Sovereignty increasingly depends on control of infrastructure

The phenomenon of hybridization between states and tech companies - where one takes on the characteristics of the other and vice versa - is recent but rapidly evolving. Of course, the State still aspires to participate in the capital of companies - the entry of the US government into the shareholding of Intel is the most recent example - or acts as venture capitalist, as it selects technological projects (France with Mistral).

But now tech companies are transforming: from mere contractors they become private state infrastructure. This applies to the prospects of war: thus the Pentagon's recent relations with OpenAI, now a true military infrastructure of the state apparatus, or with Google, Microsoft, Amazon, Nvidia, SpaceX, or (in a more conflictual way) with Anthropic. But it also applies to times of peace: in many countries Palantir is a transversal component of the state security, administrative, healthcare and energy apparatus, as is Deepseek in China, both of which are now, in fact, cognitive infrastructures of the welfare state.

The relevance of the phenomenon of hybridization between public powers and private infrastructures is significant on the level of both international relations and state systems. In fact, the notion of sovereignty of modern international law, based on the classic triad of people, government and territory, evolves in contemporary international law through the partial replacement of territory with infrastructure. Hence the notion of digital sovereignty.

But as mentioned, state infrastructure is becoming increasingly private, and the definition of public control must change parameters and tools accordingly. If Starlink turns the signal on or off at any given time, it determines the outcome of a battle between Russia and Ukraine. Even an electoral result depends on the solidity of the technological infrastructure. Therefore, the sovereignty of the state - in peace and war - depends on its ability to control the increasingly private infrastructure.

The control of the technological infrastructure is achieved through various methods: agreements with companies or closure orders (most recently, Autistici Inventati) in the United States, their integration into the state apparatus (in China), or the definition of a regulatory structure (in the EU), which can also be asymmetric, precisely because it is not just the market and competition that are at stake, but state sovereignty. Each modality affects the effectiveness, efficiency and development capacity of the system in a different way, as well as the defense or promotion of the sovereign state.

Now, the need for a system of common rules is shared by big tech, to an even more convinced extent by those that develop artificial intelligence (Anthropic first), who actually invoke them regularly to overcome the 'prisoner's dilemma' that is bringing technological developments to extreme dystopian scenarios.

One way to reverse this trend or at least slow down the trend could consist in defining the rules of technological developments (artificial intelligence, blockchain, among others) in international agreements involving states, international organizations and big tech: let's call them mixed interinstitutional agreements. In this way, the legitimation of big tech resulting from their recognition as signatories would correspond to their responsibility for compliance with the agreements. This is moreover what happened before the birth of nation-states, when the various Indian companies - of which big tech seems in many ways to be the heirs - signed treaties with empires.

Such sharing of rules would allow companies to develop technology in a competitive but controlled environment, and the EU to orient the development of technology towards a model consistent with continental European welfare, while overcoming information asymmetries and regulatory capture.