A new analysis from Governance AI has quantified what many in the tech industry have long suspected: European data protection rules are significantly slowing the rollout of advanced artificial intelligence models. The report, which examined 375 large language models (LLMs) released between June 2018 and May 2026, found that at least 11% of models from companies like Meta, Google, OpenAI, and Anthropic were either delayed or never released in the European Union compared to the United States. The United Kingdom fared slightly better, with 7% of models delayed or withheld.
The study identifies data protection regulations—primarily the General Data Protection Regulation (GDPR)—as the main culprit. Of the 68 documented instances of delays or non-releases, regulatory factors were the primary cause in 56 cases. Non-text modalities, such as image, audio, and real-time video models, faced even greater barriers than text-based ones, highlighting the tension between Europe’s privacy framework and the data-hungry nature of modern AI.
GDPR Enforcement and Uncertainty
While the UK and EU share similar data protection laws—the GDPR was adopted before Brexit—the barriers are more pronounced in the EU. The report attributes this to the EU’s “more aggressive enforcement” and the slow clarification of how data protection rules apply to training and deploying LLMs. For example, the EU experienced a 71-day delay in the web app release of Claude 3 Opus, while Meta had the highest overall rate of delays and non-releases, with more than a quarter (26%) of its models delayed or withheld in the EU and 15% in the UK.
The study notes that the Digital Markets Act (DMA) and the AI Act, which were only enforced or adopted in 2023 and 2024 respectively, have yet to show their full impact. However, the authors warn that these newer regulations could compound existing barriers. The EU is currently considering the Digital Omnibus, a legislative package aimed at making data rules more workable for AI development, but it is also reviewing the EU Copyright Directive and the AI Act’s copyright provisions to protect authors’ rights. If applied too rigidly, these could further limit access to the most advanced AI models.
“It’s important that policymakers in the EU and the UK are calibrated to the risk of regulatory barriers causing delays for their citizens and businesses in accessing the latest AI models,” said John Lidiard, a UK AI policy researcher at Governance AI and one of the report’s authors. “Our report finds that European regulation, primarily the GDPR, led frontier AI companies to sometimes delay model releases or in some cases not release models at all to the EU and UK. Policymakers should consider delays to model access as a factor when implementing and designing AI-related regulations.”
The findings come as European policymakers grapple with balancing innovation and regulation. The EU’s approach has been praised for prioritizing citizen rights, but critics argue it risks leaving the continent behind in the global AI race. The report’s data suggests that the current framework, while well-intentioned, may be creating a two-speed AI landscape where European users and businesses get later access to cutting-edge tools.
This is not the first time regulatory divergence has caused friction. The EU auto sector is split over local content rules to counter China, and similar debates are playing out in digital policy. Meanwhile, a UN panel has warned that the window for AI governance is closing as inequality risks grow, underscoring the urgency for Europe to find a workable balance.
For now, the study serves as a clear signal to Brussels and London: if they want their citizens and businesses to benefit from the latest AI advances, they may need to recalibrate their regulatory approach—or accept that the most powerful models will arrive elsewhere first.
