This grant aims to develop new approaches, tools, and techniques to overcome the limitations of current centralized AI compute techniques, such as energy and compute capacity shortages, lack of chip diversity, data quality, security, and latency issues. The target is to mitigate the EU’s AI compute capacity bottlenecks by enabling alternative decentralized and sustainable AI compute models that exploit diverse hardware architectures and scaling approaches. The focus is on enabling AI data processing to use distributed compute resources across the cloud and edge continuum for all phases of the AI model lifecycle, including data collection, training, fine-tuning, and deployment. Research areas include distributed, decentralized, and federated AI architectures, model compression, handling diverse computing infrastructures, incorporating innovative hardware paradigms, improving data management and privacy, and enhancing energy efficiency and sustainability. Projects must demonstrate the developed solutions in at least two complementary real-world use cases across different domains to show impact and value for Europe’s competitiveness.
| Agency | HORIZON (EU) |
|---|---|
| Region | EUUKother |
| Country | ATBEBGHRCYCZDKEEFIFRDEGRHUIEITLVLTLUMTNLPLPTROSKSIESSEISNOCAILKRNZCHGB |
| Legal form | Company, NGO, Government organization, Other |
| Start date | 17 Nov 2026 |
| End date | 18 Mar 2027 |
| Max funding | Up to 17,500,000 EUR per grant (expected 2 grants awarded under this topic) |
| Budget | 35,000,000 EUR (for this topic in 2027) |
| Coverage | Full details and percentages depend on Horizon Europe RIA rules. Typically, 100% of eligible costs for RIA (Research & Innovation Actions) for beneficiaries. |