Best of LinkedIn: Sustainability & Green ICT CW 34/ 35

The period centered on making AI measurable and manageable, with clearer efficiency metrics, stronger carbon observability, and practical engineering guidance. Water and cooling impacts remained in focus, while early investments in greener AI infrastructure and concrete enterprise modernizations signaled execution beyond strategy.

Date

August 20, 2025

Sustainability & Green ICT

Thomas Allgeyer

Methodology: Every two weeks we collect most relevant posts on LinkedIn for selected topics and create an overall summary only based on these posts. If you´re interested in the single posts behind, you can find them here: https://linktr.ee/thomasallgeyer. Have a great read!

‍

Listen to our podcast

AI Transparency and Efficiency Metrics

Google’s energy breakdown for Gemini apps highlighted rapid prompt-level efficiency gains and a sharper view on training versus inference impacts
Experts stressed standardized disclosure, pointing to comparable metrics and an emerging AI Energy Score to align reporting
Salesforce’s AI Sustainability Outlook emphasized trusted, reliable, sustainable AI practices as a prerequisite for scaled adoption
Guidance focused on right-sizing models, minimizing unnecessary compute, and aligning performance with footprint

‍

Data Centers, Water, and Thermal impacts

Water use surfaced as a key variable, with calls to clarify WUE and to distinguish training from inference consumption
Cooling choices and siting were underlined, including the need to manage local externalities such as noise near residential areas
Investment momentum appeared with Vesper Infrastructure Partners backing Terakraft to develop a green AI data center
Workload placement in facilities with strong renewable supply and advanced cooling was reinforced as a near-term lever

‍

Green Software and DevOps in Practice

DevOps was framed as a driver of sustainable IT through waste reduction, resource optimization, and automation discipline
Green software practices were tied to tangible business benefits via lower energy use and operational efficiency
Carbon-aware design patterns were promoted, including efficient code paths and demand shaping for lower-carbon execution windows
Frugal AI training initiatives supported a culture shift toward responsible compute consumption

‍

Measurement, Tooling, and Observability

The tool stack featured AWS Compute Optimizer, AWS Customer Carbon Footprint Tool, Google Carbon Sense, Microsoft Cloud for Sustainability, Scaphandre, and Kubernetes Vertical Pod Autoscaler
Carbon observability was positioned as first-class, tracking energy per request and enabling footprint-informed SRE decisions
Practical checklists emphasized monitoring, eliminating idle capacity, and continuously tuning workload-to-infrastructure fit

‍

Hardware Footprint and Lifecycle Insights

A comprehensive LCA of the NVIDIA A100 accelerated the conversation on hardware-level impacts and data collection rigor
Collaboration with lifecycle experts reinforced the need for component-level transparency to inform procurement and design choices

‍

Enterprise Moves and Reference Cases

Port de Barcelona modernized on IBM LinuxONE to consolidate workloads, cut energy use, and harden security while scaling open, containerized platforms
An AI plus Energy startup cohort was announced to drive grid modernization and clean energy innovation across emerging use cases

‍

Community, Governance, and Knowledge Sharing

The Jisc Digital Sustainability newsletter curated global signals, helping teams track policy, practice, and technology shifts
Accenture introduced the SAIQ metric to quantify AI’s return on sustainability and guide investment decisions
ASML hosted a meetup on embedding sustainability into software engineering, linking practitioners across disciplines
Green IO London announced a program on responsible technology with hands-on use cases and sector-specific GreenOps insights

‍

Subscribe to newsletter

Subscribe to receive the latest blog posts to your inbox every week.