Serbia rolls out SAIFA to plug into Europe’s emerging AI Factory supercomputing network

Serbia’s decision to launch SAIFA marks a notable shift in how the country wants to participate in Europe’s artificial intelligence build-out—moving beyond low-cost outsourcing toward deeper integration with the continent’s high-performance computing and sovereign AI infrastructure. For investors and technology stakeholders, the key question is whether Serbia can translate that access into durable capability-building across industry, research and government.

Launched under EuroHPC JU framework

The SAIFA—Serbian Artificial Intelligence Factory Antenna—project was formally kicked off on 14 May 2026 at the Palace of Science in Belgrade. The event took place under the framework of the European High Performance Computing Joint Undertaking (EuroHPC JU), which is central to Europe’s efforts to expand AI-optimized supercomputing capacity.

Serbia’s entry point into the “AI Factory” network

SAIFA is designed as Serbia’s national entry point into Europe’s new “AI Factory” architecture, a continent-wide network intended to provide access to AI-optimized supercomputing infrastructure, advanced datasets, model development tools and specialized computing services. As part of this integration, Serbia’s system will connect directly with the Pharos AI Factory in Greece and the IT4LIA AI Factory in Italy, embedding Serbian institutions into Europe’s broader sovereign AI infrastructure strategy.

Funding scale and consortium structure

Financially, SAIFA is described as one of the largest dedicated European AI infrastructure projects currently underway in Serbia. Earlier project disclosures put total funding at approximately €3.8 million through European Commission mechanisms.

The consortium approach also signals an ecosystem-building intent rather than a standalone research effort. The project is coordinated by the University of Belgrade School of Electrical Engineering and includes partners such as the Institute of Physics Belgrade, Mihajlo Pupin Institute, Institute for Artificial Intelligence Research and Development of Serbia, the University of Kragujevac and Serbia’s Office for IT and eGovernment.

Why it matters for Serbia’s economy and talent pipeline

SAIFA is expected to provide centralized access to computing resources that would otherwise be prohibitively expensive for researchers, startups, public institutions and companies. Beyond cost access, it is positioned as a way for Serbia to move further into Europe’s AI value chain as artificial intelligence increasingly shapes industrial competitiveness and applications ranging from defense technology and healthcare analytics to energy optimization and digital governance.

The project focuses on applications linked to healthcare, energy, sustainability, culture and language technologies, agriculture and public administration modernization. Energy-related use cases are highlighted as particularly relevant within Serbia’s context—covering areas such as AI-assisted electricity system optimization, renewable forecasting, industrial efficiency and grid-balancing models that align with broader energy-transition priorities.

The timing also coincides with Serbia’s expanding ICT sector. The source notes that Serbia’s ICT exports already exceed $4 billion annually, while Belgrade and Novi Sad are emerging regional centers for AI engineering, gaming, enterprise software and deep-tech development. In that framing, SAIFA functions as infrastructure support for a wider economic transformation already underway.

Interoperability goals—and an “AI Gateway” built for broader access

Technically, SAIFA aims to integrate Serbia’s national high-performance computing systems with European AI infrastructures using interoperable standards, secure data-sharing frameworks and containerized computing environments. Coordinators state that it will support model training, benchmarking, testing and deployment across multiple sectors.

A central element is accessibility. Unlike traditional supercomputing systems historically reserved for elite scientific institutions, SAIFA is structured around a centralized “AI Gateway” model intended to simplify access not only for academia but also for startups, SMEs and smaller research teams. The initiative could be significant given ongoing constraints on domestic access to advanced computing capacity and late-stage technical infrastructure within Serbia.

Strategic implications inside Europe—and risks ahead

At a geopolitical level, SAIFA carries symbolic weight: Serbia is not an EU member state, yet participation in the EuroHPC AI ecosystem effectively embeds parts of its digital infrastructure strategy into Europe’s technological framework. The source notes this occurs even as broader accession negotiations remain politically complex.

The project also faces substantial challenges. Building sovereign AI infrastructure requires long-term capital expenditure alongside stable energy supply, advanced cybersecurity protections and continuous hardware modernization. Competition from hyperscale US cloud providers and Chinese AI ecosystems remains intense. Talent retention is another concern: while Serbia continues producing strong engineering and mathematical talent, senior AI expertise still migrates toward Western Europe or North America.

Even with those constraints, SAIFA signals that Serbia intends to compete not only as a low-cost software outsourcing market but as a participant in Europe’s emerging high-performance AI infrastructure layer—an approach that could reshape how local firms develop capabilities if sustained investment keeps pace with demand.