its Lattice platform integrating the company's own hardware — including drones, autonomous vehicles, and sensor towers — into a unified command network. Shield AI, another firm in this space, is developing autonomous AI pilots for military aircraft and drones.
Big Technology Companies
Google, Meta, and Amazon Web Services together control roughly 67 percent of global cloud infrastructure, and their computing capacity has made them increasingly attractive partners for defense agencies. The Pentagon's recent agreements with several major technology firms have cleared these companies for access to IL6 and IL7 networks — the most sensitive classified computing environments the Department of Defense operates.
Foundational AI Model Providers
The final category consists of companies building the underlying AI models that other firms integrate into military tools. In the United States, OpenAI and Anthropic are prominent examples. In Europe, Mistral AI has emerged as a notable player. Chinese researchers have also been active in this space, reportedly using Meta's open-source Llama model as a foundation for a People's Liberation Army tool called ChatBIT, designed to support intelligence gathering and field decision-making.
Ethical Tensions and Shifting Policies
The Pentagon estimates Chinese military AI spending at roughly $1.5 to $2 billion annually, comparable to US investment. Several Chinese technology companies — including Alibaba, Baidu, and Unitree — have been added to the Pentagon's list of firms believed to be supporting China's military.
Ethical concerns have also emerged within the industry itself. Many frontier AI laboratories initially opposed allowing their technologies to be used for military purposes. However, OpenAI reversed its ban on military applications in 2024. In the same year, Palantir and Anthropic entered a partnership, though that relationship has since been complicated by legal disputes over how Anthropic's Claude model was being used by the US military.
Source: Al Jazeera
