Eroding Sovereignty in the Age of “War as a Service”

[Dr Rupert Barrett-Taylor is a Research Fellow at the Alan Turing Institute focused on the ethical, operational, and epistemological implications of Artificial Intelligence in military operations.

Dr Matthew Ford is an academic in the Department of War Studies at the Swedish Defence University (FHS) focusing on war and the data-saturated battlefields of the 21st century]

The state’s exclusive control over the legitimate use of force is quietly eroding—apparently with its own consent. Yet this development challenges one of the most enduring principles of political order. For Max Weber, the state is a community that claims a monopoly on the legitimate use of violence within a defined territory. The authority to wage war, command armed forces and decide on the use of lethal force has long defined state sovereignty. This monopoly forms a cornerstone of international relations, where states underpin the global order. In the twenty-first century, however, the exercise of violence is increasingly reliant on digital systems of data, code, and infrastructure owned and managed not by states, but by the technology industry.

The conduct of warfare has become dependent on networks of commercial platforms and cloud services. This transformation is not merely technical. It represents a fundamental shift in power relations between state and market, and a growing challenge to the legal and political concept of sovereignty. When the data flows that inform decisions about war and peace travel through private networks and are processed by proprietary algorithms, the claim of the state to exclusive authority over violence becomes conditional.

Institutions of war are increasingly mediated by corporations whose Venture Capital funders and their interests are divergent from the states they serve. Prominent tech figures fund ‘Network State’ linked projects. These foreground market-driven visions of governance blur traditional political sovereignty, extending their reach into statecraft.  This post sets out the urgent need to study this phenomenon and the means by which to undertake such study.

The Marketisation of War

The state has always relied on private industry to equip and sustain its armed forces. From shipbuilders to arms manufacturers, markets have long been partners in war. But the current transformation differs in kind, not just degree. Where past partnerships involved the production of physical materiel, contemporary relationships have emerged around the sovereign production and control of information.

The conduct of “digital warfare” rests on the collection, transmission, and analysis of vast quantities of data. These processes take place through what Bratton calls “the stack”: a layered ecosystem of hardware, software, and cloud infrastructure that supports everything from intelligence gathering to battlefield communication. Once, much of this infrastructure was owned and operated by states (further elaborated by Rogers and Bienvenue). Today, it is dominated by corporations from Amazon Web Services, Microsoft, Google, as well as a constellation of data analytics firms such as Palantir and Anthropic.

These companies provide computational power and algorithmic tools “as a service” to their state clients. These alliances form the technological and economic geography of the modern military cloud: a global assemblage of commercial actors who process sensitive information on behalf of governments. Defence ministries no longer own their data environments. Instead, they rent them. The sovereignty of the state extends only as far as the limits of these service contracts.

The Martial Imaginary

The roots of this dependency lie in a powerful shared cultural vision within military institutions of how war should be fought. Since at least the Cold War, NATO-aligned allies have dreamed of perfect surveillance and instantaneous destruction derived from the ability to see everything and strike instantly. From the AirLand Battle doctrine of the 1980s to the “Revolution in Military Affairs” of the 1990s and the contemporary “Multi-Domain Operations” this fantasy has persisted in different guises.

In each generation, technology is imagined as the key to omniscience and precision. Drones, satellite networks, hypersonic weapons, and now AI-assisted decision systems are all heirs to the same belief that war can be won through information dominance. Yet the martial capacity to realise this fantasy is limited by its lack of technical expertise. The tools needed to fulfil the dream lie increasingly in the hands of software engineers, data scientists, and private Big Tech businesses headquartered thousands of miles from the battlefield.

The convergence of martial desire and Big Tech has produced a symbiotic relationship. Tech companies see in defence contracts a lucrative market and a chance to validate their products with military credentials. Military leaders in turn see Big Tech as the embodiment of innovation and efficiency and a way to burnish their technical legitimacy. Both communities share an ideological faith in optimisation and data-driven decision-making. What emerges is not just cooperation but an alignment of worldviews that places computational control at the centre of modern warfare.

Behind the relations of martial and private sector actor thus lies an ideological convergence. Silicon Valley operates under what Richard Barbrook and Andy Cameron famously termed the “Californian Ideology” which blends libertarian economics with techno-utopian optimism. It celebrates decentralisation, innovation, and disruption, often at the expense of collective or state control. When this ideology meets the martial obsession with optimisation and precision, the result is a shared belief that technology can perfect war.

This belief recasts warfare as a domain of technical management rather than political judgment (forthcoming from Barrett-Taylor). The language of operational advantage and efficiency obscures the moral and legal dimensions of state violence. As militaries integrate AI systems trained on civilian data and hosted in commercial clouds, they internalise the subscription-based, data-driven, and perpetually upgradeable logic of the market. This embeds civilian infrastructure into military systems offering greater reliability and resilience yet conversely enmeshes citizens in targeting cycles in which they actively or unwittingly participate. Consequently, violence is delivered by the market as a service, not as a sovereign act.

The Stack as Sovereign Infrastructure

How this alignment erodes sovereignty can be traced through the political nature of the digital stack. It is more than just technical architecture but a system of layered physical, digital, and legal dependencies. Fibre-optic cables, data centres, cloud servers, software platforms, and user applications form horizontal layers, each operated by different corporate actors. Data passes vertically through these layers, often across borders and jurisdictions, to reach the user whether a civilian uploading a file or a military user targeting an enemy position.

From an international law perspective, this distribution of control creates a profound ambiguity. The physical servers may be located in one state, the company headquarters in another, and the end-user, a soldier, deployed abroad. The chain of command that governs military decision-making becomes interlaced with the corporate hierarchies that govern cloud infrastructure. Sovereignty becomes entangled. The stack is an assemblage that binds states, corporations, algorithms, and infrastructures into a constantly shifting configuration of power. The capacity to wage war emerges not solely from the state but from the interplay of these distributed actors.

This has practical consequences. When a defence system depends on a privately hosted algorithm or an outsourced AI model, operational continuity relies on the provider’s willingness and ability to maintain access. Software licences, subscription models, and data-sharing agreements become instruments of control. The same mechanisms that manage consumer applications such as Digital Rights Management, server authentication, update cycles now govern the functioning of military tools. A single expired domain or server outage can paralyse critical systems. This is seen when commercial platforms temporarily lock out users due to expired verification keys, or platforms such as Amazon Web Services have outages. Such vulnerabilities illustrate how the power to wage war is conditioned by commercial infrastructures that operate beyond state authority.

From Procurement to Dependence

Historically, defence procurement was a bounded relationship. The state specified, funded, and owned the systems it deemed necessary for war. The contemporary model of War as a Service (WaaS) reverses this logic. Its effect can be understood through the existing model of Software as a Service (SaaS). Rather than acquiring a tool outright, users access software through cloud-based subscriptions maintained by the provider. The convenience of constant updates and integrated functionality comes at the cost of autonomy.

When applied to the defence sector as WaaS, this model turns core military functions into rented capabilities. A command platform is rendered not a discrete product but as a node within a broader commercial ecosystem. Its analytical power depends on underlying cloud infrastructure and on the continued availability of proprietary machine-learning models. Each component of the stack belongs to a different entity, governed by distinct legal and contractual terms.

The result is a complex web of dependencies in which the state is only one actor among many. This configuration blurs the boundary between contractor and combatant. When software or hardware engineers maintain civilian data centres or update code for systems integrated into military targeting or threat assessment, they effectively become the digital munitions workers of the twenty-first century. Yet international law continues to classify them as civilians. Legal frameworks governing war—rooted in territorial jurisdiction and state accountability—struggle to accommodate a world where the critical infrastructures of combat are privately owned and globally distributed.

Virtual Sovereignty and the State in Retreat

The erosion of sovereignty in cyberspace has been a concern for decades, but WaaS exposes a deeper, structural problem. The withdrawal of the state from direct ownership of its technological foundations means its digital sovereignty is contingent and reversible. It exists only so long as those that host its data permit it. In physical terms, data centres are fixed installations subject to territorial law. In functional terms, however, they form a transnational terrain of power that no single state controls. A classified dataset may reside on servers located in multiple jurisdictions, connected by fibre networks crossing oceans.

Encryption and access protocols create enclaves of virtual territory that can be expanded, contracted, or deleted at will. Sovereignty in this context is fluid, and a matter of negotiated access rather than absolute control. This new geography of war challenges the basic assumptions of international law. Territorial integrity and jurisdictional exclusivity presuppose that the means of violence are materially contained within the state. Yet in the cloud, authority is dispersed. The virtual battlefield is hosted, not owned. When lethal decisions depend on data processed by a commercial algorithm, questions of accountability become opaque. Who is responsible if an AI-generated target list proves faulty? The state that acted on it, the contractor that supplied it, or the cloud provider that hosted it?

Such ambiguities echo older debates about Private Military Contractors, but the digital dimension adds an unseen layer of dependency. Unlike the mercenary, the platform provider cannot simply be dismissed or replaced because its infrastructure underpins every function of the modern state, from tax collection to warfighting. The line between civilian and military technology has collapsed, and with it the legal clarity that once underwrote the state’s monopoly on force.

With the erosion of sovereignty comes a redefinition of the relationship between soldiers and the state. Treating warfare as an optimisation problem transforms the soldier into both a battlefield target and a blind executor of technical demands. Deskilling, already a feature of the “uberisation” of many sectors of the tech economy, also reshapes the military profession, bringing a loss of status and purpose. The soldier of the future will increasingly function as an extension of the machine—part of a target set designed to provoke the enemy into revealing themselves for automated destruction.

Contesting the Erosion

If the state wants to retain its Weberian position, then the challenge it faces is as much cultural and political as technical. The allure of “WaaS” lies in its promise of frictionless efficiency, an illusion that flatters both generals and technologists. To resist it requires reimagining the relationship between technology, power, and the public good. The tools of digital war cannot remain the private property of corporations whose accountability lies to shareholders rather than citizens. Analysis of “WaaS” through the political-technological intersections of the stack reveal the mechanisms through which the private sector erodes state sovereignty and its consequences. The state was once the primary actor capable of imposing order on the chaos of war. Work must be done to protect the state from tech sector ambitions. For unless sovereignty is reimagined to encompass digital infrastructure, control over warfare may in the future belong not to nations but to networks. The digital stack is the place to start this work.