A lighter web, and lighter AI behind it The first workshop at The Web Conference on the sustainability of the web itself
GreenWeb brings together researchers, practitioners, and standards contributors working on measuring and reducing the energy and carbon footprint of websites, web systems, and the AI services behind them.
The web as it is, and the growth it could have.
Why GreenWeb, and why now
Every page load, video stream, model query, and background sync draws electricity somewhere, in the device, in the network, and in the data centre. For a long time this cost was treated as an operations detail rather than a research question for the web community. That position no longer holds.
The International Energy Agency projects that electricity demand from data centres will roughly double by 2030, with AI the main driver. The web itself keeps getting heavier too. Pages have grown for two decades, generated interfaces add weight that nobody reviews for efficiency, and LLM-backed features put an inference cost on actions that used to cost almost nothing.
The research is out there, but it is scattered. Green software engineering, systems, HCI, and the W3C sustainable web community each hold a piece of the problem. GreenWeb gives that work a home inside the web research community, where it meets the measurement, mining, and modelling expertise The Web Conference is known for.
Why Dublin is the right place to start
Ireland hosts one of the largest data centre clusters in Europe, and the sustainability of digital infrastructure is a live national debate. Figures from the Central Statistics Office put it plainly.
Data centres now use more electricity than every urban household in the country combined.
Topics of interest
We read the scope broadly. If your work speaks to the energy or carbon cost of the web or of web-facing AI, it probably fits. Open a topic to see what it covers.
Measuring web energy
Energy consumption of web pages, web applications, and browsers, on desktop, mobile, and constrained devices.
Energy cost of AI on the web
LLM inference and other AI services in web products, including agentic web systems, and methods to estimate, attribute, and reduce their cost.
Carbon-aware web systems
Scheduling, serving, caching, content delivery, and edge computing that respond to grid carbon intensity.
Sustainable web engineering
Design and engineering practice, page weight reduction, framework and protocol efficiency, and the effect of ads, trackers, and third-party content.
Datasets and benchmarks
Reproducible methodology for web and web AI energy studies, including mining large corpora of pages for sustainability signals.
Auditing tools
Automated auditing for web sustainability, and how it relates to established performance and accessibility auditing.
Trade-offs and co-benefits
How energy interacts with performance, accessibility, privacy, and cost, and where improving one helps the others.
Standards and policy
The W3C Web Sustainability Guidelines, European sustainability reporting rules, and what they mean in practice.
Experience reports
Industrial case studies of reducing the footprint of production web systems, including what did not work.
An interactive half day
GreenWeb is built for discussion, not just presentations. Roughly a third of the time goes to breakout groups and a closing panel.
- KeynoteAn invited talk from the web sustainability standards community or from industry.
- Paper sessionsSeven talks, twelve minutes each plus questions.
- Posters over coffeeFurther accepted papers on display through an extended break, so authors get real discussion time.
- Breakouts and panelThree facilitated groups on measurement, AI energy, and practice and standards, then a closing panel.
A community research roadmap
Notes from the breakout groups become a short roadmap of open problems, published after the workshop and sent to everyone who took part.
An open tools and data catalogue
A curated, public list of measurement tools and datasets for web energy research, kept in an open repository that anyone can add to.
Call for papers
Three kinds of submission in the ACM format, with single blind review. Every submission receives at least three reviews from the programme committee. Accepted regular and short papers appear in the Companion Proceedings of The Web Conference 2027, published by ACM.
Regular papers
Original research, empirical and measurement studies, datasets, and benchmarks. Up to two extra pages for references and an optional appendix.
Short papers
Work in progress, position and vision papers, tool and demo papers, and industrial experience reports.
Extended abstracts
Early ideas, practitioner case studies, and tools you want feedback on. Presented as posters or lightning talks.
Non-archival\documentclass[sigconf]{acmart}, also available on Overleaf.Extended abstracts are not published in the proceedings, so the work can still go to a full venue later. Papers that have been published before, or are under review elsewhere, cannot appear in the proceedings either, but they can be presented as non-archival contributions. Please say so when you submit.
Important dates
Deadlines follow the workshop timeline of The Web Conference 2027. All deadlines are anywhere on Earth.
- Workshop accepted
- Paper submission deadline
- Author notification
- Camera ready deadline
- The Web Conference 2027, Dublin. Workshop day to be confirmed
Anywhere on Earth means the deadline passes when it is midnight in the last time zone on the planet.
Organisers
Eleven organisers from four countries on three continents, with backgrounds in web energy, software engineering, data science, machine learning, natural language processing, and computer vision.
Programme committee
The programme committee is being formed now, drawing on green software engineering, web systems, AI efficiency, standards, and industry practice. If you would like to review for GreenWeb, please get in touch.
Volunteer to reviewGet in touch
Questions about the workshop or the call are very welcome.
muhammadsohaib.ayub@universityofgalway.ieIf you have access needs for the day, or ideas for the tools and data catalogue, write to us at any point before or during the workshop.
Email the organisers