Brief Author Note

You probably all saw it… Last week, Google released a paper on the energy and environmental costs of its Gemini service.

Industry leaders hailed it as revolutionary and a sign that AI is “green” after all.

Researchers and skeptics called it “AIwashing” and a green PR move illustrating the flaws in Google’s methodology.

I did a post on LinkedIn curating some of the criticisms from my favourite researchers in the space if you’re interested in learning more about this.

I also did a fun little exercise with Perplexity Labs to create a dashboard comparing what we have from OpenAI, Google, and Mistral with environmental metrics here.

BEFORE you send me an email blasting me about the flaws with my dashboard, trust me I know!

Methodologies vary, highly inaccurate numbers from OpenAI, etc. It’s just meant to be a fun exercise and be educational. 😁 

Chau,

Nate

Quick Links!

• Podcast Episodes

• Scientific Papers

• Policy Documents

• Multimedia

• Organizations

• Tools

• Fellowships Corner

• Submit Resources

Sustainability at Scale in the GenAI Era (AI & Environment): Dr. Katia Chaban shares her journey into sustainable IT, tackling the thorny issues of e-waste, embodied carbon, and the rapid scaling of AI. She offers a candid look at how GenAI is reshaping infrastructure decisions, where the biggest sustainability challenges lie, and how innovators are finding opportunities for more responsible growth.

Using AI for Sustainability and Building Profits for Businesses (AI & Policy/ESG): Sheryl Barbosa breaks down how AI is helping consumer-facing companies cut waste, streamline packaging and logistics, and enhance supply-chain resilience. Through case examples, she highlights how sustainability and profitability can move in lockstep when digital tools are leveraged to forecast inventory and drive smarter operations.

Rethinking AI’s Physical Footprint (AI & Data Centers/Hardware): This episode asks what happens when AI’s soaring appetite for energy outpaces grid capacity. Industry leaders weigh in on the physical realities of scaling AI, from construction bottlenecks to infrastructure strain, and explore how sustainability-focused design and innovation can prevent today’s digital boom from becoming tomorrow’s resource crisis.

Measuring the Environmental Impact of Delivering AI at Google Scale (AI & Data Centers/Hardware): A first-party Google study quantifying the energy, carbon, and water impacts of Gemini Apps inference at production scale, highlighting significant year-over-year efficiency gains and setting a benchmark for large-scale AI sustainability reporting.

Not All Water Consumption Is Equal: A Water Stress Weighted Metric for Sustainable Computing (AI & Water Resources): Wu et al. introduce SCARF, a framework that factors in spatial and temporal water stress to assess the true water footprint of AI workloads, datacenters, and chip fabs, providing a more context-sensitive metric for sustainable computing.

Permafrost Wildfire Interactions: Active Layer Thickness Estimates for Paired Burned and Unburned Sites in Northern High Latitudes (AI & Geospatial Analysis): Talucci et al. present the FireALT dataset, measuring permafrost thaw dynamics after wildfire events in Arctic and boreal regions, offering a critical tool for improving climate models and environmental monitoring.

Sustainability at a Crossroads (AI & Sustainability): GlobeScan’s 2025 research findings explore how businesses, policymakers, and civil society are navigating an inflection point in global sustainability. The report highlights shifting public expectations, the role of AI in both accelerating and complicating sustainable transitions, and the urgent need for cross-sector collaboration.

Chipping Point: Tracking Electricity Consumption and Emissions from AI Chip Manufacturing (AI & Infrastructure): Greenpeace investigates the hidden energy and carbon costs of AI hardware, tracing electricity demand and emissions across the chip manufacturing supply chain. The report underscores the environmental footprint of advanced semiconductor production and calls for greater transparency and regulatory accountability.

AI & Climate Change: Opportunities and Challenges for a Sustainable Future (AI & Climate): Mirova examines how AI can both drive and derail the transition to a low-carbon economy. The white paper outlines opportunities in renewable energy, adaptation, and risk management while cautioning against rebound effects, resource strain, and governance gaps that could undermine climate goals.

What Happens If There’s an AI Crash? (AI & Environmental Impact): Arvind Narayanan explores the possibility of an AI bubble bursting, drawing parallels to past tech crashes and asking what ripple effects it could have on industries, infrastructure, and environmental commitments built on AI’s rapid growth.

Earth Observed | Episode 1: Bill Greer – How Nonprofit Satellites Could Change Everything (AI & Multiple Applications): The debut of Earth Observed features Bill Greer, Co-Founder of Common Space, who discusses building an open-access satellite constellation for humanitarian use. The episode highlights how nonprofit-led geospatial innovation could democratize access to Earth data and transform climate and disaster response.

Environmental Group Launches Campaign to Raise Awareness About Data Center Impacts (AI & Environmental Impact): The Piedmont Environmental Group has launched “Virginians for a Smarter Digital Future,” spotlighting how unchecked data center growth could strain local energy grids and raise household bills. The campaign calls for smarter planning as digital infrastructure expands across Virginia.

Gaia3D Inc. (AI & Geospatial Analysis): Based in Asia, Gaia3D develops advanced weather data visualization tools that run seamlessly in the browser and beyond, making complex atmospheric information accessible for climate monitoring and decision-making.

Mapbox (AI & Geospatial Analysis): Headquartered in North America, Mapbox powers navigation for people, packages, and vehicles worldwide, using AI-driven geospatial technology to improve maps, logistics, and real-time mobility.

StackAI (AI & Infrastructure): Operating across industries from healthcare to defense, StackAI provides a horizontal platform for building and deploying AI agents and automations. Its tools help organizations accelerate workflows, scale efficiently, and integrate AI into core operations.

Surya Heliophysics Foundational Model (AI & Geospatial Analysis): A collaboration between NASA and IBM, Surya is an AI model trained on 14 years of solar observations from NASA’s Solar Dynamics Observatory. Built on open science principles, it enables researchers worldwide to accelerate heliophysics discoveries and advance climate-related space science.

Clara Research (AI & Workforce Development): Developed by Oxford’s Sustainable Law Programme, Clara is a litigation and accountability assistant that allows users to query climate litigation archives. It delivers citation-rich, AI-generated answers, helping lawyers, researchers, and advocates navigate complex climate cases with greater precision.

Eco Pal (AI & Sustainability): Eco Pal provides a collaborative platform for sharing knowledge, identifying systemic barriers, and designing solutions to sustainability challenges. By fostering collective problem-solving, it supports individuals and organizations seeking to accelerate environmental action.

AI x Animals Fellowship (Open to global applications): The AIxAnimals fellowship is an 8-week program preparing learners for impactful careers at the intersection of AI & animal welfare. As part of the fellowship, learners will join an industry-leading community working at the forefront of AI and sentient welfare for expert-facilitated discussions following our curated curriculum. 🗓️ Deadline: October 8, 2025.

Miller Center Global Social Benefit Accelerator (Open to Global South applications): An investment-readiness accelerator for mission-driven social enterprises. Open to CEOs, Executive Directors, or senior leaders of organizations operating in the Global South with ≥ $50K annual earned income, ≥ 3 full-time staff, and a proven impact model. 🗓️ Early Decision Deadline: September 10, 2025 | Final Deadline: November 1, 2025.

Climate Adaptation Creatives – British Council Program (Open to UK & Wider Europe applications): A peer-learning programme for experienced creative practitioners (≥ 5 years of practice) based in the UK and Wider Europe (including Albania, Armenia, Georgia, Turkey, Ukraine, and others). Open to artists and cultural practitioners developing creative responses to climate adaptation. 🗓️ Deadline: September 7, 2025 at 23:59 BST.

If you have resources that you wish to submit to the AI & Environment Resource Hub, fill out the Google Form! Thank you for suggesting resources to add to the Resource Hub. 🙃 

That’s it for this week.

Thanks for reading the 15th issue of The Climate Code! It means a lot to me and we have more coming in the future, so definitely stick around! 💚 

Nate

P.S.

New here? Check out the AI & Environment Resource Hub if you have no idea what it is. 😆