What's next for AI in Europe post-US election?
Introducing AI Now's newsletter on AI and industrial policy in Europe
We’re excited to launch a newsletter on industrial policy and AI in Europe, i.e. public investment, regulatory, spending, and procurement strategies designed to promote the EU’s AI economy. A few times a month, we’ll share our analysis of the latest developments in the European AI market, track Commission initiatives such as on AI factories, and monitor if and how the recommendations made in the Draghi report are being implemented in practice. Europe’s newfound interest in industrial policy, the outsized emphasis placed on AI, and the substantial funds that could be mobilized as a result have the potential to shape European societies for years to come. This is where our work fits in: to ask critical questions about how resources are allocated, the processes by which priorities are determined, and, most fundamentally, to scrutinize the assumptions underpinning this vision.
This monitor is for you if you work in or invest in AI, innovation strategies, tech policy, antitrust and competition, defense, open-source, or public digital infrastructure.
Issue #1
It’s been a busy week in Brussels, from Commissioner hearings to heated discussions about the potential impact of the US election on AI inEurope.
Here’s what we’ll cover today:
Highlights from the post-US election AI landscape
What we learnt from the Commissioner hearings
In other news: AI disillusionment, the environmental impact of AI, the latest news on AI Factories, lessons from past EU funding, and public compute
Highlights from the post-US election AI landscape
As we learnt last week, Republicans will have a governing trifecta in Washington come January, giving President Donald Trump a strong mandate to advance his agenda. Trump’s cabinet picks have triggered “unease” in Europe, but what exactly could the new US administration mean for AI governance and industrial policy in Europe? We’ve been exploring these themes:
(1) Deregulation and China arms race narrative on AI in the US
Trump’s personal views on AI have oscillated between fear and awe, while those of his most influential allies on tech, most notably Vice President elect JD Vance, Trump’s controversial pick for attorney general Matt Gaetz, and Elon Musk paint a picture with multiple, contradictory signals. For example, despite the broad brush de-regulatory stance, Musk has long been in the AI existential risk camp (offering support for California’s controversial SB 1047 law), signaling that these narrow range of safety concerns might continue to capture the attention of political elites. Meanwhile, Vance has previously characterized AI regulation as primarily benefitting Big Tech incumbents, while Matt Gaetz is an outspoken GOP supporter of antitrust legislation.
Here’s what we can expect: Trump has pledged to repeal Biden’s Executive Order on AI, and the future of the U.S. AI Safety Institute (AISI) remains uncertain (read our explanation of how AI is currently governed in the U.S.). The executive discretion driving actions by the DOJ, FTC, and SEC is a key target for the GOP, so we can anticipate further weakening of these agencies. When it comes to promoting global AI norms in international governance fora, the U.S. is unlikely to play a leading role going forward. Instead, expect even more emphasis on the arms race narrative with China, a framing AI Now has been tracking for years. Given the expected negative effects of a Trump administration on climate change, it’s unlikely that the climate and environmental impacts of large-scale AI will gain much attention.
Issues to monitor: the future of U.S. state-level AI legislation; whether the deregulatory shift in the U.S. will impact the implementation and enforcement of frameworks like the AI Act in Europe; and whether Big Tech hawks within Trump’s movement will prevail over the Silicon Valley corporate interests aligning with Trump (the latter being more likely).
(2) Ramping up of competitiveness and sovereignty narrative in Europe
We are already seeing the Commission amplify its rhetoric on sovereignty and competitiveness, likely driven by concerns about the economic impact of potential U.S. tariffs and the possibility of an ‘America First’ president being a less reliable partner in trade and security. Whether this increases the likelihood of collaboration on previously intractable issues, such as joint debt, a complete budget overhaul, Draghi’s vision, and plans for unified defense and security, depends on whether the new U.S. administration brings EU member states closer together, or whether it creates new divisions through transatlantic alliances with individual, Trump-friendly EU members. Musk’s appointment to a government position also raises questions about possible backlash against EU enforcement of regulation such as the DSA against platforms like X.
Issues to monitor: indicators of growing collaboration versus the potential formation of a global nationalist alliance between EU member states and a Trump administration; the U.S. response to EU tech sovereignty initiatives and enforcement actions; and whether Europe is prepared to take the lead on competition and antitrust, should the U.S. deemphasize the issue.
(3) Boost in defense spending in Europe may further securitize AI discourse
One thing that’s certain is that the US election will increase the pressure to spend on defense. The EU has already freed up billions of euros for defense and security by giving member states more flexibility in spending so-called cohesion funds (about a third of the EU’s budget) to support defense industries and military mobility projects, including boosting the production of weapons and ammunition. The AI Now institute has recently published on the dangers of using commercial AI in a military context.
Issues to monitor for the EU: It remains to be seen how the expected boost in defense spending will impact the European AI sector. We’ll be watching this space with a special focus on the use of commercial AI in military contexts.
What we learnt from the Commissioner hearings
Last week, we’ve also seen key Commissioner hearings in the European Parliament that can give some indication of where Europe’s nascent industrial strategy on AI is heading.
The key hearing for the tech-bubble was held on Tuesday evening. The incoming Executive Vice-President, Finnish ex-MEP (EPP) Henna Virkkunen, answered the questions of the parliamentarians for three hours.
Beyond occasional soundbites signaling the importance of the cutting red tape during her administration, the hearing added little substance to what we knew already on the basis of Commission President von der Leyen’s political guidelines and the mission letter she wrote to outline her expectations for Virkkunen. The portfolio assigned to her is daunting, ranging from AI industrial policy to content moderation and defense. Virkkunen will also be a key node in the larger power-play in the European Commission. Coming from a traditionally frugal member-state (just like her superior von der Leyen), she might be a key pawn in the effort to mediate more muscular industrial policy proposals coming traditionally from France and the southern member states.
The AI Factories initiative, Apply AI Strategy and EU Cloud and AI development are the first AI policy initiatives we are expecting from her term. Thanks to a push from a German MEP Geese during the hearing, we learnt that the EU Cloud and AI Development Act will extend computing capacity beyond training to also cover inference and fine-tuning. This highlights the increasing emphasis on infrastructure in the thinking of EU policymakers. However, control of infrastructure is only one step in the effort to tackle the ecosystemic dominance by big tech corporations.
We are expecting: Clear trajectories forward are hard to see at the moment. The current political situation in Europe is chaotic, to say the least, with the German elections looming in February. The traditional power-balance in the European parliament and Council are shifting, which is making cooperation difficult. The political and economic implications of Trump’s return to office will take time to sink in, and the implications for the export-oriented European economy might force difficult political decisions and changes.
Issues to monitor: The AI Factories initiative will likely be the first initiative from Virkkunen’s office. How various industrial policy instruments, the coalescence of data, talent and compute to innovation clusters will survive the political process will be interesting to watch. We are expecting a call for input from the Commission early next year, with a potential proposal released in the late spring.
In other news
AI disillusionment: Multiple sources are now reporting that LLMs aren't scaling as hoped. Open AI, for instance, is reportedly shifting its strategy as the rate of GPT AI improvement slows. Brian Merchant argued that calls for even more investment suggest that the goal is to make AI too big to fail.
Environment: France's state-owned EDF- the world's largest producer of nuclear power - is in talks on deals to power three 1GW data centers. In our EU AI industrial policy report we’ve featured an essay by Fieke Jansen and Michelle Thorne on how large-scale AI may be incompatible with Europe’s environmental goals.
AI Factories: The Commission has received seven proposals for artificial intelligence (AI) Factories, designed to “create a thriving European ecosystem for training advanced AI models and developing AI solutions”, built around the EU's network of European High-Performance Computing (HPC) supercomputers. Commission President von der Leyen promised to deliver on AI Factories within the first 100 days of the new Commission. We’ll be monitoring this closely.
Lessons from past EU funding: Open Future published a new report analyzing EU Research and Innovation funding. A key insight: a disconnect between research and innovation funding and the everyday needs of people in Europe.
Public compute: The Ada Lovelace Institute published an interim report that maps global approaches to public compute.