Building Sustainable Software: Architectural Foundations & AI Frontiers

Green Software meets Artificial Intelligence: This meetup focuses on sustainability in software engineering. We’ll dive into two talks that offer comprehensive insights, starting with an introduction to the architecture of green software and concluding with a look at how the race toward ever-larger AI models is colliding with ecological limits. Can smarter designs, local models, or new energy strategies help keep AI’s carbon cost in check?

*** AGENDA ***

• 6.30 PM | Doors open. Grab some snacks & drinks.
• 7.00 PM | Welcome & Intro - Start of hybrid meetup part
• 7.10 PM | Architecting Green Software - An Introduction - Nils-Oliver Linden & Alexander zur Bonsen
• 8.00 PM | From Chat to CO₂ -The environmental cost of AI - Sebastian Macke

afterwards: More snacks & drinks, and a lot of time for networking with all attendees and speakers.

*** ABSTRACTS ***

Architecting Green Software - An Introduction

Nils-Oliver Linden (Principal Consultant) and Alexander zur Bonsen (Senior Consultant) at TNG

Software architecture and design greatly influence the ecological impact of IT systems. In our talk “Architecting Green Software - An Introduction” we discuss what it means to architect green software and how to start with it. We’ll explore the key decisions for green architects across all layers of the stack, from compute infrastructure, over hosting solutions, up to how one can architect for elasticity and handle data economically. If you are interested in developing sustainable software, this talk provides you with first steps for building a greener digital future.

***

From Chat to CO₂ -The environmental cost of AI

Sebastian Macke, Lead Software Engineer at QAware

Google searches once raised concerns about energy use. Today they seem trivial at 0.3 Wh each. AI, however, is a different story. One server rack can draw as much power as a village, and mega-projects already rival nuclear power plants. In just three years, generative AI has grown to claim an estimated 1–2% of global electricity. Beyond model training, even agent-style workloads like SWE-agent can add significant ongoing costs, as they keep large models active for extended reasoning and tool use. We explore how the race for ever-larger models is colliding with ecological limits and ask: can smarter designs, local models, or new energy strategies keep AI’s carbon cost in check?

*** REMOTE ATTENDANCE ***
Link will be published later.

*** CODE OF CONDUCT ***
Please comply with our Code of Conduct.

*** PHOTOS / VIDEO RECORDING ***
Please note that photos are taken during the meetup. To offer a hybrid meetup, we record the meeting on-site. If you do not want to appear in the photos / video recording, please contact the meetup team at the beginning of the event.

*** SPEAK on our Meetups ***
Do you have a topic you would like to present?
Please contact the Organizer Team!