Sustainable AI Insights
Deep dives, case studies, and practical guides on reducing the carbon footprint of Artificial Intelligence computing.
How Much Water Does ChatGPT Drink? LLM Water Footprints Under the UN’s Warning
We analyze the hydrological footprint of major LLMs—including reasoning models like o1, o3-mini, and DeepSeek-R1—in light of the UN’s 2026 environmental cost report.
Building a Green Software Business Line with Embeddable AI Carbon Badges
A developer’s guide to embedding real-time AI sustainability badges on websites, and how to monetize carbon-aware software as a service (SaaS) or FinOps business.
Claude Code & Agentic Dev AI: Tracking Carbon Costs Under the UN’s 2026 Warning
As developers adopt Claude Code and agentic CLI tools, we analyze the environmental footprint of autonomous software development loops in light of the UN's recent sustainability warnings.
The UN’s June 2026 Warning: AI’s Environmental Cost Goes Beyond Carbon
The United Nations University has released a landmark report on the environmental footprints of AI compute. It warns that by 2030, AI energy consumption could reach 945 TWh, with water footprints equivalent to 1.3 billion people, and a massive physical land demand.
AI Weather Forecasting: Can Machine Learning Save the Climate?
AI weather forecasting models like Google’s GraphCast and NVIDIA’s Earth-2 are revolutionizing weather prediction and climate modeling. But can these fast, efficient neural networks help us manage climate risk, and at what environmental cost?
How to Report AI Compute in Scope 3 Under SEC Climate Rules
The SEC climate disclosure rules require large filers to report Scope 3 emissions. Here is how to account for AI compute, data center energy, and water usage in your filings using research-backed data.
CSRD and AI: How European Companies Should Report AI Emissions Under ESRS E1
The EU Corporate Sustainability Reporting Directive requires Scope 3 emissions disclosure starting 2026. For companies using AI heavily, here is how to account for compute emissions under ESRS E1.
El Nino 2026: What a Potential Super Event Means for the Planet
The Pacific is shifting from neutral to El Nino, and forecasters warn we could be heading for a super event by winter. Here is what that means for global temperatures, food security, and the climate targets we set on Earth Day.
The Gulf Stream is Shifting North: An Early Warning for AMOC Collapse
New high-resolution modeling reveals that the Gulf Stream is shifting northward. This observable change provides a critical early warning signal for the collapse of the Atlantic Meridional Overturning Circulation (AMOC).
Run AI Models Locally with Ollama: Greener, Faster, Private
Every prompt you send to a cloud API spins up a GPU farm you cannot see. Ollama lets you run powerful open-source models on your own machine, cutting cloud energy use and keeping your data off someone else's server.
Understanding the AMOC Tipping Point
A breakdown of how the Atlantic Meridional Overturning Circulation works, its relationship with climate change, and what recent science tells us about its future.
The AMOC Collapse Debate: What CMIP6 Models *Actually* Tell Us About Resilience
A data-driven analysis of CMIP6 models revealing why some predict a wind-driven AMOC remnant, while others show a catastrophic failure.
The Carbon Footprint of AI Image Generation
Why generating one image costs as much energy as charging your phone, and how designers can work sustainably.
Claude 3.5 Sonnet vs GPT-4o: Which AI is More Energy Efficient?
A technical deep dive comparing the energy usage, water consumption, and carbon footprint of OpenAI's GPT-4o versus Anthropic's Claude 3.5 Sonnet.
5 Ways to Reduce Your AI Carbon Footprint
You don't have to give up AI to be sustainable. Here are 5 actionable tips for developers and users to use AI responsibly.
How Much Energy Does ChatGPT Really Use?
Understanding the hidden carbon footprint of standard LLM queries compared to traditional search and why it matters for the future of the planet.