The European Commission this week launched a tender for up to seven AI gigafactories across Europe, backed by $11.5 billion in public funding, in a bid to build sovereign AI infrastructure and narrow the gap with the United States and China. The initiative, unveiled on July 31, 2026, marks one of the bloc's most aggressive moves yet to secure its own computing muscle for advanced artificial intelligence.
The program aims to expand access to computing power for training, fine-tuning, and running sophisticated AI models. It also targets a reduction in Europe's overreliance on AI infrastructure built outside its borders. The commission frames the gigafactories as a cornerstone of its push for technological sovereignty, a theme that has gained urgency as global AI competition intensifies.
A Strategic Necessity
Henna Virkkunen, Executive Vice President for tech sovereignty, security and democracy at the European Commission, framed the move in stark terms. "Access to the raw scale of computing power within AI gigafactories is a strategic necessity for Europe as AI development accelerates," she said. Her remarks underscore a growing consensus in Brussels that Europe cannot depend on foreign cloud providers and chipmakers indefinitely.
The tender is not just about hardware. It is about positioning Europe as an "AI continent," a phrase the commission has used to describe its ambition. The facilities will bring together advanced AI processors, software, cloud technology, high-speed connectivity, and energy-efficient data center infrastructure. That combination, the commission argues, will give European researchers and businesses the tools they need to compete globally.
The scale of the investment is notable. The $11.5 billion in public funding could attract at least $23 billion in private investment, according to the commission's estimates. That would bring the total potential outlay to roughly $34.5 billion, a figure that signals how seriously Brussels takes the AI race. The program was initially planned for five gigafactories, but strong interest from EU member states pushed the number up to seven.
The urgency behind the plan is rooted in a simple calculation. Europe currently imports most of its advanced computing hardware and relies on non-European cloud platforms for heavy AI workloads. That dependency creates vulnerabilities, both economic and strategic. The commission has warned that without domestic capacity, European AI developers would remain at the mercy of foreign suppliers and their pricing, availability, and policy decisions.
Building on Existing Infrastructure
The new gigafactories will not start from scratch. They will complement an existing network of 19 AI factories already operating across Europe. Those facilities have laid the groundwork, but the commission says they lack the industrial scale needed for frontier AI development. Gigafactories, by definition, are industrial-scale plants that produce AI compute, EV batteries, and EVs. The new sites will be built to handle massive workloads.
The facilities are expected to provide computing capacity to a wide range of users. AI startups, scale-ups, small and medium-sized enterprises, industrial companies, researchers, and public authorities will all be eligible to tap into the resources. That broad access is a deliberate choice, aimed at democratizing AI development across the continent rather than concentrating it in a few tech hubs.
The commission also insists that AI systems developed under this program will comply with EU data protection, security, and ethics regulations. That is a significant selling point for European companies that face strict legal requirements. It also differentiates the EU approach from that of the United States and China, where regulatory oversight is often lighter or less predictable.
The existing 19 AI factories have proven that the concept works, but they operate at a smaller scale. They have supported research projects, public sector applications, and early-stage commercial work. The gigafactories are meant to take that experience and multiply it many times over. The commission expects the new sites to handle workloads that would overwhelm the current network, including large-scale model training runs that require tens of thousands of processors working in parallel.
Energy efficiency is another pillar of the plan. The gigafactories are expected to use energy-efficient data center infrastructure, a requirement that aligns with the EU's broader climate goals. Data centers are notoriously power-hungry, and the commission wants the new facilities to set a standard for sustainable computing. That means advanced cooling systems, renewable energy sourcing, and smart power management will likely be part of the technical specifications.
Procurement and Partnerships
The operational details are still taking shape. The EuroHPC JU, the European High-Performance Computing Joint Undertaking, will jointly procure compute access from the selected gigafactories alongside participating member states. That joint procurement model is designed to ensure that public money translates into real, usable capacity for European entities.
Eighteen member states have already signed a joint procurement agreement with EuroHPC JU. The signatories include Germany, France, Italy, Spain, Sweden, Poland, and Ireland, a mix that spans the continent's largest economies and several smaller ones. The breadth of participation suggests that AI infrastructure is seen as a shared priority, not a pet project of a few wealthy nations.
Interestingly, the facilities will not be limited to European hardware. Consortia bidding for the tender can procure processors from suppliers in Europe and other countries. That pragmatic approach acknowledges that Europe does not yet have a homegrown chip industry capable of meeting all its needs. Part of the procurement could also be directed towards European startups and scale-ups, giving local firms a chance to break into a market dominated by giants.
The joint procurement agreement is a practical mechanism, but it also carries symbolic weight. It binds member states together in a common purchasing effort, which is rare in a bloc where infrastructure decisions are usually made at the national level. The agreement means that a startup in a smaller member state will have the same access to gigafactory compute as a large corporation in Germany or France. That equal access is a core goal of the program.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
The commission has also left room for flexibility in how the gigafactories are built and operated. Consortia can propose different ownership structures, partnership models, and technical configurations. The goal is to encourage innovation in the delivery of the facilities, not to impose a one-size-fits-all template. That openness could attract a diverse range of bidders, from established data center operators to newer entrants with fresh ideas.
Chipmakers Sign On
The commission has already secured commitments from major chipmakers. AMD, Nvidia, and Qualcomm have all signed letters of intent with the commission to ensure hardware access for the gigafactory projects. The involvement of these non-European firms is notable given the sovereignty goals behind the initiative. It suggests that Brussels is willing to work with global suppliers in the short term while building its own capabilities over time.
The letters of intent are not binding contracts, but they signal that the world's leading chipmakers see the EU program as a serious opportunity. For AMD and Nvidia, the European market represents a growing source of demand for high-end AI processors. For Qualcomm, the deal could open doors in data center markets where it has historically been less dominant.
The chipmaker commitments are a practical necessity. Europe's own semiconductor industry is still developing, and the commission has acknowledged that it cannot wait for domestic suppliers to catch up. The letters of intent ensure that the gigafactories will have access to the latest processors from day one, rather than relying on older or less capable hardware. That access is critical for training frontier AI models, which demand the most advanced chips available.
The involvement of Qualcomm is particularly interesting. The company has traditionally focused on mobile processors, but it has been expanding into data center and AI markets. A role in the European gigafactory program could accelerate that expansion and give Qualcomm a foothold in a market where it has historically been a minor player. For the commission, having three major chipmakers on board reduces the risk of supply chain bottlenecks.
The tender closes on Nov. 12, likely in 2026, and successful bidders are expected to be named in early 2027. Once contracts are signed, the selected facilities are expected to become operational within 18 months. That timeline is ambitious, but the commission appears confident that the groundwork laid over the past year will pay off.
The selection process will be competitive. Consortia will need to demonstrate technical expertise, financial stability, and a credible plan for delivering compute capacity quickly. They will also need to show how they plan to meet the EU's energy efficiency and data protection requirements. The commission has said it will evaluate bids on a range of criteria, not just price.
A Competitive Landscape
The EU is not acting in a vacuum. The United States has poured billions into AI infrastructure through both public and private channels, and China has made AI a national priority. Europe's response has been slower, but the gigafactory program suggests a new sense of urgency. The commission's own analysis points to the risk of falling behind if Europe does not act now.
The increase from five to seven gigafactories is a telling detail. It reflects strong interest from EU member states, many of which see hosting a gigafactory as a way to attract investment and jobs. The competition among countries to land a facility could be fierce, and the selection process will likely be closely watched.
The program also has implications for the broader tech ecosystem. By creating a network of high-performance computing sites, the EU hopes to foster a virtuous cycle. More compute capacity could attract more AI talent and companies, which in turn could generate more demand for European-made technology. That is the theory, at least.
The timing of the tender is no accident. The global AI race has intensified over the past two years, with major breakthroughs in model capabilities and a surge in demand for computing power. The commission has been working on this program since 2022, when it first floated the idea of large-scale AI infrastructure. The launch this week represents the culmination of that planning.
The $410 million figure often cited in early discussions of the program is now dwarfed by the final $11.5 billion commitment. That growth reflects both the rising cost of AI infrastructure and the increasing political will to fund it. The commission has argued that the investment is justified by the economic and strategic returns, which it estimates could be substantial over the coming decade.
The program is also a signal to the private sector. By committing public money and securing chipmaker support, the commission is de-risking the investment case for private firms. The expectation is that private capital will follow public money, multiplying the impact of the initial outlay. If the commission's estimates hold, the total investment could reach $34.5 billion, a figure that would make the program one of the largest public AI initiatives in the world.
The road ahead is not without challenges. Building seven gigafactories across different member states will require coordination across borders, legal systems, and regulatory frameworks. The commission will need to manage expectations, timelines, and budgets carefully. But the launch of the tender marks a concrete step forward, and the coming months will reveal which consortia are ready to take on the task.