🔍 What's in today's flow
💧 UC Riverside–Caltech study reveals AI data centers could require $10B–$58B in new water infrastructure
🤖 California water agencies deploy AI for leak detection, wastewater optimization, and smart metering
📊 Anthropic launches Claude Code Channels - AI coding via chat platforms with potential for water utility operations
🔧 Itron's AI-powered leak sensor platform transforms non-revenue water management
⚡ New research maps AI's role in stormwater management: opportunities, barriers, and equity
────────────────────────────────────────────────────────
AI research spotlight: data center water demands could cost billions in new infrastructure
A landmark study published on March 9 by researchers at UC Riverside and Caltech has quantified the hidden water cost of the AI boom and the figures are staggering.
The paper, titled "Small bottle, big pipe: Quantifying and addressing the impact of data centers on public water systems," finds that without new efficiencies, data center cooling systems could require an additional 697 million to 1.45 billion gallons of peak water capacity per day by 2030 , roughly equivalent to New York City's entire daily water supply. The estimated infrastructure cost: $10 billion to $58 billion.
The study highlights a critical but overlooked challenge: peak demand. While annual data center water usage may appear modest, daily demand from evaporative cooling systems can spike 6–10 times the average — and in some planned facilities, over 30 times. A single large data center can withdraw over a million gallons per day on hot summer days.
Key recommendations:
Data centers should report peak water use, not just annual averages
Developers should partner with communities to fund infrastructure upgrades
Companies should offset water capacity to preserve community supplies
Adaptive cooling strategies should toggle between water-based and dry cooling based on local grid and water system stress
Why it matters for the water sector: This research directly challenges the narrative that AI's water footprint is negligible. For water utilities, the study underscores the urgency of planning for peak demand scenarios driven by data center growth, particularly in water-stressed regions.
👉 Full study
────────────────────────────────────────────────────────
Case study: California's water agencies begin looking to AI
California's water agencies are emerging as early adopters of AI, not for hype, but for measurable operational gains. A report from the Public Policy Institute of California (PPIC) published this week profiles several agencies deploying machine learning across their operations.
Technologies in action:
1. Wastewater treatment optimization — The Eastern Municipal Water District (EMWD) in Riverside County uses machine learning models to optimize treatment processes, saving an estimated $100,000 per year in energy costs alone.
2. AI-powered pipeline maintenance — EMWD and Tucson Water have partnered with VODA.ai to deploy AI for predicting pipeline failures and prioritizing repairs. The system analyzes pipe age, material, soil conditions, and break history to forecast where failures are most likely.
3. Smart metering and water use modeling -The Moulton Niguel Water District uses machine learning combined with advanced metering infrastructure (AMI) to model water consumption patterns and detect faulty meters, improving demand forecasting accuracy.
4. Real-time operational dashboards - The Los Angeles Department of Water and Power (LADWP) has implemented an AI-powered platform integrating diverse operational data streams into centralized dashboards for informed management decisions.
Key takeaways:
AI is not replacing water sector workers - it's augmenting their decision-making
High-quality, "AI-ready" data is the foundation for effective deployment
Organizational policies for privacy, cybersecurity, and workforce training are essential prerequisites
Impact: These deployments demonstrate that AI adoption in the water sector is no longer theoretical. From energy savings to predictive maintenance, California's agencies are proving the business case for AI-driven water management.
────────────────────────────────────────────────────────
Latest in AI
What's new
Anthropic launched Claude Code Channels (March 23), enabling users to control its AI coding agent via Discord and Telegram. The feature directly counters the viral open-source tool OpenClaw, which lets users create personal AI agents through chat apps. Claude Code Channels adds enterprise security with admin-managed allow-lists and Model Context Protocol integration.
Why it matters for the water sector
Water engineers and operators can now interact with AI coding assistants through familiar chat platforms like Discord, simplifying SCADA script development, sensor data queries, and operational troubleshooting — without needing terminal access. The built-in security controls make it viable for utility IT environments.
────────────────────────────────────────────────────────
AI tool of the week: Itron leak sensor + AI pipe asset management
Itron's NRW platform combines acoustic leak sensing with AI/ML-powered pipe asset management to help water utilities detect losses and protect revenue in real time. The Temetra Analysis software prioritises both real and apparent water losses, while open APIs ensure smooth integration with existing billing, GIS, and asset management systems. Rogers Water Utilities cut NRW from 18% to 5%, saving ~$200,000 annually.
Ease of Use - 4/5 The platform is well-structured and designed for utility operators, not data scientists. The learning curve exists but is manageable for teams with basic technical familiarity.
Cost - 3/5 Enterprise pricing makes this a significant investment. Strong ROI is achievable, as Rogers Water proved, but upfront costs place it out of reach for smaller utilities.
Security & Privacy - 4/5 Built for critical infrastructure, with robust data handling and network-level controls. Falls just short of a perfect score as third-party integration points introduce standard dependency risks.
Integration -5/5 A standout. Open APIs and native compatibility with AMI, GIS, billing, and CMMS systems mean it slots into existing utility infrastructure with minimal disruption.
Overall -16/20 A mature, proven platform that delivers measurable results. The integration strength and real-world performance make it a compelling choice for utilities serious about tackling water loss if the budget allows.
🔌 Try it
────────────────────────────────────────────────────────
The shadow of AI: stormwater, equity, and the barriers to adoption
A comprehensive review published in SN Applied Sciences (Springer) this month offers a sobering assessment of where AI stands in stormwater management - and why the technology's promise remains unevenly distributed.
The paper maps AI applications across three domains observation, analysis, and governance and finds that while machine learning and hybrid models show strong potential for real-time monitoring, predictive analysis, and improved system efficiency, significant barriers persist.
Key findings:
AI enables real-time monitoring, automated inspections, and continuous water quality tracking , generating high-resolution datasets that traditional methods cannot
Predictive modeling for flooding, infrastructure optimization, and scenario simulation are AI's strongest analytical contributions
Deep reinforcement learning is being explored for real-time control of stormwater systems
Critical barriers identified:
Data quality gaps and inconsistent standards across jurisdictions
Model accuracy and scalability challenges when moving from pilot to deployment
Institutional and regulatory frameworks not designed for adaptive AI systems
Ethical concerns about fairness and transparency in AI-driven decision-making
Water sector implications:
The review calls for governance frameworks that explicitly prioritize equity ensuring AI-driven stormwater solutions don't disproportionately benefit wealthier communities
Proposed solutions include hybrid physical-AI models, standardized data protocols, and equity-centered pilot projects
Workforce capacity building remains essential: AI augments but cannot replace the expertise of water professionals
The paper serves as both a roadmap and a reality check — AI can transform stormwater management, but only if deployment addresses data, governance, and equity gaps simultaneously.
Thanks for reading! I hope you’ve enjoyed this week’s edition and look forward to seeing you next week!
Dr. Andrea G.T