Explore our projects

SHARC

Enabling a disaggregated, multi-tenant and AI-driven 6G RAN over a secure, hardware-accelerated compute continuum.

Official Logo

Multi-Tenant, Disaggregated 6G RAN Over a Secure Hardware-Accelerated Compute Continuum

Project Details

Funded under: Horizon Europe, Smart Networks and Services Joint Undertaking (SNS JU), Call HORIZON-JU-SNS-2025-01-STREAM-B-02

Project ID: 101290994

Start Day: 01/09/2026

Project Duration: 36 months

Tags: 6G, Open RAN, Hardware Acceleration, AI-Native RAN

Social Media

Learn more about this project

Flaticon-internet Linkedin Twitter Youtube

Let’s Work Together for Development

Let’s explore how we can support your next project.

Head Office
208, Kifisias Avenue, Chalandri, 15231
Technical Branch
32, Vosporou, Elliniko, 16777
Contact Us
+(30) 213 088 9302 research@fourdotinfinity.com
Get in Touch With Us

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.

The Story

The move from 5G to 6G is not only about faster speeds. It also demands a rethink of how the radio access network (RAN) is built and operated. The RAN is already the biggest consumer of energy in a mobile network, with studies placing it at roughly 70–75% of the total. However, scaling today’s architecture to 6G traffic is economically and environmentally unsustainable.

Furthermore, 6G workloads such as massive MIMO, beamforming, and real-time AI go beyond what general-purpose CPUs can handle. Today’s hardware acceleration is vendor-specific and statically integrated. Similarly, Open RAN has brought modularity, but also more management complexity and a larger attack surface, while AI in the RAN still lacks a telecom-grade way to be deployed and managed.

SHARC takes on these intertwined challenges of energy efficiency, orchestration, AI lifecycle management and trust, to define a RAN that is intelligent by design, sustainable, and aligned with Europe’s goals for digital sovereignty and the green transition.

Project Overview

SHARC develops the Intelligent Acceleration and Orchestration Framework for next-generation 6G RAN. Specifically, it creates a disaggregated, AI-driven, shared, and accelerator-enhanced architecture. To achieve this, the framework relies on four core technical pillars:

Profiling key RAN functions across GPUs, FPGAs, CPUs and neuromorphic hardware, and developing accelerator-aware xApps and dApps that offload heavy tasks such as beamforming, channel coding and precoding.
Investigating neuromorphic and analog processors for energy-efficient radio processing.
A micro-orchestrator that bridges RAN software and heterogeneous hardware, combined with a hardware-based trust fabric (remote attestation, hardware roots of trust and trusted execution) for trust-aware scheduling.
Extending the O-RAN SMO and RIC to be aware of the accelerator fabric, turning operator intents into policies enforced across the RAN.
An MLOps pipeline for automated lifecycle management of AI models in the RAN, supporting different paradigms such as reinforcement learning and generative AI.

The 6G Sustainability Barrier: Energy & Architectural Limits

The results are validated in four use cases, covering non-linear physical-layer processing, regenerative 6G non-terrestrial network payloads, accelerator-aware execution of RAN functions in the RIC, and orchestration of the RAN across the cloud–edge continuum. SHARC targets TRL 4 by the end of the project.

The project is coordinated by Space Hellas and brings together 14 partners from industry, SMEs, research centres and universities across Europe.

Our Role in SHARC

Four Dot Infinity (FDI) leads the development of accelerator-aware RAN workloads in SHARC, bringing its AI expertise to the project.

FDI leads the work package that adapts RAN functions to run efficiently on heterogeneous hardware, from profiling the workloads to benchmarking and optimising their execution.
FDI leads the selection and profiling of RAN workloads and defines the APIs of the Hardware Abstraction Layer (HAL). It builds an AI-driven pipeline that optimises how tasks are balanced across accelerators and conventional functions, and how fast offloaded tasks execute.
FDI leads the benchmarking of AI pipelines with and without HAL deployment options, comparing execution time, model accuracy and energy consumption. The outcome is a set of blueprints showing when and how accelerators should be used.
FDI supports the integration of profiled workloads into accelerator-aware xApps and dApps, contributes to the validation of HAL-based offloading of L1/L2 tasks (such as beamforming, FEC and scheduling) to GPUs and FPGAs in Use Case 3, and takes part in dissemination, exploitation and liaison activities with other SNS projects.

Impact and Expected Outcomes

SHARC delivers a validated foundation for an energy-efficient, intelligent and trusted 6G RAN.

Key Results
A unified hardware abstraction layer (i-HAL) with a hardware-based trust fabric
Accelerator-aware xApps and dApps for offloading RAN workloads
AI-driven, multi-layer orchestration of the RAN across the compute continuum
A hardware-aware MLOps pipeline for 6G RAN AI models
Energy-efficient deployment blueprints, targeting for example a 30% energy reduction for optimised workload execution compared to a CPU-only baseline

Project’s Progress so far

5%
PREVIOUS PROJECT
NEXT PROJECT
Cart (0 items)

Reach out, your message is the first dot.

Head Offices
208, Kifisias Avenue, Chalandri, 15231, Athens, Greece
Technical Branch
32, Vosporou, Elliniko, 16777, Athens, Greece
Contact with us
Call Consulting: +30 210 723 6195 Call Cooperate: +30 213 088 9302
Working Hours
Mon - Fri: 9.00 - 17.00 EET
Facebook
Twitter/X
LinkedIn