ABOUT / THE PROJECT
A different way to deploy intelligence.
The f-inference project is making foundation models more efficient, inclusive and sovereign across cloud, edge and IoT.
INSIDE THE PROJECT
Why a new approach is needed
Foundation models can work with language, images, audio, video and time-series data. Yet many deployments still rely on remote cloud infrastructures.
This dependence creates challenges: sensitive information may leave organisational boundaries; specialised computing raises costs and energy demand; and remote processing can add latency to time-critical services.
Four limitations to overcome.
Keep data in your control
Run models where sensitive data is generated, reducing reliance on external infrastructure.
Smaller footprint. Smarter inference.
Reduce energy through model optimisation and resource-aware placement.
Powerful AI for more organisations
Bring efficient foundation models within reach of SMEs and resource-limited settings.
Real-time intelligence
Move inference closer to people, machines and sensors for responsive applications.
THE VISION
Our vision
f-inference envisions a Resource-driven Computing Continuum (RCC) spanning cloud, edge and IoT. Models can be distributed, adapted and coordinated across these tiers instead of running only in distant data centres.
The goal is to place intelligence as close as practical to the sources of data and the people who use it, balancing performance, security, energy use, affordability and ownership of data.
Optimisation strategies
Design-time profiling, compression and device-aware model adaptation.
Adaptive inference
Distributed inference, routing, caching and cross-platform execution.
An open framework
Deployment pipelines and policy-aware orchestration across the continuum.
Open research results
Open tools, datasets, academic publications and policy recommendations.
Project at a glance.
A Research and Innovation Action supported by the European Union.
Field | Value |
|---|---|
Full title | Efficient Foundation Model Inference Across the Computing Continuum |
Grant Agreement | 101298393 |
Programme | Horizon Europe, Cluster 4 (Digital, Industry and Space) |
Topic | HORIZON-CL4-2025-04-DATA-02 |
Type of action | Research and Innovation Action (RIA) |
Duration | 1 October 2026 – 30 September 2029 (36 months) |
Coordinator | Institute of Communication and Computer Systems (ICCS), Greece |
EU contribution | €7,984,003.75 |
Granting authority | European Health and Digital Executive Agency (HaDEA) |