Hm, I was sorta joking around with this post that I made earlier this week but this might actually become a thing at this rate lol. Earlier this week, the President made a post on Truth Social about a Super Intelligence (SI) Force. I wonder if we will see the equivalent of a Data Center Intelligence Force with some of the big gals and guys?

Other news that I spotted:

Cya alligator 🐊 ,

Nate

How much are AI data centers adding to your electricity bill? (Governance, justice, and political economy): NPR examines who pays when AI data centers require new power lines and grid upgrades. Scott Neuman and Elissa Nadworny discuss how those costs are divided between developers and other electricity customers, alongside questions about reliability and potential benefits from new infrastructure.

The Midwest’s Response to Data Centers (Governance, justice, and political economy): Sci on the Fly host Emilie Lozier speaks with Anna Haensch and Kavitha Chintam about data center expansion in Wisconsin and Illinois. The conversation explores how communities and policymakers are responding to the energy, water, and local governance questions that accompany these projects.

How Tech Billionaires Broke Democracy | The Spill (Governance, justice, and political economy): Amy Westervelt and Gil Durán discuss tech-billionaire influence, democracy, and the climate crisis in this episode of The Spill. The conversation offers a wider political lens on the people and ideas shaping the technology sector and its growing power.

Improving the Energy-Efficiency of the Code Generated by LLMs through Effective Prompting (Measurement and sustainable practice): This arXiv preprint tests eight prompting strategies across ten models generating Python and C++ code. The reported energy savings vary by model and language. Its findings concern the energy used to execute the generated code, which is one part of the broader environmental footprint of using AI.

Containing AI: From Responsible to Critical AI Governance (Governance, justice, and political economy): Paul Schütze and Benedetta Brevini argue for moving from responsible AI to critical AI governance. Their article in AI & SOCIETY brings environmental and social costs into the discussion and asks a question that can get lost in implementation debates: should a particular AI system be developed at all?

The Cost of Intelligence: Impacts of AI Data Centers on US Electricity Prices (Environmental footprint of AI): This working paper examines US data center additions and electricity prices from 2020 to 2025. The authors estimate price increases for industrial and residential customers, with results varying across developers and utilities. It adds to the debate over who bears the electricity costs of AI infrastructure, with findings that remain working-paper estimates.

Alternatives to Concentrated AI: A Layer-by-Layer Map of What Is Being Built (Governance, justice, and political economy): The Decentralization Research Center maps alternatives to concentrated AI across the technology stack, drawing on 36 interviews, three builder case studies, and nearly 30 systems, institutions, and standards. The report surveys what is being built today, with policy recommendations planned separately for 2027.

The Real Cost of Power (Environmental footprint of AI): This S&P Global Energy whitepaper, commissioned by Fluence, compares battery energy storage with diesel and gas generation for data centers. Its published overview introduces lifecycle costs and operational considerations, offering an industry perspective on the power choices facing developers as computing demand grows.

Khazna ESG Report 2025 (Environmental footprint of AI): Khazna Data Centers sets out its first standalone ESG reporting baseline, covering energy, cooling, water, emissions, and organizational oversight during 2025. The report provides a look at how a data center operator measures and discloses its impacts, including estimates where complete data were unavailable.

EASY PEASY: Data Center (Environmental footprint of AI): Legrand offers a roughly two-minute introduction to data centers and their role in everyday digital services. This company-produced explainer is a quick starting point for readers getting familiar with the physical infrastructure behind the cloud.

The AI Data Centres Situation is Crazy (Environmental footprint of AI): Internet Education examines the AI data center building boom and the infrastructure problems surrounding rapid expansion. The video follows the race for computing capacity and the enormous commitments behind it, asking what happens when those ambitions meet the practical challenges of building data centers.

Oracle Trucking Gas to Keep Data Centers on Track (Environmental footprint of AI): Bloomberg Tech reports on Oracle trucking natural gas directly to data centers to keep projects moving. Brody Ford joins Ed Ludlow to discuss the approach, including its potential use at a New Mexico development affected by a delayed pipeline. A short look at the power constraints shaping the AI buildout.

American Infrastructure Alliance (Governance, justice, and political economy): The American Infrastructure Alliance advocates for data center development with protections for ratepayers, water resources, and local communities. Its stated principles emphasize developer-funded infrastructure and local accountability, making it an organization to follow in debates over the terms of new projects.

Studio Decarbonade (Measurement and sustainable practice): This Montreal-based digital sustainability studio helps organizations understand, measure, and reduce digital and AI impacts. Studio Decarbonade contributed impact models and best-practice guidance to DGC Green’s GenAI toolkit, connecting digital sustainability work with practical decisions in creative production.

Centered Around You (Governance, justice, and political economy): Centered Around You shares community perspectives on data center jobs, taxes, electricity, water, noise, and land use. The campaign is paid for by Infrastructure for America and supported by the data center industry, context to keep in mind when reading its account of development benefits and tradeoffs.

legoESM (AI for environmental action): legoESM is a differentiable Earth system modeling tool built with JAX. It couples physical components and supports gradient-based analysis and hybrid AI-physics research. The software is available for noncommercial use under its current license; commercial use requires a separate paid license, and computing costs are additional.

Kolibri (Measurement and sustainable practice): Aleph Alpha’s German- and English-language model supports reasoning, retrieval, and agentic workflows, with an emphasis on inference efficiency. Its model weights are available under Apache 2.0. The developer’s lower-energy comparisons remain vendor-reported, while hardware, hosting, and enterprise services can add costs.

tret (Measurement and sustainable practice): tret is an Apache 2.0 AI workbench for auditable knowledge work, with a climate-risk assessment pack and human approval of draft outputs. It also reports estimated energy and carbon for model runs. Those figures are token-based estimates; model-provider and hosting costs are separate.

May I introduce the new water cycle? LOL.

Thanks for reading the 62nd issue of The Climate Code! It means a lot to me and we have more coming in the future, so definitely stick around! 💚

Nate

New here? I would recommend checking out the InfraPrism if you’re interested in tracking data center policy at the state or federal layers or the AI & Environment Resource Hub open-access database.