AI Gigafactories: What is really at stake?

In Europe, the conversation about artificial intelligence has focused heavily on regulation, while its position in the technological race will ultimately be decided by other factors. One way to look at them, reflected in our CEE AI Action Plan, is to think in terms of five pillars: infrastructure, data, human capital, regulation and financing. It is in this context that the Baltic AI GigaFactory project emerges, developed by Poland and the Baltic states, which has recently been joined by Czechia. This is not just a diplomatic gesture from a neighbour, but an important signal that the region wants to play in the top league of European AI infrastructure.

What exactly are AI gigafactories? They are large-scale, specialised computing facilities designed from the ground up for artificial intelligence. Instead of being just traditional server rooms, they bring together very significant numbers of advanced AI chips, ultra-fast networks and sophisticated cooling and security systems to train and run cutting-edge models at scale. In practice, an AI gigafactory can be implemented as a single major campus or as a tightly integrated, distributed infrastructure spanning several locations within a consortium, but operated as one coherent AI computing facility.

The difference between a traditional data centre and an AI gigafactory is roughly like the difference between a local workshop and a jet engine factory. In the former you can assemble something; in the latter you build a power unit that makes it possible to take off in the global race in the first place. Without such facilities, Europe would be forced to rely on infrastructure in the US and China, even in areas that are particularly critical for the EU in terms of security and data protection.

In this context, Czechia’s decision to join the Baltic AI GigaFactory has a very practical dimension. Instead of two competing projects from Central Europe, a single, larger and more substantial consortium is emerging, bringing together Poland, Lithuania, Latvia, Estonia and Czechia. From the perspective of the European Commission, this is a story about genuinely cross-border infrastructure: spanning the Baltic states, a strong Czech industrial base and a large Polish economy. It is easier to defend such a project than several separate initiatives that end up undermining one another.

For our region, the stakes are not just about prestige. We are talking about an investment counted in billions of euros, new jobs for highly qualified specialists, and the attraction of funds and companies that want to develop advanced AI solutions “close to the source of compute”. The creation of a gigafactory in Central and Eastern Europe could help turn the region into one of the natural AI infrastructure hubs in the European Union.

A second dimension is real outcomes in science and innovation. Today, many European research teams and companies have the skills and the data, but hit a wall when it comes to convenient and affordable access to large-scale computing infrastructure. A gigafactory can shift that boundary: it enables experiments with models requiring thousands of processors and enormous datasets, work that until now has usually meant renting capacity from global hyperscalers or moving projects to where such infrastructure is available. The point is not that this was impossible before, but that with a regional gigafactory companies and labs can do it closer to home, under European rules and with fewer frictions around cost, latency and data flows.

The third element is digital sovereignty in the economy, understood in a very practical way: whoever controls the infrastructure decides on the terms under which AI models are developed and deployed. If key computing power remains solely in the hands of a few global providers from outside Europe, we will be subscription-paying users of someone else’s technology. If we build part of this infrastructure in Europe, we become co-owners of the value chain, with greater influence over standards, the location of intellectual property, and tangible benefits for the whole continent.

In the end, a fundamental question remains: does Central and Eastern Europe want to be only a sales market for AI models developed elsewhere, or an active developer and co-host of the infrastructure on which those models are built? The Baltic AI GigaFactory, now substantially bolstered by Czech accession, is one of the first serious attempts to choose the latter path. Whether in a few years we hail this project as a breakthrough or write it off as a missed opportunity will depend on strategic decisions taken both in Brussels and in the participating states. Obtaining the financing and building the factory is just the first step – this infrastructure must be competently managed, the impressive, but still limited resource, put to the most effective use possible. Plenty of challenges await the consortium if its funding bit is successful, but for now let’s celebrate this rare and valuable example of forward-looking and pragmatic political decisions.