AI/ML Driven Software Optimisation to Reduce Cost and Climate Impact. A first-of-a-kind software decarbonisation pipeline that reduces the climate impact of computing at scale.
Video created by explainapaper.com.
For the latest on GreenCode, visit greencode.ai and follow the project on YouTube.
Addressing the root cause of IT emissions: software efficiency.
In 2022 the IT sector's footprint was about 860 million tonnes of CO2e, and on published estimates it now exceeds civil aviation's. Legacy systems make up 60 to 80% of all software, carrying technical debt and energy waste. GreenCode tackles this at the source.
The Problem
The world's IT infrastructure consumes a rapidly growing share of global energy, predicted to reach around a fifth of global electricity by 2030. The root cause is not hardware -- it is how software is written, maintained, and deployed. Legacy codebases, inefficient algorithms, and unoptimised infrastructure create compounding energy waste at scale. Current approaches focus on hardware efficiency and renewable energy supply, but leave the software layer largely untouched.
The Solution
GreenCode builds a software decarbonisation pipeline that uses generative AI, sustainability best practice, stress testing, and software quality assurance to optimise software for energy efficiency. The pipeline maps codebases for energy hotspots, applies AI-driven refactoring, validates improvements through automated performance testing, and certifies the results. The outcome: reduced carbon footprint, developer productivity gains, increased reliability, and reduced operating costs.
"GreenCode addresses the elephant in the room: software itself. You can power your data centre with 100% renewables, but if your code is inefficient, you are still wasting energy at scale. We need to optimise what runs on the hardware, not just the hardware it runs on."
Nine pillars of software decarbonisation.
GreenCode delivers a comprehensive pipeline covering codebase analysis, AI-driven optimisation, performance validation, developer training, and green certification.
Comprehensive Codebase Mapping
Deep analysis of software architectures to identify energy hotspots, inefficient patterns, and optimisation opportunities across entire codebases.
Green Optimised Generative AI
AI-driven code refactoring that targets energy efficiency -- transforming wasteful patterns into optimised implementations while preserving functionality.
Automated Software Maintenance
Continuous automated maintenance that keeps codebases lean, removes technical debt, and prevents energy waste from accumulating over time.
Energy and Cost Reduction
Measurable reductions in energy consumption and cloud compute costs through targeted software optimisation -- targeting a measured energy reduction of more than 15% per system.
Automated Performance Testing
Stress testing and benchmarking that validates energy improvements and ensures optimised code meets reliability and performance standards.
Developer Sustainability Training
Upskilling programmes that embed energy-aware coding practices into development teams, creating lasting cultural change in how software is built.
Green Software Certification
A certification framework that provides verifiable evidence of software energy efficiency -- giving organisations a credible sustainability credential.
Sustainability Led Code Quality Assessment
Code quality assessment that prioritises energy efficiency alongside traditional metrics like maintainability, readability, and test coverage.
Infrastructure Carbon Footprint Assessment
End-to-end carbon footprint analysis of IT infrastructure -- from cloud configurations and container orchestration to CI/CD pipelines and deployment patterns.
The scale of the problem -- and the opportunity.
Tonnes CO2/Year
Tonnes of CO2e from the IT sector in 2022 (GISWatch). On published estimates the sector now emits more than civil aviation.
Legacy Software
Of all software is legacy code, carrying technical debt and compounding energy waste.
Of Global Electricity by 2030
Wider estimates put the IT sector at around a fifth of global electricity by the end of the decade.
Target Energy Saving
Measured per system. Published studies of refactoring report savings of 10 to 30%.
21 partners across 8 countries.
A consortium of leading universities, research institutes, and technology companies supported by six national funders: Innovation Fund Denmark, BMFTR (Germany, administered by DLR), ANI, CDTI, TÜBİTAK and Innovate UK. The Austrian and Romanian partners take part at their own cost.
Published research.
The GreenCode State of the Art Review (Deliverable 2.1, December 2025) surveys the software energy-measurement and optimisation landscape across 182 pages, laying the scientific foundation for the project's decarbonisation pipeline. It was edited by Chris Dean of Digital Tactics (Work Package 2 lead), drawing on inputs from more than 70 contributors across the partner organisations. It is published through ITEA at itea4.org and on greencode.ai.
Two further deliverables are also public: the GreenCode Project Vision Statement (Deliverable 2.3, 2025), authored by Chris Dean, and the Work Package 3 State of the Art Review (Deliverable 3.1, 2025) on static code analysis and quality tools, to which Digital Tactics contributed.
Related R&D Projects
View all R&D projects across Eureka, Horizon Europe, and internal research.
GIGI
Generative AI Inference Governance and Impact Measurement. Provider-agnostic AI usage control plane for enterprises.
GLEAM
Green Learning for Eco-Conscious Medical AI. Energy-efficient AI for medical imaging with federated learning.
Horizon CSA
Using AI to optimise scientific computing for performance and energy efficiency. In development.
Interested in GreenCode?
Visit greencode.ai for the latest project updates, or get in touch to discuss collaboration opportunities.
AI software engineering and innovation
for ambitious organisations