EUROPE 2030

Idealism is killing Europe. Idealism is why Europe must exist. Achieving Europe's ideals in the areas that matter will require Europe to become realist in most others. This starts with accepting the diagnosis: Europe needs to recognize that unless it drastically changes its trajectory in the coming decade, it's about to become a third-world country, at the complete mercy of the US and China. The ongoing transition towards increasingly capital-intensive economies driven by the AI revolution means that we may be in the last decade or two where it is possible for Europe to become a sovereign superpower. The symptoms, once the object of scrutiny, reveal how deep some of the challenges are and that the trajectory change Europe will need to take to succeed won't be easy and will come with trade-offs. We outline a path to victory for Europe to be an independent superpower capable of reaching the ideals it aspires to and becoming a safe harbor for everyone who aspires to such a life: privacy-preserving technologies, safe AI development, and a high quality of life for everyone.

By Europe, we refer to a set of countries sharing humanist values which includes the EU, the UK, Norway and Switzerland; we don't refer to a specific jurisdiction. These uniting values include reason, compassion, ethical responsibility and a repeated commitment to human flourishing and dignity. While these countries are part of different political jurisdictions, they have shared challenges and much to win by acting together.

Since WW2, Europe has consistently been a humanist global power. As we are developing civilization-bending technologies such as AI or brain technology, Europe needs to play a role and needs to develop the means to do so. If we want Europe to be able to remain a beneficial power, we need to reestablish strong foundations.

Symptoms & Root Causes

As the thin-veiled fiction of the rules-based order is falling, hard power is retaking the center stage in bilateral negotiations. This was made painfully clear for Europe. It no longer has the means to pursue its value-focused ambitions, when negotiating with China or the US on issues such as supply chain sovereignty or digital rights. The symptoms are obvious for those looking in the right areas:

• Economic stagnation: Europe has seen little economic growth for the past 15 years, with even economic powerhouse Germany showing a decline in exports; the US effectively became twice as rich in nominal GDP per capita as Europe in this short period. While GDP is not a biblical truth, such a drastic change reflects a tremendous loss in relative power. The day-to-day translation of this statistics is dire: European companies unable to match top US salaries lose in competitivity; European defense and intelligence budgets growing less than peers leave us vulnerable to the whims of other countries; EU states unable to maintain balanced state budgets face tantalizing choices regarding protective labor laws and pension systems. These effects reinforce each other, creating a vicious cycle hard to escape.
• Technological dependence: The European modern economy is built on US and Chinese foundations. In hardware like in software. Besides sovereignty implications, this induces ecosystem vicious cycles making building in Europe a superhuman endeavor in comparison with mature tech ecosystems.

Europe's Digital Stack is Built on Foreign Foundations

Market share held by non-European companies, by layer.

How did this happen? Europe missed the golden goose of the tech wave.

Europe Missed the Tech Wave

Europe overwhelmingly missed the unique fruits of the tech wave: large-scale mega-corporations. Ranked by market capitalization, the first European company, ASML, is the 21st most valuable company, about 10x less valuable than NVIDIA, the world's most valuable.

The first wave started in the 1970s and thrived with the boom of the micro-computer and the internet, in large part thanks to the “Fairchildren”, the companies like Intel and AMD founded by employees of Fairchild Semiconductor, enabled by the absence of enforcement of non-compete agreements in California. Fairchild alumni also played a key role in seeding the nascent venture capitalist (VC) industry, i.e. highly risk-tolerant investors who help companies scale much faster than self-funded companies. This early virtuous cycle consolidated Stanford and Silicon Valley position to become a core engine of innovation. Because software can scale drastically faster than industry and hardware, the advantage of having a strong VC industry would provide outsized returns to California in the decades to come. The winners of this first wave now dominate the economic world: Meta, Google, Amazon, Microsoft, and Apple.

A second wave of companies appeared in the 2000-2010s, supported by the increasingly structured Bay Area VC industry: new social media such as Instagram, Snapchat, and Reddit; industry disruptors like Uber or Airbnb; and internet infrastructure companies like Stripe or GitHub. A large fraction of these companies would eventually be acquired by the giants from the first wave.

We’re now in the third wave that started in the late 2010s, characterized by AI as a new general-purpose technology. This includes foundation model companies such as OpenAI and Anthropic, and vertical-focused AI companies such as Waymo or Isomorphic Labs. The ultimate winners from this wave remain to be seen, but the US has a significant headstart.

Europe cannot afford to miss the AI wave. It lost a crown jewel with the DeepMind acquisition in 2014. The window for change is shrinking, but Europe must try. Reasons for pessimism include the massive economies of scale from which AI is benefitting, making it incredibly hard to catch up. Reasons for hope include historical examples of newcomers overtaking incumbents in domains with tremendous economies of scale. The case of TikTok, created in 2016, is a powerful example. The social media industry is one where economies of scale are pervasive and where Facebook (now Meta) was an established monopoly. More recently, DeepSeek managed to efficiently fast-follow US AI companies despite significant computing power constraints, demonstrating the tremendous contribution of top talent to AI innovation.

Europe wasn’t able to capitalize on these waves so far primarily due to three core factors:

• Lack of structured VC and late-stage funding ecosystems. Europe has long lacked a risk-taking early-stage funding ecosystem. As a result, many promising ideas from the European R&D ecosystem either never left the lab or got scaled in the US. This lack of supply of early-stage investors gives European VCs significant bargaining power, leading them to offer significantly lower valuations and much more extractive investment deals. This creates a natural incentive for the best European founders to raise funds or start companies in the US. Even when companies scale to the $100M or $1B level, the absence of European Big Tech with substantial treasuries and the scarcity of late-stage funds capable of acquiring or funding companies for $10B+ creates a trend where successful European companies get acquired by foreign superpowers. Skype, a market leader, was acquired for $8.5B by Microsoft; DeepMind, an AI leader, was acquired for $650M by Google. Redirecting European savings towards filling this gap is critical for European sovereignty.
• Underleveraged global clusters of talent. Clusters of talent create massive positive externalities. The US has the Cambridge area with MIT, Harvard & Northeastern, and the Bay Area with Stanford & Berkeley; China has Tsinghua and PKU. While Europe has Oxford and Cambridge, it needs to further scale their mid and late-stage startup funding ecosystem to match the Bay Area (Dealroom). Building technology hubs is essential to benefit from network effects; by not having hubs nearly as dense as other superpowers, Europe curtails the massively superlinear effects of talent clustering.
• Lack of a large-scale domestic market. Startups benefit massively from a large domestic market as an initial source of demand. Scaling across jurisdictions increases operational complexity and slows growth. To achieve a domestic market of a similar size to the US or China, Europe must unify its markets. This entails enabling a single legal entity recognized across all European countries as is currently being done, limiting national preference policies at the expense of European competitors, and harmonizing tax regimes for startups.

Europe is Punching Way Below Its Weight on AI

Despite its economic weight and top STEM talent, Europe is currently on track to become a minor player in AI. As a striking example, the currently planned European GPU datacenter buildout is smaller than that of the UAE, a country with a 70x smaller population. Adoption of AI among European companies is among the lowest globally. Despite its massive talent pool, Europe has very few frontier AI developers. It lost ownership of Google DeepMind and has a single company which can be considered close to the frontier. This contrasts sharply with the Chinese ecosystem, which has at least three developers close to the frontier and more than ten tier-2 companies catching up fast. The problems compound: low local demand for GPUs makes investors cold for the local infrastructure buildout; the lack of frontier datacenters risks making it impossible for Europe to train its own frontier models, creating permanent dependency.

Europe Has a Big PR Problem and a Localized Regulatory Problem

Europe has a deep public relations problem around regulation. While partly warranted, this extends far beyond the proportional disadvantage of its regulations. The US and China also have significant regulatory overreach. In fact, administrative and regulatory burden for businesses, burden for new firms and tax compliance are all higher in the US and China than in Europe according to the OECD Product Market Regulation; San Francisco, the epicenter of the AI revolution, is an extremely politically dysfunctional city, but does not suffer from a comparable perception problem. This perception problem likely makes it harder to attract talent and reduces the propensity of founders to build in Europe.

Regulatory Burden: A Multi-Domain Comparison

The story is more complex than social media pundits suggest.

Europe United States China France Germany UK

Europe’s Industrial Edge is Under Threat

We see the early signs of China’s industrial base outcompeting European competitors on every axis that matters: quality, price, and innovation. In stark contrast with many European industrial monopolies, which have often been protected by favorable relationships with states, China has implemented policies in key sectors that incentivize ruthless internal competition. The winners that emerge are battle-hardened and crush international competition once they begin exporting thanks to their very low prices. Europe can learn from the way China performs industrial policy, replacing the envy to pick national champions with the urge to catalyze competitive European ecosystems. From such ecosystems, strong global champions can then emerge.

Europe’s Energy Policy Disaster is a Ticking Bomb

The disastrous energy policy of many European countries in the 2010s, most prominently Germany, is a prime example of idealism leading Europe to shoot itself in the foot. By shutting down more than 50GW of capacity, including >20GW of nuclear reactors, without appropriately scaling other energy sources, Germany created a tremendous dependency on Russia and the US and induced a massive energy price hike. As a result, Europe’s power generation is roughly the same as it was in 2000. Energy is the lifeblood of modern society. Cheap energy enables manufacturing to thrive, is a key bottleneck in the race to AI dominance, and tremendously increases international influence.

Total Electricity Generation

China surged past 10,000 TWh while Europe remained flat around 4,600 TWh.

Europe’s Outsourcing of Its Defense is an Existential Risk

Since the 1990s, many European powers have outsourced their defense, assuming alliances would remain stable. By spending significantly less as a % of its GDP than other superpowers, it has fallen behind on many key military technologies. By failing to get the economies of scale that a truly European defense industry would provide, it has outsourced its supply chain to other countries. The result has been a continuous decrease in deterrence, leading to the unpunished 2014 Russian invasion of Crimea, followed by the full-scale invasion of Ukraine in 2022.

Defense Spending as % of GDP

Where Defense Supplies Come From

The Intelligence Gap is Technology, Not People

Europe collectively employs ~114,000 intelligence personnel — roughly matching the US core IC (~100,000 government employees). But the US spends ~6x more per person.

Foundations: The Remedies

Leaving "Everything Bagel" Europe and Embracing Realism

The New York Times columnist Ezra Klein coined the term "Everything Bagel" to describe the phenomenon of applying highly stringent ethical values to all projects, at the expense of achieving the core project objectives. European idealism has translated into a similar epidemic. Environmental reviews and court challenges have drastically slowed the development of renewable energy; environmental concerns have stalled the installation of air conditioning, leading to hundreds of thousands of preventable deaths from heatwaves. To succeed, Europe must become laser-focused on achieving the core objectives of its policies. It is behind on so many sources of hard power that it can no longer afford to systematically impose on itself harder ethical constraints than other superpowers. It must pick its battles and regain its strength before it can afford to be ambitious in how it achieves its goals.

Making European Governments World-Class

Europe's governments are significantly involved in managing society. This can be a strength, but only if those governments are highly competent, especially in expertise-heavy domains like technology policy. A few key policies could facilitate this:

• Increase government wages to match the private sector. To attract and retain top talent, compensation must be competitive. Considering the scale of government spending, the return on investment from hiring more competent decision-makers would be immense.
• Encourage specialization in expertise-heavy roles. Technology policy benefits from deep understanding, which requires specialized career paths within government, rather than promoting generalists.
• Let talent thrive and take risks. Technical experts who join government need to be given significant ownership and respect for their expertise, akin to methods in the private sector, to prevent them from leaving due to an inability to execute their mandate. Importantly, they must be able to take non-consensual decisions and be rewarded if these pay off.
• Implement periodic reviews of existing regulations. Europe, and the EU in particular needs a systematic process to review all regulations for necessity, redundancy, and unintended consequences. This could be inspired by institutions like the UK's What Works Centre, ensuring the legal apparatus does not become obsolete or counter-productive.

Furthermore, AI must be implemented widely in government to massively increase efficiency and deliver better, more customized services to citizens. This requires avoiding overly burdensome deployment conditions that stifle adoption and ensuring civil servants are continuously trained to recognize and adopt useful AI applications as they emerge.

Resolving Europe’s PR Problem

Jointly with a policy aimed at more systematically making European governments competent, Europe should aim to remove symbols of regulatory mistakes. This is an irreplaceable part of a policy aimed at making Europe a stronger option for entrepreneurs and top talent:

• Remove symbols of regulatory failure as soon as possible. Europe’s regulatory reputation is disproportionately shaped by a few highly visible cases. Identifying and rapidly removing these is a low hanging fruit for its impact on perception. A prominent example is removing and replacing the cookie banner with a browser-level standard that achieves the same GDPR goal, i.e. letting users deliberately choose what data to share, without the massive user experience cost and reputational damage.
• Appoint highly technically savvy representatives on technical topics. To combat the perception of an incompetent bureaucracy, external-facing policymakers on topics like AI, health or biotechnology must be deeply knowledgeable and credible.

These efforts should go along with policies that make Europe a better home to fast-growing companies that are the defining feature of economic superpowers in the 21st century.

Market Unification

To equate the depth of the US and Chinese markets, European market unification is a necessity. The Draghi report on EU Competitiveness goes into great depth on the problems and solutions. Without attempting to replace this report, we highlight some of the most consequential measures:

• Unleash private capital: Complete the EU Capital Markets Union (CMU) to create a deep, integrated pool of capital. This includes measures needed to convert Europe’s vast savings pools into investment in venture capital and growth equity to fund the next generation of European companies. The untapped potential is massive: while Europe has 5x less pension fund assets and 7.5x less VC funding deployed than the US, it has 1.7x more household savings per year.

Europe's Capital Allocation Gap

Europe saves 65% more than the US. But that capital never reaches startups.

Europe United States

• Market unification: Create a single market for goods, services, and labor, removing the national preferences and frictions that still fragment the continent and slow down scaling companies. The IMF estimates that existing barriers are equivalent to a 44% tariff for manufacturing and 110% for services.
• Debt unification: Move towards a common debt market at the European level, as proposed by O. Blanchard & A. Ubide. A common European bond market would dramatically increase liquidity, strengthening the euro’s position as a credible global reserve currency and lowering borrowing costs for all member states.

Along with paving the way for high-growth companies, Europe should invest in protecting its core historical strength in industries and manufacturing. This requires investing in energy, the lifeblood of societies, but also rethinking the approach to economic policy.

Pumping Energy in Europe’s Veins

The failure of European energy policy in the 2010s is such that Europe cannot afford any more idealism in that realm. To remain competitive and reduce its massive external dependence, it must scale energy generation of all kinds as quickly as possible. This will entail trading off some of its values and require a return to realism in the short-term:

• Scale solar, batteries, and wind turbines. Catching up to the rate of China and the US will require removing environmental reviews and other permitting regulations that drastically slow energy and renewable growth in Europe, often by years. As an example, wind turbine projects can take up to nine years to receive permitting while the project takes two years to build. Importantly, solar power has been a renewable technology that has reached extremely low prices, enabling a wide deployment and rapid scaling of energy supply. The bottlenecks to deployment are largely regulatory and related to permitting. As a reference, China has installed more than four times more solar capacity in the first half of 2025 than Europe in 2025.
• Scale fossil fuels. Fossil fuels are a necessary transitional fuel to ensure grid stability and energy strength in the short to medium term. The voluntary decrease of energy supply driving an increase in energy prices has been a major driver of tearing apart the German economy, which used to be the most thriving ones in Europe. By making the entire German industrial basis significantly less competitive, it is a major pull on European relevance.
• Scale nuclear. Nuclear power has been a historical advantage for Europe, can be made cheap and provides substantial sovereignty thanks to its energy density which allows long-term stocks. To rebuild this capability, a pragmatic approach could be to conduct joint ventures between national suppliers and South Korean firms, who have demonstrated the ability to build new reactors in as little as four years. Additionally, shortening the regulatory review will be key to rebuilding domestic capacity, as diagnosed by the recent analysis from the UK government of the root causes of nuclear decline, which found that it induced £700M spent to save a few fish per year on average.

The more we scale solar, the less we need other sources. But we are installing 5x less per year than China and cannot afford to rely only on renewables alone to remain competitive in the short run.

Strengthening the Industrial Basis

Restoring Europe’s competitiveness in the industrial realm requires action across several fronts:

• Foster competition: Numerous European industrial leaders are exposed to little competition, sometimes because of government expenditures shielding them from such competition, like in the case of the tariffs on Chinese vehicles. While such protections can make sense to protect a nascent rising industry, it is simply postponing the inevitable decline of industrial leaders at great cost to the taxpayer if not accompanied by a credible strategy towards increasing their competitiveness. A key focus of industrial policy should be to support thriving and competitive ecosystems rather than picking winners who, shielded from competition, are quickly outpaced internationally. A key ingredient essential to the success of Chinese exports is the ruthless competition local companies have been exposed to in their growth stage, leading the winners of the competition at the consolidation stage to be extremely competitive on international markets.
• Invert the trend: Europe’s most competitive industrial players, including German manufacturers, are losing ground internationally. Rather than propping up incumbents with subsidies and tariffs, Europe should leverage one of its strong advantages: the rapid industrial growth in Eastern member states like Poland, Romania, and the Czech Republic, which have built significant manufacturing capacity at competitive costs. Joint ventures pairing Western European engineering and brands with Eastern European production capabilities could create industrial champions more competitive on quality and price. Building joint industrial defense capacity is a strong area to target policies aimed at incentivizing such positive-sum relationships.

Manufacturing Production: Western Decline, Eastern Rise

Industrial production index (2021 = 100), seasonally adjusted.

• Make energy cheap: One of the key costs of productions of most manufacturers and industrial players is energy. As a result, driving energy prices down is a key enabler for national champions to be competitive internationally. Symmetrically, unpredictable or high energy prices tend to kill many manufacturing industries, given how low-margin manufacturing typically is and that it requires substantial upfront investments that become too risky if energy costs are unpredictable. Thus, European countries should have among their top priorities to be scaling production of energy, including power, through all types of sources and make their best effort to keep prices low in the long-run.
• Cut red tape: Numerous regulations and measures that were developed in an era where international competition was less significant and where velocity mattered much less than it does today prevent Europe from being an attractive place to build factories, invest or scale up a manufacturing company. Robotics is one area benefitting tremendously from trial and error and whose growth in Europe might be curtailed by burdensome regulatory requirements from the EU AI Act, the Machinery Regulation or GDPR, that don’t exist in China or the US and enable faster development.
• Reduce external dependence on minerals: By making it essentially impossible to open new mines on its soil by a mix of permitting, environmental review and protective measures, Europe effectively delocalizes mining of many core raw material and minerals critical to its supply chain, thus outsourcing the negative externalities of mining in countries most often less equipped to deal with such externalities. To change this trend, Europe will need to relax successfully all of these policies and bear the cost of technology development on its own soil, regaining elements of sovereignty by the same token. Concretely, an effective enforcement of the Critical Raw Materials Act across member states would be a strong step.

Critical Minerals: Europe Fails Its Own Mining Target on Most Strategic Materials

EU domestic extraction vs. imports for 8 Strategic Raw Materials under the CRMA.

Securing Europe

The current fast evolution of military strategy we’re living in are presenting an exceptional opportunity for Europe to build back its lost deterrence by focusing on being excellent in a few key strategic areas:

• Drones & military technology: With >60% of Ukrainian casualties on the Ukraine-Russia front coming from drones, it’s clear they are reshaping warfare. The core of Western military doctrine, focused on a few highly expensive units, is being questioned. Whether it is €15M ballistic missiles that can take down >€10B aircraft carriers, €250k marine drones that can take down €100M ships, or €1k drones that can take down >€1M tanks, the problem is pervasive. The novelty of this evolution presents an opportunity for European defense to offset some of its weaknesses in other domains by building strength in this one. Europe must rapidly innovate and adopt new military technologies in these areas, leveraging its proximity with Ukraine to its advantage.
• Nuclear Triad: In a world increasingly brutal, maintaining a fully sovereign supply chain for the nuclear triad is essential. While France has emphasized this, the UK may need proactive policies to restore it. As an example, the UK Trident missiles are leased from the US & the UK has failed to perform a successful Trident launch since 2012.
• Intelligence & cybersecurity: The Snowden affair revealed the extent to which European powers were penetrated by foreign intelligence agencies. To maintain an IP edge in any immaterial domain and have credible deterrence in the cyber realm, reducing this asymmetry is crucial. This requires both substantially increasing resources for intelligence agencies and improving the talent funnel from top universities. As demonstrated by Mythos being capable of finding thousands of zero-days vulnerabilities, this domain is increasingly affected and dominated by AI, furthering the need for strategic independence.

Fostering Innovation

Europe’s innovation business ecosystem is weak compared to that of the US. Despite the significant basic research and R&D capabilities of European countries, Europe struggles to convert that into successful companies. Ten years ago, most European countries barely had a startup ecosystem. Over the past decade, several countries such as France have grown a promising ecosystem, reaching over 30 unicorns. While the trend is positive, Europe is still lagging behind in part due to remaining frictions that have to be addressed to increase Europe’s competitiveness in highly competitive sectors:

• Reduce labor costs: State-induced labor costs remain very high, making wages uncompetitive internationally and hindering talent attraction. The net cost to the employer for an employee to receive $150k net is $213k in Austin and $354k in Paris. These differences should be reduced, especially in high-growth tech sectors where employees already have significant bargaining power and do not need the same level of protection as in other industries.
• Reduce employment frictions: Highly protective labor laws, designed for a low-skill job era where the employer-employee bargaining relationship was highly asymmetric, penalize dynamic companies. Making it difficult to lay off employees during downturns makes startups riskier and encourages slower, less ambitious growth. Similarly, long notice periods for managers can cripple an early-stage company’s agility at critical moments. Targeted measures that enable high human capital sectors such as tech to be more agile would be transformative for building successful companies in Europe.
• Introduce EU Inc & cut red tape: Enable the incorporation of a single legal entity recognized by all European countries and eliminate uselessly complex processes, such as the German requirement for in-person notary readings of investment deals, that add friction and slow down capital allocation.

To further increase its odds, Europe should lead a policy of high-skill visa enablements, streamlining the process and making it extremely easy for international talent to enter. It is a cheap intervention that can differentiate Europe, in a context where visas are a major issue preventing talent from easily working in existing superpowers.

Startups are inherently risky for employees, but that risk becomes tolerable when a dense local ecosystem of other active and hiring startups provides a safety net. As a result, for top talent to be more willing to take that risk, a full ecosystem is typically needed and facilitated by physical clustering of startups, VCs, talent and a highly dynamic employment market. Hubs like London and Paris would benefit from defining targeted AI zones for AI companies and housing enabling network effects to be concentrated in a few square kilometers rather than split across these megacities, replicating the benefits that San Francisco gains from its size.

Riding the AI Wave

Developing the AI Infrastructure

As AI is about to reshape the economy, Europe is currently barrelling towards being an irrelevant player on the international stage. As a point of reference, US hyperscalers aim to build 80-100GW of frontier AI datacenters by 2030. Europe currently has only 2-3 GW planned. To remain sovereign, Europe should aim for a quarter of that, 20 GW, i.e. roughly 19 million next-generation GPUs, requiring approximately €1T of investment. This is achievable, but requires streamlining four key inputs simultaneously:

• Unlock grid connections. Europe already has tens of GW of retired power plant grid connections sitting idle[1]. Germany alone has >30 GW of retired capacity, of which at least 10 GW is uncommitted and could be leveraged for datacenters. Each country should replicate France’s initiative of systematically identifying adequate sites for AI datacenters, with fast-track grid permitting to accelerate deployment.
• Unlock power generation. Grid connections are the veins; power generation is the heart. The first source of novel power generation is to max out the existing power capacity. Beyond capacity coming from net exports such as France and Sweden, Germany has 32 GW of installed natural gas capacity running at only 28% capacity factor. Simply scaling this to 80%, typically feasible for gas plants, would unlock over 15 GW of power. Making sure the pricing doesn’t significantly hurt non-AI industries may be feasible using targeted policy measures that ensure any cost surplus is incurred by datacenters. To complement existing capacity, a solar/battery plus gas-powered buildout will likely be necessary[2]. Spain is particularly well-suited for performing parts of this buildout, leveraging its 3 GW of retired coal plants and its excellent sun exposure.
• Fix permitting & legal uncertainty. The current permitting timeline for energy infrastructure in Europe is 4-9 years, which is incompatible with a realistic catch-up timeline. Grid permissions and environmental assessment must be streamlined and fast-tracked. Detailed proposals of Special Compute Zones have been made to address these issues in a unified way. A €1T buildout will require a significant number of skilled construction and operations workers, which may require immigration fast-tracks for datacenter technicians. Finally, enabling a buildout of that magnitude to happen requires legal certainty on elements that are important to AI development such as copyright law[3]. Given the economic and national security implications of AI, recently demonstrated by frontier models’ capability to find thousands of software vulnerabilities, an emergency legislative procedure under Article 122 TFEU may be warranted for these.
• Catalyze capital. High-tier GPU demand far exceeds the supply in the current market, representing a lucrative investment opportunity for private capital. The value of deploying AI has increased to a level such that even previous-generation hardware H100s have increased in rental value. As a result, the core policy goal should be to leverage public action to catalyze private investment. Tools could include sovereign guarantee on a fraction of loans used for the buildout, tax alleviations to make European investments competitive with foreign datacenter buildouts, or the enablement of accelerated depreciation.

Managing Risks to National Security and the Human Species

Europe has a strategic interest in ensuring AI is developed safely and reliably. This means preventing rival powers from infringing on our sovereignty by developing uncontrolled superintelligence, but it also means contributing to global safety to avoid large-scale risks, no matter where they originate. The EU’s work on risk management, such as the AI Code of Practice, is a helpful start and should be built upon. To be effective, this work must be backed by a credible AI position and hard power.

Conducting Industrial Policy Along the AI Value Chain

Europe must build expertise and capacity in all key areas of the AI supply chain, even if full autonomy is out of reach. Having companies like Zeiss and ASML in Europe is a tremendous lifeline that significantly reduces the leverage that other superpowers have on Europe. Reinforcing areas of strength and reaching a minimal level of production in areas of dependency should be a key part of the European industrial policy:

• Machine-tools (EUV): Europe has a monopoly via the Netherlands-based ASML produced of EUV technology which is considered one of the most sophisticated technologies humanity has ever developed. This critical advantage must be preserved and expanded into a broader ecosystem of innovation in semiconductor manufacturing equipment. Beyond supporting ASML, it should be a priority to try to turn this single company advantage into an ecosystemic advantage. Europe should encourage the development of EU-based EUV companies such as Lace pushing the frontier of the technology that could potentially benefit from Joint Ventures with ASML to innovate on the basic technology. Otherwise, there might come a time where Europe loses its key edge in this technology.
• Chip assembly: Taiwan has a monopoly of the assembly of chips through TSMC, closely followed by the South Korean Samsung. While being fully autonomous in building frontier AI chips at the relevant volume is out of reach for Europe, building at least one fab capable of assembling frontier AI chips in Europe would be highly relevant to building expertise in the domain and reducing the pressure that Europe could be subject to without manufacturing capacity in the domain.
• Chip design: NVIDIA’s lead is formidable, but its primary moat is its CUDA software stack, not just hardware. Europe should encourage an ecosystem of competitors, focusing on highly efficient or specialized niches where new players such as Groq or Cerebras can thrive.
• Foundation models: Europe possesses top talent. The government can facilitate the formation of new world-class AI labs by connecting top researchers and providing backing for crucial datacenter access. Facilitating events that help such talent to get to know each other and offering backing for permissioning regarding AI datacenter buildout may increase the odds that new relevant AI developers emerge in Europe. There is no talent bottleneck to the EU becoming a frontier AI center as dynamic as China is.

Enabling AI Adoption

The future competitiveness of the European economy will depend on its ability to integrate AI into its core industries. Removing bottlenecks to adoption should be a key objective of EU AI policy in the coming years to ensure this happens.

• Minimize the burden of the EU AI Act and GDPR. The broad, use-case-focused parts of the AI Act risk creating a significant compliance burden that slows adoption. Its impact must be closely monitored, and its scope potentially reduced to minimize the drag on industry. Similarly, GDPR must be interpreted in ways that don’t have a chilling effect on AI adoption in the EU.
• Accelerate AI for industry & life sciences. European leading companies are global leaders in industry and in life sciences, both of which have a high bar for reliability. Mainstream AI models are not yet reliable enough for these use cases. Industrial policy should focus on enabling the development and deployment of AI systems reliable enough that they can be deployed in these core sectors, turning a European strength into a technological advantage.

Looking Ahead

Idealism is Europe’s gift that makes it special. Idealism is Europe’s curse that renders the path hard. Navigating this duality will require to reckon the most self-destructive aspects of it and embrace pragmatic idealism. From where we stand, all scenarios are extreme, whether in means or in outcomes. Our current trajectory is extreme in outcomes: Europe is at risk of becoming irrelevant in the next decade on all the key features of hard power, whether economic, military or industrial might. Pulling ourselves out of this scenario will require an appropriate level of urgency, extreme means, and moving far from the characteristic consensus-driven policy habitus dominating large parts of our countries. This is the price of two decades of excessive idealism. This is the path to make Europe’s idealism count again.