As the saying goes, all good things must come to an end. This June 2026 marks the final month of the EXIGENCE project, wrapping up three years of intense, groundbreaking research dedicated to driving energy efficiency and carbon reductions across information and communication technology (ICT) ecosystems. But before we officially cross the finish line, our team gathered for one last major event together: the EuCNC & 6G Summit 2026 in beautiful Malaga, Spain, held from June 2–5.
As a proud member of the Smart Networks and Services Joint Undertaking (SNS JU), EXIGENCE made an impactful final appearance, demonstrating how concrete innovations can advance true operational sustainability in upcoming 6G systems. From highly interactive live booth demonstrations to detailed technical posters and high-profile workshop sessions, here is a look at how we spent our grand finale!
At the Booth: Bridging the Gap Between 6G Infrastructure and Everyday Consumers
At Exhibition Booths 1 & 2, the EXIGENCE project hosted a bustling, multi-screen live demonstration. Traditional green network paradigms often focus on isolated, per-domain energy optimisations. The core mission of EXIGENCE is to shatter those silos by creating an architecture capable of measuring and reducing end-to-end energy usage across heterogeneous domains while closing the loop through user awareness.
To show the public how transparent ecological data can shift user behaviour and optimise network resources, we brought interactive screens using highly relatable digital services:
Screen 1: Interactive Video Streaming
Users experienced hands-on interaction with a live video stream running over a real 5G network. They could track real-time, end-to-end energy consumption and dynamically adjust the bitrate, instantly seeing how their choices impacted both video quality and carbon emissions.
Screen 2: AI/LLM Inference Tracking
Users accessed a Large Language Model (LLM) powered by our GPU infrastructure, where they could input custom prompts. The interface provided more than just answers; it calculated the exact carbon footprint for every interaction. Furthermore, users could activate the EXIGENCE “low-carbon” mode, which offloads energy-intensive prompts to GPU servers powered by green energy, effectively reducing the carbon impact of their AI usage.
Screen 3: Gamified Leaderboard
Integrated with the video stream, this dashboard translated eco-friendly user selections into “environmental points,” using peer competition to steer users toward green habits, and integrating users’ individual characteristics, e.g., selfishness and altruism.
Screen 4: Federated Learning Simulation
This network-level visualisation tracked distributed user equipment battery levels alongside grid versus renewable energy consumption, proving how our system lowers grid dependency without sacrificing performance.
The Innovations: Breaking Down Our Result Posters
Backing up these live experiences were our newly published technical results, proving that application-level sustainability is achievable without ballooning provider costs.
Incentives for Green Video Streaming
To address the digital streaming carbon footprint, this research done by our partner the STEcon group AUEB, maps the behaviour of 1,000 heterogeneous users using a Stackelberg game model. Operating as a continuous loop, the framework nudges consumers to opt for energy-efficient bitrates based on their personal “greenness factor”, a private incentive threshold, and peer recognition. By introducing serious-game mechanics – ranking users into top and bottom tiers where top-performers get social rewards (H) and bottom-performers face penalties – the framework achieves remarkable results.
When asking users to switch from Ultra-HD (4K at 20 Mbps) to Full-HD (1080p at 5 Mbps), the mechanism secured a massive traffic reduction of up to 67.2%, capturing nearly 90% of the theoretical maximum for emission savings. Crucially, the data proved that monetary budgets suffer from diminishing returns; instead, under tight budgets, social motivation and community recognition completely dominate financial perks, boosting user adoption from 9.4% to 20% and multiplying network energy savings more than fourfold.
End to end power consumption measurements for user traffic streams in a 5G network
This research done by our partners from TNO, builds a comprehensive telemetry and attribution framework to map power consumption directly to specific network functions or individual users. Tested across a physical Radio Access Network (RAN) and an OpenStack virtualised Core Network, the study isolated how energy scales with operational loads. In the physical RAN, baseline testing clocked an idle power consumption of 79 W at zero load, transitioning into a strictly linear relationship as downlink traffic rose to 30 Mbps.
Remarkably, both environments revealed that serving multiple simultaneous user streams incurs almost zero additional power overhead. By feeding these precise cross-domain power attributions into a Double Deep Q-Network (DQN) reinforcement learning algorithm, the team successfully automated user network flow distributions across a 50-node edge compute infrastructure, improving operational power consumption by 2% to 5% over standard baselines.
GreenFlag: A Green Agentic Approach for Energy-Efficient Federated Learning
As mobile networks transition toward 6G, integrating distributed intelligence is crucial for driving system autonomy and real-time adaptability. Federated Learning (FL) has emerged as a frontline technique in this space, enabling on-device AI model training while preserving strict user data privacy. However, training distributed models across thousands of devices introduces massive energy demands that traditionally rely on carbon-heavy grid power. To solve this, a collaborative effort between partners NKUA and HWDU resulted in GreenFLag – a novel, agentic resource orchestration framework designed to minimise grid power reliance and lower the overall carbon footprint of FL workflows without sacrificing model performance.
The framework operates by incorporating renewable energy sources directly into the system’s operational logic. Using a sophisticated Soft Actor-Critic (SAC) reinforcement learning approach, GreenFLag intelligently and jointly optimises both computational and communication resources. The AI agent dynamically accounts for real-time network communication contention among devices while tracking the highly volatile availability of local green energy.
When evaluated using real-world open meteorological data from Copernicus, the results were nothing short of spectacular. GreenFLag successfully reduced grid energy consumption by an average of 94.8% compared to three state-of-the-art baselines. By shifting the heavy lifting of localised AI training to periods and areas with peak renewable availability, the framework proves that future 6G distributed intelligence can run almost entirely on green power-effectively neutralising the environmental impact of next-generation edge AI.
In the lecture Halls: Championing “Operational Sustainability”
Beyond the exhibition floor, EXIGENCE co-anchored Workshop 11: Sustainable by Design, Sustainable in Operation: The 6G Perspective on June 2. Co-organised by Mir Ghoraishi (Gigsys Solutions) and Christoph Schmelz (Nokia), this session united expertise across 12 distinct SNS projects to explore how sustainability can be engineered directly into runtime decision processes, control loops, and live 6G architectures.
Green Orchestration & Long-Term Ecosystem Evolution
EXIGENCE partner Artur Hecker (Huawei Munich Research Centre) delivered a compelling joint presentation titled “Green Orchestration and Energy-Aware AI”. He tackled the massive sustainability challenges posed by rapid transitions into large-scale AI workloads, distributed training, and inference services. Hecker illustrated how green orchestration can dynamically align AI workload placement with real-time energy conditions, using cross-domain data exchange and federated coordination to unlock carbon-aware operation across organisational boundaries.
Following the theme of long-term survival, EXIGENCE partner Prof. Rui Aguiar (Institute of Telecommunications Aveiro) gave an insightful keynote address, “For the Long Run: Telecom in a Changing Ecosystem”. Prof. Aguiar detailed the operational pressures telecom operators face during the 6G transition, emphasising how software-driven complexity, cloud-native architectures, and intelligent automation must be leveraged to control operational costs and maintain resilience over multi-stakeholder lifecycles.
The Ultimate Highlight: A Standout Commendation from the EU
The absolute standout of our week occurred during the concluding panel discussion, “Beyond Greenwashing: What Does Sustainable 6G Really Mean?” Moderated by Dr. Mir Ghoraishi, the panel featured elite industry voices alongside Chiara Mazzone from the SNS JU Office.
While discussing the rigorous EU Code of Conduct for sustainability measurements, Chiara Mazzone specifically singled out the EXIGENCE project among all other active SNS initiatives!
She lauded the project for introducing the vital concept of overavailability at the service layer. As Mazzone astutely pointed out to the audience, measuring energy consumption strictly at the isolated network or hardware domain level does not truly empower or benefit the end-user. By elevating the sustainability conversation directly to the application and service layer, EXIGENCE has introduced a total game-changer for how the telecommunications industry will validate green metrics moving forward.
Leaving a Lasting Legacy for 6G
As the sun sets on the EXIGENCE project at the end of June 2026, we leave the European telecom ecosystem with our heads held high. Our final showcase at EuCNC & 6G Summit proved that achieving real-world sustainability in 6G requires moving past hollow vision statements and static efficiency metrics. By combining multi-domain power telemetry, intelligent green orchestration, and gamified user incentives, we have successfully demonstrated that networks can drastically mitigate their environmental impact while staying commercially viable and operationally resilient.
Though the project is concluding, the frameworks, relationships, and innovations built by the EXIGENCE partners over the last three years will continue to serve as a foundational blueprint for the cleaner, smarter, and truly sustainable 6G networks of tomorrow.
A massive thank you to the entire EXIGENCE consortium for their dedication, brilliant research, and hard work over the last three years. What a way to cross the finish line!
Author
F6S Innovation
Melissa Tang, Communication Manager at F6S Innovation, plays a pivotal role in various EU-funded initiative. Her expertise, built on a foundation of a BA in Design and Advertising and a Master’s in Communication and Media Studies, is further strengthened by hands-on experience leading communication, dissemination and exploitation efforts in over 15 research and business projects.