Inside AquAI - Bring AI to every drop

Welcome back. This week exposed AI's hidden costs. SpaceX acquired xAI and immediately announced plans for space-based data centers because Earth's infrastructure "cannot meet AI demand without imposing hardship on communities." That's a problem for water utilities competing for the same resources. Meanwhile, researchers cracked natural language control for water network simulations with 100% accuracy, and NSW's new cyber strategy put utilities on notice with mandatory incident reporting. AI is moving faster than most utilities realize.

💧 SpaceX acquires xAI in $1.25 trillion merger aimed at building orbital AI data centers powered by solar energy to meet computing demands.

🔬 Researchers developed EPANET-Agentic, an AI system that controls water network simulations using natural language with 100 percent accuracy.

⚙️ LeakZon's AI platform helps utilities reduce non-revenue water by up to 66 percent through automated leak detection and virtual metering zones.

🛡️ New South Wales launches 2026–2028 cyber security strategy with tighter protections for water utilities and critical infrastructure.

Researchers developed EPANET-Agentic, a multi-agent AI system combining large language models with EPANET simulation software. The system enables water network modelling through natural language commands while maintaining full computational accuracy.

This breakthrough removes technical barriers to hydraulic modelling, allowing operators to run complex simulations without specialised software training. It positions AI-driven systems as practical tools for digital twin applications and autonomous water network management.

👉 Full paper

SpaceX acquired Elon Musk's xAI in a $1.25 trillion deal announced February 2, creating the world's most valuable private company. The merger aims to build space-based AI data centers using solar power to address terrestrial energy constraints. SpaceX filed plans to launch up to one million satellites configured as orbital compute nodes operating between 500 and 2,000 kilometres altitude.

The acquisition reveals a harsh reality: AI infrastructure demands are outpacing Earth's capacity to support them sustainably. Water utilities deploying AI for optimization, leak detection, and forecasting must weigh these tools' resource costs against benefits. As data centers compete for water and energy in utility service areas, understanding AI's infrastructure footprint becomes essential for long-term planning.

👉Full story

LeakZon deployed its Water Efficiency Analytics & Detection (WEAD) platform across multiple water utilities and multi-family properties to combat non-revenue water. The AI system analyses data from smart meters, automated meter reading systems, GIS maps, and environmental sensors to create virtual district metering areas without physical infrastructure changes, automatically detecting leaks, theft, and meter failures.

Beta program results show up to 66 percent reduction in water loss rates, translating to hundreds of thousands in annual savings for utilities. The platform integrates with existing metering infrastructure, eliminating costly hardware upgrades while delivering real-time anomaly detection, automated resource allocation, and continuous network monitoring for improved operational efficiency.

👉Full story

ElevenLabs released v3, its most expressive AI voice model, featuring emotional depth, multi-speaker dialogue mode, and inline audio tags for controlling tone and delivery. The alpha model supports 70 languages and generates natural conversations between speakers. Water utilities can use it for automated customer notifications, training materials, and public communications.

Ease of use: 4/5
Intuitive text-to-speech interface with straightforward audio tag syntax, though advanced features require learning tag vocabulary.

Cost: 3/5
Subscription-based pricing with free tier available, but production-level usage requires paid plans for volume and commercial applications.

Security and privacy: 4/5
Cloud-based service with standard encryption, though utilities handling sensitive communications should review data handling policies.

Integration: 4/5
RESTful API supports most platforms, with SDKs for JavaScript and Python, though some legacy utility systems may need custom development.

Overall comment (15/20)

Strong option for utilities needing multilingual outreach, automated alerts, or training content. Best suited for external communications and public education rather than internal operational systems requiring strict data governance.

🔌Try it

The New South Wales Government released its 2026–2028 cyber security strategy in January, tightening expectations for critical infrastructure protection. Water utilities face stricter requirements for operational technology security, third-party vendor oversight, 24-hour incident reporting, and continuous risk management.

As water systems adopt AI-driven monitoring, SCADA automation, and digital networks, cyber vulnerabilities directly threaten service reliability and public health. The strategy signals regulatory pressure to embed security into procurement, asset management, and daily operations, not just IT departments. Utilities can no longer treat cyber protection as optional or separate from operational planning.

👉Full story

Thanks for reading! I hope you’ve enjoyed this week’s edition and look forward to seeing you next week!