D2.2 Partner Collaboration and Liaison Framework

Over the last five years Serbia has built a national AI platform whose performance has grown from 5 to 37 PFLOPS, with a further 22.5 PFLOPS system contracted for 2027. It has a second-generation national AI strategy running to 2030, ethical guidelines, a dedicated AI law in preparation, around three hundred AI companies and a dense layer of science and technology parks, hubs, clusters and funds. The Serbian AI Stakeholder and Market Landscape Report (D2.1) documents these strengths in detail. However, Serbia faces economic, structural and institutional deficiencies that we would like also to mention. Gross expenditure on R&D stands at 0.94% of GDP, well under half the EU level. One Serbian enterprise in ten uses AI, about half the EU share. Venture capital intensity is 8% of the EU benchmark. AI firms, enablers and computing capacity are concentrated in Belgrade and Novi Sad. The binding constraint on Serbia’s AI ecosystem is therefore no longer capacity. It is connectivity: between research and deployment, between technology developers and holders of domain data, between national programmes and European infrastructure, and between Belgrade and the rest of the country. D2.2 responds to this diagnosis with three complementary instruments, as foreseen in the Description of the Action: A strategy for aligning national R&D and DIH investment with EU-level AI Factory goals (Chapters 3–5). It sets out five strategic lines — federated compute access, data readiness, lab-to-market continuity, regional diffusion, and skills with trustworthy AI — together with concrete alignment mechanisms, a roadmap to M36, indicators and risks. Coordination guidelines between Serbian partners (Chapter 6). They define how the eight SAIFA partners present a single, coherent interface to the ecosystem: one relationship owner per stakeholder, a shared relationship register, a common user request pathway and a predictable reporting rhythm. A framework Memorandum of Understanding with national initiatives (Chapter 7 and Annex C). It is tiered, non-binding and designed to mirror the Activity Sheets of the agreements with the AI Factories, so that national and European cooperation follow the same logic. The first signature is due by M16 (Milestone 16). The international dimension — the Extended Partnership Agreements with the Pharos AI Factory (Greece) and the IT4LIA AI Factory (Italy), which convert the Letters of Intent into operational instruments — is summarised in Chapter 8, with the negotiated texts in Annexes A and B.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-03
DOI
https://doi.org/10.5281/zenodo.23917673
Primary Topic
University-Industry-Government Innovation Models
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

D2.2 Partner Collaboration and Liaison Framework

АЛЕКСАНДАР ВУЧКОВИЋ, Ernest Vlačić, Katarina Đulić, Marina Šestović
Zenodo (CERN European Organization for Nuclear Research)
University-Industry-Government Innovation Models
article

D2.2 Partner Collaboration and Liaison Framework

АЛЕКСАНДАР ВУЧКОВИЋ, Ernest Vlačić, Katarina Đulić, Marina Šestović
article en

Abstract

Over the last five years Serbia has built a national AI platform whose performance has grown from 5 to 37 PFLOPS, with a further 22.5 PFLOPS system contracted for 2027. It has a second-generation national AI strategy running to 2030, ethical guidelines, a dedicated AI law in preparation, around three hundred AI companies and a dense layer of science and technology parks, hubs, clusters and funds. The Serbian AI Stakeholder and Market Landscape Report (D2.1) documents these strengths in detail. However, Serbia faces economic, structural and institutional deficiencies that we would like also to mention. Gross expenditure on R&D stands at 0.94% of GDP, well under half the EU level. One Serbian enterprise in ten uses AI, about half the EU share. Venture capital intensity is 8% of the EU benchmark. AI firms, enablers and computing capacity are concentrated in Belgrade and Novi Sad. The binding constraint on Serbia’s AI ecosystem is therefore no longer capacity. It is connectivity: between research and deployment, between technology developers and holders of domain data, between national programmes and European infrastructure, and between Belgrade and the rest of the country. D2.2 responds to this diagnosis with three complementary instruments, as foreseen in the Description of the Action: A strategy for aligning national R&D and DIH investment with EU-level AI Factory goals (Chapters 3–5). It sets out five strategic lines — federated compute access, data readiness, lab-to-market continuity, regional diffusion, and skills with trustworthy AI — together with concrete alignment mechanisms, a roadmap to M36, indicators and risks. Coordination guidelines between Serbian partners (Chapter 6). They define how the eight SAIFA partners present a single, coherent interface to the ecosystem: one relationship owner per stakeholder, a shared relationship register, a common user request pathway and a predictable reporting rhythm. A framework Memorandum of Understanding with national initiatives (Chapter 7 and Annex C). It is tiered, non-binding and designed to mirror the Activity Sheets of the agreements with the AI Factories, so that national and European cooperation follow the same logic. The first signature is due by M16 (Milestone 16). The international dimension — the Extended Partnership Agreements with the Pharos AI Factory (Greece) and the IT4LIA AI Factory (Italy), which convert the Letters of Intent into operational instruments — is summarised in Chapter 8, with the negotiated texts in Annexes A and B.

Zenodo (CERN European Organization for Nuclear Research)
Metropolitan University (RS), NOVAMINA Innovative Technology Centre (HR)
Openalex Percentile: Top 11%
University-Industry-Government Innovation Models
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon