# Couldn’t attend Green IO New York?… Nolwenn got it covered for you

Green IO New York was nothing short of energizing — a gathering filled with insight, candor, and real-world impact. Our speakers (from veterans to trailblazers) didn’t just share best practices, strategies, and tools — they brought vulnerability, hard-won lessons, and truth about the real challenges our community is facing. Sometimes with humor 😀. Below are my main takeaways. For additional details, check the presentations on the Green IO New York website and check the conference booklet for an overview.

🌱 From Awareness to Action: Operationalizing Sustainable Tech

Sustainability is not just an ethical concern — it's a practical, financial, and strategic imperative. Moving from intention to implementation requires measurable actions, tools, and organizational buy-in.

• GreenOps as a repeatable operational model (Thoughtworks, Mastercard)
• Learn–Measure–Reduce frameworks (Green Software Foundation)
• Tools like Greenpixie, Cloud Carbon Footprint, Electricity Maps, Boavizta, CodeCarbon, AI Energy Score, Ecograder and AIWattch are useful to raise awareness of compute intensity, track and reduce emissions
• "Sustainability isn’t a checklist — it’s a mindset embedded in decisions and systems."

🖥️ Web & Cloud Optimization as a Climate Lever

What is the carbon impact of deploying in this vs. that region or with this or that provider? What are the carbon trade-offs of your data storage policy? The digital footprint of websites and cloud infrastructure is both measurable and reducible. Optimization = emissions reduction + cost savings.

• Ecograder and WSG provide standards and metrics for web sustainability (Tim Frick, W3C).
• Cloud optimization strategies:
• Autoscaling & right-sizing, turning off unused instances
  • Eco-routing of traffic to cleaner data centers (more below)
  • Serverless + storage tiering
  • Off-peak scheduling
  • Software-level optimization
• Results: Heroku, Mastercard, and Leboncoin each demonstrated tangible savings in both cost and carbon.

“Move slow and plant things” replaces the Silicon Valley “move fast and break things.”

🛰️ Carbon-Aware Scheduling & Location-Based Optimization

Carbon emissions from digital systems vary dramatically by region and time. Latency vs. carbon is often a worthwhile trade-off.

• Dynamic scheduling using real-time emissions data (grid carbon intensity fluctuates)
• Route workloads to greener grids (i.e. with lower carbon intensity)—even if it increases latency slightly (e.g., Paris vs. New York example)
• Use tools like Electricity Maps, Greenpixie and Boavizta,
• Embodied emissions matter too: using older hardware efficiently can avoid emissions from manufacturing

“Where and when you compute is as important as how.”

🧠 Open-Source AI and Efficient AI Practices

Open-source AI brings innovation and access—but must be managed responsibly to avoid massive energy waste.

• The 3Rs of Sustainable AI: Reduce, Reuse, Recycle (Sasha Luccioni, Hugging Face)
  • SmolLMs (a family of smaller models), distillation (transfers knowledge from large models to smaller ones reducing deployment costs), quantization (lower precision reducing  compute and memory needs)
  • Energy-aware development (Hugging Face’s AI Energy Score)
  • Using pre-trained, efficient models
  • Fine-tuning with LoRA and tools like AutoTrain
  • Deploying in renewable-powered regions
• Efficient prompting and model design reduce inference-phase emissions (Hugging Face, AIWattch)

“Bigger isn't always better” — especially in energy-hungry AI systems.

🧩 Integrating FinOps and Sustainability for Scalable Impact

When cost optimization (FinOps) and carbon reduction align, organizations gain traction and support for sustainable practices.

• Mastercard: $28M saved and 410 MTCO₂e avoided through integrated action
• Aligning reporting, tooling (Greenpixie), and recommendations to both financial and sustainability metrics
• Sustainability becomes a core performance indicator, not just an add-on

“Every cost-saving action can also be a carbon-saving one—with the right visibility.”

⚖️ Responsible Tech as a Holistic Framework

Sustainability sits within a larger Responsible Tech ecosystem, encompassing ethics, equity, accessibility, privacy, security and governance (All Tech Is Human, NTEN, Green Software Foundation, Sustainable Architectures expertise).

• UN–Thoughtworks Playbook outlines principles across AI ethics, inclusion, security, and sustainability
• Design choices affect social and environmental outcomes
• Responsible AI must account for who benefits and who bears the cost, as well as unintended impacts (useful tools Tarot Cards of Tech)

“We already have the tech to solve climate change. What we lack is the will to challenge the status quo.”

🔥 AI's Role in Fossil Fuel Expansion Must Be Challenged

AI is not neutral. Without governance, it is actively fueling fossil fuel extraction and delaying climate action.

• Per Enabled Emissions (Will Alpine), AI is helping Big Oil scale drilling and fracking
• Corporate sustainability pledges are being undermined by high-carbon use cases
• Real change requires regulation of AI use, not just optimization for lower environmental impact

“We must govern how AI is used—not just how efficiently it runs.”

📣 The Path Forward: Cultural Shift + Regulatory Readiness

Sustainability must be embedded into culture, design practice, regulation, and education.

• Standards, frameworks and regulations are emerging (e.g., SCI, WSG, CSRD, REEN, RGESN)
• Wording matters: reframe sustainability in terms of efficiency, resilience, cost reduction, i.e. business value (to avoid potential resistance in politically sensitive environments).
• Equip teams with knowledge, tools, data, and autonomy to act in addition to education on best practices on sustainable content/ product/ design/ software development
• Don't wait for regulation — lead now

“Visionaries don’t just imagine a better future—they build it.”

🧭 Conclusion

To scale sustainable digital practices, we need a multi-dimensional strategy:

• Govern AI’s purpose, not just its power draw
• Connect financial incentives to environmental impact
• Equip teams with real-time data and tools for actionable insights
• Optimize both cloud infrastructure and web experiences
• Embed responsibility into culture, policy, and product design

The digital sustainability movement is here. It’s practical. It’s measurable. And it’s urgent.