Sanchez Says AI Self-Regulation Has Failed in Europe

Artificial intelligence cannot be left to regulate itself by the companies that control it, Spanish Prime Minister Pedro Sanchez said on Monday, according to reporting by Reuters from Madrid. The statement puts one of Europe's most vocal critics of big technology at the center of the global debate over AI self-regulation, at a moment when governments are racing to write rules for systems they increasingly view as too powerful to be governed by their makers.

Speaking at an event on AI regulation in Madrid, Sanchez argued that AI self-regulation has failed: the industry's voluntary guardrails have fallen short against the speed and scale of frontier models. According to Reuters, he described some advanced systems as aggressive, and said his government would reinforce its defensive cybersecurity measures in response, a concern underscored by recent real-world tests in which AI systems broke into corporate networks. The prime minister, an outspoken critic of what he calls tech oligarchs, framed the issue as a question of democratic control: the direction of a transformative technology should not be decided in boardrooms alone.

A 12-month roadmap with a gigafactory at its core

The remarks were paired with policy. Sanchez outlined a 12-month roadmap meant to balance responsible deployment with economic growth, placing public investment and research institutions at the center of Spain's AI strategy. The plan includes a proposed AI gigafactory, a large-scale compute facility intended to give researchers and companies access to the kind of processing power that is currently concentrated in a handful of private labs.

The gigafactory effort has been taking shape for months. According to reporting by Diari de Catalunya, Sanchez has met with business leaders including executives from Telefonica, Banco Santander and ACS to advance the project in Mora la Nova, in Catalonia, with San Fernando de Henares near Madrid also proposed as a partner site. The consortium behind the facility is already established, with the state holding 47.99 percent, the Catalan regional government one percent, and the private sector the majority stake. Government sources quoted in the report estimate roughly five billion euros of investment and a first operational phase targeted for 2028.

The roadmap also calls for new AI models developed with the Barcelona Supercomputing Centre, aimed at climate, health and energy applications. The emphasis is on public-interest research rather than commercial chatbots: systems designed to help model extreme weather, accelerate medical discovery and optimize energy grids. It is a deliberate contrast with the corporate labs that dominate headlines, and it reflects a bet that Europe can compete on applied scientific computing, an area where voluntary industry standards were not delivering public goods on their own.

Data centres, energy and local communities

Sanchez also turned to the physical footprint of AI. Data centres, he said, must meet high environmental and energy standards, protect data autonomy and benefit the local communities that host them. The point echoes a concern that has followed the gigafactory project from the start: the Terres de l'Ebre region would see enormous new power demand, and officials have pointed to closed-loop cooling and renewable energy synergies to argue the impact can be managed, according to Diari de Catalunya.

The prime minister's broader message was about who the technology serves. He argued that AI should be developed openly, with citizens seeing it as an opportunity rather than a force that quietly depletes scarce public resources. That framing turns AI self-regulation into a democratic question: if citizens bear the costs of data centres and the risks of opaque models, they should have a say in the rules. The argument lands at a time when the costs of AI are increasingly visible: electricity grids strained by training runs, water concerns around cooling, and communities questioning what they get in return for hosting server farms.

Governments step in as trust in self-regulation fades

Sanchez's comments fit into a wider shift. Governments around the world are drafting rules as worries mount over privacy violations, biased decision-making, job displacement, misinformation and the concentration of computing power among a small number of technology giants, according to Reuters. The era in which platforms could promise to police themselves is giving way to formal oversight, and AI self-regulation now faces scrutiny similar to what financial and pharmaceutical industries confronted decades ago. Brussels has already moved first with its AI Act; Madrid's roadmap suggests member states are now racing to fill the gaps with national enforcement, public compute and mandatory standards rather than trusting industry pledges.

The timing is also significant. Sanchez spoke just days after the heads of leading United States AI labs called for a slowdown in development, warning that models could soon improve on their own and slip beyond human control. The convergence of voices is striking: even the builders of frontier systems now describe limits to their own oversight, which undercuts the case that the industry can manage the risks alone. For readers following Washington's own moves, GenZ Newz recently covered plans for a US AI Force and AI czar, a sign that state power is entering the AI race on both sides of the Atlantic.

Spain's approach is distinct from outright bans or sweeping prohibitions. The roadmap treats regulation as an industrial strategy: set strict standards, fund public compute, build domestic research capacity, and require AI projects to earn their social license. It is the opposite bet from AI self-regulation, replacing trust in corporate promises with public infrastructure and enforceable rules. Whether that balance works will depend on execution, and on whether the gigafactory delivers computing power fast enough to matter.

For now, Sanchez has drawn a clear line. AI self-regulation, he argues, amounts to letting the companies that profit most from AI write the rules for everyone else. According to reporting by StratNews Global, the plan's details will roll out over the coming year, giving regulators, researchers and local communities time to shape how the technology lands. The next twelve months will test whether a mid-sized European country can chart a middle path between unchecked private expansion and heavy-handed state control, and whether the world follows Madrid's lead away from AI self-regulation toward enforceable public standards.