And whether Bulgarian firms will use it at all.

Kristalina Georgieva opened her speech to EU finance ministers in Dublin on 19 September with a personal recollection: her first purchase in hard currency, in a Bulgaria where membership of the European Union still seemed unthinkable, was a can of Guinness. The point she drew from it was that Europe has repeatedly proved capable of transformations that few anticipated.

Two days later, on 21 September, Prime Minister Rumen Radev addressed the AmCham business breakfast in Sofia and set out what Bulgaria intends to offer in the AI economy. In the preceding week, Christine Lagarde and Mario Draghi had each published their own assessment of Europe’s position.

Read together, the four interventions address different parts of the same problem. Kristalina Georgieva is concerned with where Europe should build, and points to where energy is cheapest. Christine Lagarde explains why part of the capacity needs to be European, citing data and continuity of access. Mario Draghi sets out how the build-out might be financed and which uses require European control. Rumen Radev presents Bulgaria as a location for it. Taken together, they suggest that the question facing Bulgaria is less whether to host AI factories than on what terms, and for whose benefit.

What Europe’s leaders are proposing

Kristalina Georgieva argued that Europe needs to be a provider, a builder and an adopter of AI, since an economy of the EU’s size cannot realistically outsource all of it. On energy, her prescription was deeper integration of national grids through an energy union and, in the meantime, siting data centres where electricity is cheapest.

Christine Lagarde, speaking in Vienna, was careful not to argue against imports: importing technology, she noted, is not inherently harmful, and Europe benefits from American investment in AI. She identified, however, three respects in which AI differs from earlier technologies. The first concerns data: a cloud service stores information, whereas a model reads it, and almost half of the firms that considered AI and decided against it cited data protection. The second concerns access: once AI underpins border controls, tax audits, rail operations and payments, any withdrawal of access, or change in its terms, would affect every sector simultaneously. The third concerns the frontier, since in highly competitive sectors such as pharmaceuticals, finance and defence a modest advantage in capability can translate into disproportionate gains.

Her proposed response proceeds in three steps: more European computing capacity, models that are “good enough” for most purposes and run on European infrastructure, and continued access to frontier systems. She notes that the gap between European data centre demand and installed capacity already stands at around 3 GW and is projected to reach roughly 20 GW by 2036.

Mario Draghi’s article in the Financial Times is the most specific of the four, in that it combines quantitative evidence, a principle for deciding which uses require European control, and a financing model already in operation. The EU hosts less than 5% of the world’s AI computing capacity, compared with some 75% in the United States. As sovereignty considerations gain weight, he argues, compute ceases to be fully interchangeable, and much of the infrastructure that stores and processes European data will need to be located in Europe. American operators can supply a substantial part of this through so-called sovereign cloud services, which remain under American ownership; the most sensitive uses, however, would need to run under European control.

This tiered approach is the one set out in the Commission’s Cloud and AI Development Act, proposed on 3 June 2026 and currently under negotiation. The proposal classifies cloud and AI services into four sovereignty levels, ranging from a baseline of self-assessment and robust cybersecurity at Level 1 to full EU ownership and control, EU-cleared personnel, retention of AI inference data within the EU and audits validated by national authorities at Level 4. The level determines which public-sector workloads a provider may bid for, so that non-EU providers would remain eligible for most of the market while being excluded from the most sensitive work.

Mario Draghi also examines what is holding Europe back. Part of the problem lies on the supply side: owing to permitting and grid-connection delays, a data centre takes around 24 months to build in the United States and 42 months in Germany. In his view, however, the principal obstacle is demand. Investors require long-term offtake contracts before committing capital, while European demand remains dispersed across millions of firms, which helps explain why European AI data centre operators such as Nscale sell most of their capacity to American buyers.

His proposed remedy is the pooling of demand. A group of European companies, including ASML, Capgemini and Amadeus, has committed to multi-year purchases of compute from Mistral, intended to underwrite 1 GW of capacity by 2030. He also sets out conditions for host regions: clean energy, a fair contribution to the grid costs that data centres create, and tangible benefits for local communities.

Bulgaria’s value proposition

Rumen Radev told AmCham that companies from developed countries are interested in building large AI factories in Bulgaria because, in their home markets, they face waits of around five years for a grid connection. Bulgaria, he argued, can offer surplus energy, an increasingly green supply, a stable grid and, above all, rapid connection.

The energy claim appears well founded. Bulgaria now operates 5.4 GW of battery storage, and its grid is becoming greener, which would be consistent with Mario Draghi’s clean-energy condition.

Rumen Radev explicitly invited American investment in AI factories, large data centres, robotics and semiconductors. He also stressed that the government’s interest lies less in selling electricity than in the high-technology ecosystem that such centres could stimulate, in education, space, defence, robotics and semiconductors.

Where they meet

On the supply side, Bulgaria’s offer responds directly to the problem the others describe. Mario Draghi’s 42 months in Germany and the five-year waits reported by Rumen Radev’s interlocutors reflect the same bottleneck: permitting and, above all, grid connection. Ireland illustrates where this can lead, having restricted new data centre connections in the Dublin area because the grid could not keep pace.

Mario Draghi’s answer is to build where construction is cheap and fast. He cites Sweden, where an AI data centre costs only about a tenth more than one in China. Bulgaria is positioning itself as a location of this kind, with speed of connection as much a part of the offer as the price of power.

On ownership, Mario Draghi’s framework is more useful than a simple opposition between European and American capacity. By his standard, American-owned capacity in Bulgaria could legitimately serve most of Europe’s workloads. The more pertinent question is whether Bulgaria’s offer is confined to that tier.

Were every AI factory built in Bulgaria to be American-owned, the country would host capacity for Europe’s ordinary workloads but none for the sensitive tier: public-sector records, health data and industrial data, which is precisely the category Mario Draghi identifies as Europe’s remaining sovereign asset. Given that defence features on Rumen Radev’s own list of target sectors, the sensitive tier is unlikely to remain a theoretical consideration.

Who uses it

Bulgaria already possesses a European AI stack, much of it located at Sofia Tech Park. INSAIT develops the models, including BgGPT, an open family of Bulgarian-language models now in its third generation. Discoverer supplies the compute, including a GPU partition of 128 NVIDIA H200 units. BRAIN++, the EuroHPC AI Factory operated together with INSAIT, gives start-ups, SMEs and public bodies access to that compute and to a federated data lake. InnovationAmp, a European Digital Innovation Hub, offers firms subsidised AI and HPC services.

None of this featured in the AmCham address. Yet the ecosystem Rumen Radev expects to develop around new AI factories arguably exists in part already, having grown around a publicly funded research institute and a European AI factory rather than around foreign-owned data centres.

The stack is under European control, although its foundations are not entirely European. BgGPT 3.0 is built on Google’s open Gemma 3 model, was trained using Google Cloud credits and runs on NVIDIA hardware. Its open weights cannot be withdrawn once downloaded, but each new version depends on Google continuing to release its base models openly, which is precisely the risk Christine Lagarde identifies for open models.

The stack is also modest in scale. A 128-GPU partition is well suited to research and to smaller firms, whereas the factories discussed at AmCham would be measured in gigawatts, several orders of magnitude larger.

The more significant gap, however, concerns use. In 2025, 8.55% of Bulgarian enterprises with ten or more employees used AI, compared with 19.95% across the EU, the third-lowest share in the Union. The gap relative to the EU average has widened from 4.4 percentage points in 2021 to 11.4. Among large enterprises, the figures are 26% in Bulgaria and 55% across the EU.

Kristalina Georgieva’s central point is that productivity gains come from changes in how firms operate rather than from the technology alone, and Mario Draghi’s virtuous circle depends on domestic demand financing domestic capacity. Gigawatt-scale factories serving foreign firms that cannot connect at home would not, in themselves, raise Bulgarian adoption. Kristalina Georgieva’s own suggestion is that the state lead by example: in Ireland, four in five firms report that they would be more likely to adopt AI if the state integrated it into its own systems.

The terms

Mario Draghi’s tiered approach and his conditions for host regions, together with the European stack already in place, point towards a possible negotiating agenda for Bulgaria:

  • Equal access to fast connection for pooled European buyers and European AI data centre operators, not only for American investors, since Mario Draghi’s pooled buyers require sites that Bulgaria is in a position to offer.
  • A share of capacity under European control, large enough to host the sensitive tier and linked to the existing European stack at BRAIN++ rather than built anew.
  • Access for Bulgarian firms, with part of the new capacity offered to Bulgarian start-ups and SMEs through BRAIN++ and InnovationAmp, not only to foreign tenants.
  • A fair contribution to grid costs from those who connect.
  • Ecosystem commitments set out in the agreements themselves, rather than in public statements.

Bulgaria’s bargaining position rests on what others currently lack: time to connection. Christine Lagarde’s advice to Europe as a whole, to become a link in the chain that others cannot afford to remove, would seem to apply here as well. Speed of connection gives Bulgaria that position for the time being; the terms it negotiates will largely determine whether it retains it.

Measuring the difference

The AI Strategic Autonomy Framework (AI-SAF) distinguishes between two dimensions that this debate tends to conflate: how much AI capacity a country has, and who controls it. Physical capacity contributes to resilience; control contributes to autonomy. The framework also distinguishes capability, meaning what exists, from commitment, meaning what has been planned or announced.

On that basis, Bulgaria’s battery storage counts as delivered capability, while the ecosystem Rumen Radev describes remains a commitment until it materialises. For the AI factories, the relevant question is therefore not only how much capacity is built, but which tier it serves and who controls it. The Sofia Tech Park stack is also delivered capability. Its open model weights make it resilient, since they cannot be withdrawn, but its autonomy is bounded by the base model and the chips on which it runs.

AI-SAF does not, however, measure use. Control over AI is of limited value unless the economy actually relies on it, and on that question the Eurostat figures already provide an answer: it is the dimension on which Bulgaria most needs to improve.

Kristalina Georgieva is right that Europe has repeatedly proved capable of transformations few anticipated. Bulgaria has built a strong position in energy. What it obtains in return will depend on the terms it sets now, and on whether Bulgarian firms begin to make use of what is already in place.

Sources