- The inquiry covers AI operations broadly — not a single product or abuse, but the structural question of whether AI creates winner-take-all dynamics across multiple sectors simultaneously.
- Competition Commissioner Teresa Ribera has framed the investigation around “entrenchment of corporate power” — a concept that goes beyond traditional price and output analysis to examine whether AI creates durable advantages that competition cannot erode.
- The EU is the only major jurisdiction conducting AI-specific competition enforcement at this scale. US antitrust agencies under the current administration have signalled a lighter touch, making Brussels the primary regulatory battleground for the global AI industry.
The European Commission’s competition directorate confirmed in March 2026 that it has opened a wide-ranging antitrust inquiry covering the “entire” AI operations of major technology companies. Competition Commissioner Teresa Ribera, speaking at a regulatory event, said the Commission is examining how companies can “entrench corporate power” in AI markets in ways that standard competition analysis may not capture adequately. Nvidia and Meta are specifically named in reporting as early focal points, but the inquiry’s scope extends to Microsoft, Google, and Amazon’s cloud infrastructure divisions.
The investigation operates under the EU’s existing competition framework — Articles 101 and 102 of the Treaty on the Functioning of the European Union — rather than the AI Act or the Digital Markets Act, though findings from those regulatory processes will inform it. The Commission has the power under competition law to impose fines of up to 10 percent of global revenue, require divestiture of businesses, and mandate structural changes to how companies organise their AI capabilities. For a company like Microsoft, which has deeply integrated OpenAI’s models into its cloud and productivity products, a structural remedy could be significant.
The theory of harm
Traditional competition law asks whether a dominant company is abusing its position to harm consumers or foreclose competitors. The EU AI antitrust inquiry is grappling with a harder question: whether a company can use AI to create dominance in new markets simultaneously — using data advantages from one market to train models that win in another — in a way that foreclosure happens before regulatory intervention is possible. The UK’s Competition and Markets Authority raised the same concern in its April 2026 AI market report, noting that agentic AI systems can also reduce competitive uncertainty between rival firms by making pricing strategies predictable, creating an implicit form of coordination.
Nvidia’s position in the inquiry is structurally different from the application-layer companies. The Commission is examining whether Nvidia’s dominance in AI accelerator chips — approximately 80 percent market share — gives it inappropriate leverage over AI model development timelines, pricing, and research directions across the entire industry. The concern is not traditional monopoly pricing but infrastructure control: that whoever controls the hardware on which AI runs can exert influence over what AI is built and how.
The enforcement gap
The EU’s willingness to pursue AI antitrust is partly a product of policy divergence from the United States. The Trump administration’s antitrust agencies have signalled reduced enthusiasm for tech-sector enforcement compared with the Biden era, creating a situation where Brussels and Washington are moving in opposite directions on the same companies. For the technology industry, regulatory arbitrage between jurisdictions is a known management challenge; AI antitrust at EU scale, without a parallel US process, raises the question of whether European enforcement alone is sufficient to change behaviour in markets that are global.