By Isaac Gavaghan, University of Oxford
All opinions are those of their respective author, and are not indicative of the stance of EUPS.
This week, Europe’s leading CEOs in AI and digital industries published a declaration warning that Europe is in danger of falling behind in the global race to develop AI technology.
The declaration was published on the 9th September 2025 by 41 of Europe’s most successful CEOs It calls for ‘decisive action and results in the next one to three years’ from the European Commission, and warns that ‘Our future prosperity, technological sovereignty and collective resilience depend on Europe’s ability to scale and lead in artificial intelligence and critical technologies whilst remaining engaged with global markets.
This intervention comes just seven months after the EU held its Artificial Intelligence Action Summit in Paris in February 2025, at which the EU’s initiative InvestAI was launched. This project aims to open up to five ‘AI gigafactories’, and the EU plans to mobilise 200 billion euros of investment for the project.
A few months later, the European Commission published its ‘AI continent action plan’ on 9th April 2025, which they propose will turn Europe into ‘an AI continent’. The Commission proposes to set up at least 15 AI factories by 2026, which will include the procurement of 9 new AAI-optimised supercomputers. The Commission wrote in their press release that this would ‘more than triple the current EuroHPC AI computing capacity’.
The funding was announced by European Commission President Ursula Von der Leyen, who promised before her election in 2019 that in her first 100 days, she would put forward legislation to regulate AI. Within the first 100 days, Von der Leyen’s commission published a white paper on AI, which aimed to position Europe as a leader in ethical AI. Von der Leyen successfully guided the legislation through the processes of the European Parliament, with the EU Artificial Intelligence Act entering into force on 1st August 2024, to be implemented in two stages.
The first stage of the AI Act, which came into force on February 2, 2025, was to enforce its classification of AI in terms of risk level. Whilst most AI was classified as ‘minimal risk’ and so remained ‘unregulated’, AI that was deemed to be ‘high’ risk was highly regulated, and AI that was considered to pose an ‘unacceptable’ risk was prohibited.
The second stage of the AI Act came into force on August 2, 2025, and regulated General Purpose AI. Providers of such AI were required to publish detailed information about their AI models, and if their models were deemed to pose a ‘systemic risk,’ then they had to be tested, and any problems that arose had to be monitored and recorded.
The EU has acted quickly and has kept ethical considerations at the heart of its legislative approach. However, the geopolitical context of the EU’s legislation is a race to develop more powerful AI faster than the US and China.
China’s State Council published ‘A Generation Artificial Intelligence Development Plan’ on 20th July 2017. The document outlined China’s goals up to 2030. The document reveals the scale of China’s AI plans. By 2030, the report forecasts AI core industry scale exceeding 1 trillion RMB [119 million Euros], and with the scale of related industries exceeding 10 trillion RMB.
However, China remains behind the US in terms of private investment. In 2024, China’s $9.3 billion of private investment into AI was 12 times lower than the figure of $109.1 billion of private investment into AI in the US.
Part of this growth in the US is due to the migration of successful European companies to the US. In 2024, the European Commission published a report by former EU Commissioner Mario Draghi, which contended that whilst Europe has the skills and the resources for AI start-ups and small to medium-sized enterprises (SMEs) to grow, scaling up is easier in the US. In Europe between 2008 and 2021, almost 30% of the start-ups which were eventually valued at over $1 billion moved abroad, the great majority to the US
The Draghi report draws a comparison between the US, where six companies valued over 1 trillion Euros have been founded in the past half-century, and Europe, where no companies valued over 100 billion Euros have been founded in the same period. The report concludes that between 750 and 800 billion euros annually is the necessary investment in AI. This figure is much higher than the 200 billion euros that Von der Leyen’s Commission announced in February 2025. In March 2025, Von der Leyen announced the proposed solution to this funding shortfall, the ‘Savings and Investments Union’, which proposes to ‘develop integrated capital markets’ by promoting the investment of the 10 trillion euros held currently in retail savings accounts across the EU to be channelled into European capital markets. The Commission’s press release details how this could include investments in view of retirement, which could be held in ‘higher-yielding capital-market instruments.
This is a high-risk strategy. In the Economist on the 11th September, the leader argued that the scale of AI investment (which has now reached $3tn globally) will result in some investors losing very large amounts of money. This will either be because AI does not increase productivity by as much as is hoped, or because not all AI companies will succeed, and those investors who back AI companies which do not succeed will lose money
In terms of funding, then, the efforts of the European Commission have as yet been ineffective in their attempts to catch the US and China. However, scepticism regarding the European Commission’s AI legislation is not limited to the funding plans. It is unclear whether the EU’s ethical aims will survive the geopolitical pressure of the Trump administration.
On 11th February 2025, the European Commission withdrew its proposals to ensure that the safety and rights of individuals are not impacted by the use of AI. Whilst the European Commission maintains that these changes were not the result of US pressure, they followed comments by JD Vance attacking ‘onerous international’ regulations attacking Big Tech
Even before this watering down of the European Commission’s AI framework, there were still concerns about the effectiveness of the AI Act. The second article of the act exempts from regulation any use of AI that is justified by national security. This loophole in the legislation could result in infringements of personal privacy and allows the potential use of AI to monitor populations, public spaces, and protests
The implementation of the AI Act, too, has raised ethical questions. The investigation of Corporate Europe Observatory has revealed that the world’s largest AI companies, such as Oracle, Microsoft, Amazon, Google, IBM, and Huawei, have been ‘deeply involved in creating permissive, light-weight standards that risk hollowing out the EU’s AI Act’
The EU’s attempts at AI regulation have so far neither made the EU competitive with China and the US, nor addressed questions over how effective the attempts at ethical governance are in practice. As the declaration of the 41 leading European AI CEOs acknowledged, ‘the direction set by [the European] Commission is right’. It is clear, though, that more needs to be done for Europe to be a future global competitor in AI.
