The following are the opening remarks given by Lucy C. Cronin, Amazon’s Vice President for EU Public Policy, at the POLITICO AI & Tech Summit in Brussels, on 14th May 2025.

***

AI is the most transformational technology of our time, capable of tackling some of humanity’s biggest challenges and transforming virtually every customer experience.

At Amazon, we’ve been developing and deploying state-of-the-art AI for decades. Today we deploy AI across all of our businesses, with more than 1,000 GenAI applications being built across Amazon. And we provide European organisations and developers with affordable access to AI technology, enabling them to build their own amazing customer experiences. So, while we deploy our Nova models internally, we also make them available to third parties; and Amazon Web Services' Bedrock offers access to the broadest selection of models from different AI companies, including European models such as those of France-based Mistral AI.

According to research by Strand Partners on unlocking AI’s potential, cloud-enabled AI could add more than €386 billion to Europe’s GDP by 2030, meaning that the growth potential in the EU is greater if it can increase its competitiveness in AI. However, many hold the view that the EU lags behind in terms of AI development and implementation, while simultaneously seeking to compete on the global stage. I think it is therefore fair to say that the EU stands at a crossroads: will it harness the enormous potential of AI, and if so - how?

It’s clear that EU leaders have recognized the strategic importance of AI. The AI Action Summit in France sent a powerful signal, and Commission President von der Leyen was very clear in her own Political Guidelines that “Europe cannot afford to fall behind and lose its competitive edge.”

The manifestation of that recognition is evidenced in the planned ‘Apply AI Strategy’ which aims to promote AI adoption in critical sectors such as robotics and transport; the ‘Data Union Strategy’ which seeks to improve the availability of data for AI training; and the ‘AI Continent Action Plan’ which seeks to strengthen AI skills and talent. We are also hearing that the Commission may seek to simplify regulation and enforcement relevant to AI, including modifications to the AI Act.

This political drive could be an important catalyst for development and implementation of AI in the EU, and we certainly encourage European policymakers to continue along this path. However, these strategies must translate into concrete measures that effectively promote innovation; and this requires dialogue with industry practitioners, scientists, and other stakeholders.

And while these plans are encouraging, I note that new requirements that could deter AI development and deployment are being considered, such as those in relation to copyright or sovereign cloud. It is essential that our customers have 1) control over their data and 2) choice as to how to secure and manage this data, and our investment in the European Sovereign Cloud reflects our commitment to these needs.

In addition, it’s crucial that the interpretation of existing legislation, like the AI Act or GDPR, does not create additional barriers for innovators. For instance, the Code of Practice for AI models should focus strictly on facilitating compliance, rather than introducing additional rules. And the nomination of numerous national authorities to enforce the AI Act will no doubt create complexity, as will some Member State moves to implement provisions that exceed those set down in the Act. The risk of this regulatory and enforcement fragmentation is that it will undermine companies’ efforts to scale across the EU, which is, as I understand it, the very objective of the EU Startup and Scaleup Strategy. To sum that up, regulatory enforcement must remain aligned with overarching political objectives.

However, policy can only establish the political and regulatory framework. It is companies – small as well as large – that drive competitiveness and innovation. At Amazon we have a strong track record of actively contributing to European innovation, and we are prepared to do more. Indeed, our commitment to the EU is long-term and deeply rooted. We have invested more than €180 billion in the EU since 2010, and we currently employ more than 150,000 people across the EU.

Research and development are also key areas for enhancing AI competitiveness, and we operate eight development centres across the EU that are dedicated to cutting-edge AI research and development. One such centre sits in Gdansk, the city where we invested in IVONA Software back in 2012 to scale their text-to-speech technology, which became an important part of the voice of Alexa, with now more than half a billion Alexa devices all over the world.

I also note that Strand Partners’ research shows that startups in particular require clear pathways to venture capital and funding, with 56% saying it is critical for faster growth. Amazon manages several venture funds that help promising startups to build their cloud and AI solutions. For example, Amazon’s Climate Pledge Fund supports the development of sustainable technologies.

Let me conclude by highlighting three key steps I believe are necessary to enhance EU competitiveness in AI:

First, Europe needs a supportive and harmonised regulatory framework for emerging technologies. National implementations of the AI Act and legal specifications such as through the Code of Practice for AI models should adopt a default innovation-friendly approach. This would strengthen companies' ability to leverage the EU Single Market, particularly benefiting SMEs.

Second, the EU must address its digital skills gap. According to Strand Partners, 40% of businesses report a lack of digital skills to be an AI adoption barrier. And the private sector also has a role to play here – for example, Amazon’s “AI Ready” initiative to provide free AI skills training to 2 million people globally by 2025, through which we’ve launched free courses about AI use.

Third, AI technology must be accessible to companies of all sizes, which is a precondition for adoption. We contribute to AI’s accessibility through Amazon Bedrock and the other initiatives I’ve mentioned. Policymakers need to play their part by not making access to AI technologies difficult for customers who are seeking that access.