From cheapest to best

Public procurement is currently treated as a technical and economic field. Yet with the billions Europe spends every year, governments could push the tech giants in a greener and more democratic direction. It has almost become a political reflex to talk about the public sector as a problem: too big, too cumbersome, too bureaucratic. Digitalisation is presented as the solution — a means of downsizing the state, making it “leaner”, more efficient and less costly.

That narrow understanding overlooks how the public sector has historically been a driving force for innovative change, and why simply buying European is not enough on its own. The public sector does not merely administer society. It organises it. In a digital economy, that organisation increasingly takes place through the choices made in the development and procurement of digital systems.

The question is therefore not how much of a public sector we should have, but what role it should play in shaping digitalisation. Recently, governments at central and state level — in France and Germany respectively — have decided to gradually replace predominantly American operating systems, collaboration tools, cloud infrastructure and artificial intelligence platforms with national and European alternatives. These decisions are not merely technical; they are political. They aim to reduce dependence on American suppliers, strengthen digital sovereignty and provide greater control over data and infrastructure.

These examples illustrate how public procurement can be actively used to shape the market and set the direction for digitalisation — not merely as consumption, but as industrial policy in practice, addressing security concerns by reducing dependence on American tech giants in the process. The EU has a chance to improve the framework through the review of the Procurement Directives, as well as through the related Industrial Accelerator Act and the forthcoming Cloud and AI Development Act.

Despite this, tendering processes are still not generally perceived as — or used enough as — a policy instrument, even though more than 14 per cent of EU GDP, equalling roughly €2 trillion every year, is spent on procurement across Europe. Given how rapidly our societies are digitalising, a significant proportion of that staggering figure must be assumed to go towards digital solutions. Public procurement is therefore not merely a budget item, but a structural demand that helps to define which technologies are developed and which standards become market norms.

Digitalisation is not neutral

Yet procurement is mostly treated as a technical discipline. Criteria are formulated narrowly, evaluations are reduced mainly to price and functionality, and broader societal considerations appear secondary — or as risks to be minimised. This is a mistake. For digitalisation is not neutral. It embeds specific priorities: between efficiency and legal certainty, between innovation and control, between short-term optimisation and long-term sustainability. Procurement should respect social, environmental and tax conditionalities, while supporting the rule of law and the due process we expect from public administration and from the digital infrastructure on which it runs. It is not enough for these priorities to be reflected in legislation; they also need to be actively implemented in the specific systems that our society depends on. The question then becomes not how the public sector can buy as cheaply as possible — but how it can use its demand to set the direction.

Making public procurement a better tool

This applies, first and foremost, to the environment. Digital systems often appear intangible, but they depend on an extensive physical infrastructure: data centres, hardware production and global supply chains. Operation and development require significant amounts of electricity, and cooling entails high water consumption. The production of equipment draws on scarce resources and causes environmental impact throughout its entire life cycle.

There should therefore be a requirement for systematic environmental accounting covering the entire digital life cycle — from development and operation to upgrades and decommissioning. This will not only improve individual decisions but also shift the incentive structure in the market. Public procurement will then avoid solutions that appear effective in the short term but in reality are resource-intensive and problematic in the longer term.

Secondly, rights. Digital systems can have far-reaching consequences for the rights of citizens and public sector employees. Automated decisions, profiling and data-driven decisions can affect access to services, jobs, promotion, legal certainty and equal treatment. Even where decisions are not made entirely automatically, the systems’ recommendations can have a decisive impact on individuals or on groups that are already marginalised.

Requirements for impact assessments based on fundamental rights should therefore be introduced. These should be drawn up in collaboration with citizens’ representatives and union representatives, carried out prior to implementation, updated regularly and made publicly available.

Thirdly, transparency and governance. Developers of digital systems for the public sector should be required to publish a range of information. It is disturbing that the largest technology companies have successfully lobbied to prevent disclosure of data on the energy and water use of their data centres. Equally, it is unacceptable that they get away with not paying their fair share of tax while at the same time gobbling up public funding. The public must have insight into the digital systems they pay for and their environmental impact, and the opportunity to verify how those systems function. Systematic legality tests should also be carried out to ensure compliance with applicable law and agreements. Without transparency and access to the core of the systems, it becomes difficult to identify and correct errors, bias and unintended consequences — and thus also difficult to assign responsibility.

Fourthly, democracy and participation. Digital systems are often developed without the systematic involvement of the employees and citizens who use them or who are affected by them. This creates not only implementation problems, but a democratic deficit. If involvement is made a requirement in procurement and development, not only will the quality of the solutions change, but so will their functionality and legitimacy.

And finally, sovereignty. Dependence on private suppliers limits the public sector’s scope for action — both technically and politically. When data and systems are not under public control, the public sector’s autonomy is undermined. Data from these systems must not be exploited commercially or moved outside European jurisdiction. Setting limits here is not a technical detail, but a political priority. Public ownership of digital infrastructure should be considered, and public-public partnerships — as promoted by the Commission in the forthcoming Cloud and AI Development Act — should be supported. This should be backed up with funding to assist, for example, municipalities to develop systems focused on the orientations we set out here, and not on supporting profit-maximising and “enshittification” principles.

The issues outlined above can all be addressed within existing procurement frameworks. Procurement legislation does not preclude consideration of the environment, rights or long-term societal impacts. The key factor is how the legislation is interpreted and applied.

Today, the prevailing view is that the public sector must primarily minimise costs and risks. This reduces procurement to an administrative function. An alternative view, as presented here, is that the public sector can, through its procurement, influence the market and the solutions available. From this perspective, the public sector does not become a hindrance to innovation, but a prerequisite for ensuring that innovation moves in a direction compatible with democratic values and sovereignty, and that protects fundamental rights and the environment.

These arguments are developed further in a new EPSU policy briefing, “Promoting digital sovereignty and fundamental rights: six principles for inclusion in public service procurement contracts”, by Christina Colclough and Hannah Johnston, first published in the FEPS book Algorithmic Rule – AI and the Future of Democracy in Sweden and Beyond.

  This post is sponsored by EPSU