The tech industry now consumes an estimated 2–3% of global electricity, and AI adoption is rapidly accelerating that number. Yet many engineers are unsure where sustainability fits into their day-to-day decisions, or assume the solution lies in esoteric programming languages or full rewrites.

In reality, the biggest sustainability gains come from architecture and operations, not code micro-optimisations.

This full-day workshop provides a practical, operations-first framework for designing and running sustainable software systems. Participants will learn how reducing carbon emissions often results in a “triple win”: lower costs, better security, and simpler, faster systems—challenging the traditional “Good, Fast, Cheap: Pick Two” trade-off.

Through case studies, interactive exercises, and real-world examples, attendees will learn how to assess their organisation’s current maturity, identify high-impact improvements, and make a credible business case for change. Topics range from zombie servers and rightsizing to carbon-aware computing, cloud architecture choices, and the growing sustainability challenges of AI.

Participants will leave with concrete techniques they can apply immediately, plus a clear action plan for improving the sustainability of their systems, without waiting for perfect data or executive mandates.

What You’ll Learn

By the end of this workshop, participants will be able to:

• Understand where software systems really consume energy and carbon
• Apply an operations-first approach to sustainability
• Identify quick wins that reduce emissions and costs
• Measure impact using practical proxies and available tools
• Design cloud architectures with sustainability in mind
• Understand the unique sustainability challenges posed by AI workloads
• Build a compelling business case aligned with cost, security, and risk
• Create an actionable sustainability roadmap for their team or organisation

Workshop Format

Morning Session

• The carbon landscape: scopes, energy proportionality, and why utilisation matters
• Why operations dominate sustainability outcomes
• Measurement challenges and practical metrics
• Interactive exercise: assessing your organisation using the Green Software Maturity Matrix

Afternoon Session

• Sustainable cloud architecture patterns and trade-offs
• Demand shifting, carbon-aware computing, and graceful degradation
• AI sustainability: training vs. inference, and when not to use AI
• Making the business case: cost, security, compliance, and talent
• Interactive exercise: drafting a sustainability action plan

The workshop combines short talks, discussion, case studies, and hands-on exercises.

Prerequisites

No specific programming language knowledge required. Attendees should have:

• At least a basic understanding of cloud computing and software deployment
• Familiarity with operating production systems

Who Should Attend

• Developers responsible for system design or performance
• Platform, infrastructure, and SRE engineers
• Architects modernising legacy systems
• Tech leads and CTOs balancing cost, performance, and risk

Key Takeaways

Participants will leave with:

• A clear understanding of IT’s environmental impact and their influence on it
• Practical strategies that reduce carbon emissions without sacrificing performance
• Knowledge of available measurement tools and realistic metrics
• A draft sustainability action plan tailored to their organisation
• Resources for continued learning and implementation

Dr. Ewelina Kurtys, Strategy Advisor at FinalSpark, discusses the revolutionary potential of biological computing with host Charles Humble.

The conversation explores FinalSpark's ambitious goal to build computers using living neurons that are 1 million times more energy efficient than current digital systems. Ewelina explains the technical challenges of working with brain organoids, the ethical considerations of using human stem cells, and why biological neurons could dramatically reduce AI costs while offering unique advantages for complex tasks like generative AI. The interview covers everything from neural encoding and plasticity to consciousness and the deterministic nature of the brain.

In this State of the Art episode, Charles Humble speaks with Adrian Mouat, Developer Relations at Chainguard and author of "Using Docker", about the evolution of container security and the persistent challenge of outdated packages.

Adrian explains how traditional Linux distributions weren't designed for the immutable, frequently-replaced nature of containers, leading to security vulnerabilities that scanners detect but teams struggle to address. He discusses how Chainguard tackles this problem by building everything from source using Wolfi, creating minimal "distroless" images with near-zero CVEs, and how concepts like SBOMs, attestations, and defense in depth are reshaping security practices.

The conversation also covers major security incidents including the XZ Utils backdoor and Shai-hulud attacks, emphasizing the importance of building from source, using short-lived credentials, and replacing rather than updating containers – practices pioneered by companies like Google that are gradually spreading across the industry.