A single statistic underpins the European Commission’s new technological sovereignty package: the Union relies on non-EU suppliers for more than 80 percent of its key digital products, services, infrastructure and intellectual property. Whether the figure is read as a warning or an indictment, it frames the problem the package is meant to solve, and exposes how steep the climb will be.
Unveiled in early June, the package bundles two legislative proposals, a Chips Act 2.0 and a Cloud and AI Development Act, with an Open Source Strategy and a roadmap for digitalising energy. The semiconductor half aims to address Europe’s overdependence on foreign chip design and manufacturing, and the Commission has signalled it would prioritise building an advanced foundry on European soil. The cloud half would streamline the construction of data centres and, more consequentially, create a single EU-wide framework for assessing whether a cloud or AI service is genuinely sovereign.
The political language has been unusually blunt. The aim, as one official put it, is to ensure that nobody holds a kill switch over European infrastructure, a thinly veiled reference to the bloc’s reliance on American hyperscalers and the anxiety that access could be weaponised in a trade or security dispute. That anxiety is not paranoid. The control points of the modern digital economy, leading-edge chips, frontier model training, the largest cloud platforms, sit overwhelmingly outside the EU, and recent geopolitical turbulence has made dependence feel less like efficient specialisation and more like exposure.
The numbers attached to the ambition are sobering. The Commission’s own estimates put the investment required at roughly 120 billion euros for semiconductors, 200 billion for data centres by 2036, 100 billion for cloud and AI, and 2 billion for open-source software over seven years. These are not figures the EU budget can supply. They imply a mobilisation of private capital and national spending on a scale European industrial policy has rarely achieved, and they invite an uncomfortable comparison with the first Chips Act, whose headline targets outran the money that actually materialised.
There is also a tension inside the strategy itself, between autonomy and openness. The EU does not want to become a closed bloc; much of its prosperity depends on trade and on access to the best technology wherever it is made. So the package tries to thread a needle: reduce dependence on single foreign suppliers without erecting a wall that cuts European firms off from global innovation. The honest difficulty is that sovereignty and openness pull in opposite directions, and every concrete decision, which foundry to subsidise, which cloud to certify as sovereign, which open-source project to back, will force a choice between them.
The deeper problem is demand, not just supply. Europe can subsidise a foundry, but a foundry without customers is a museum. The continent’s structural weakness has been the relative scarcity of large domestic technology firms that consume cutting-edge chips and cloud at scale. Building capacity on the supply side without cultivating that demand risks producing expensive infrastructure that European champions are too small to fill and foreign rivals too entrenched to cede.
None of this makes the package misguided. The strategic logic, that a bloc dependent on others for 80 percent of its digital backbone is a bloc with limited room to manoeuvre, is hard to dispute, and doing nothing carries its own compounding cost. But strategy documents are cheap and foundries are not. The package will be judged not on the elegance of its framing but on whether member states fund it, whether industry builds with it, and whether, a decade from now, that 80 percent figure has actually moved. The ambition is correct. The execution is everything, and Europe’s record on execution is the reason for scepticism.




