Flowing Forward: AI and the Future of Water

EVOLVE 2026 · MIAMI · TRACK SESSION · SESSION 07

Water utility leaders discuss how AI is helping utilities shift from reactive maintenance to predictive decision-making in the face of aging infrastructure and growing demand as they tackle aging infrastructure and scarce resources.

EVOLVE is Sand’s annual summit bringing together technology leaders, policymakers and practitioners working at the intersection of AI and critical infrastructure.

For most people, water is invisible—until it isn’t.

Opening the discussion with Cape Town’s “Day Zero,” when residents counted down the days until taps were expected to run dry, the panel reminded the audience that water infrastructure is one of society’s most critical—and least visible—systems. As utilities face aging assets, workforce shortages and growing demand, the question is no longer whether more data is available. It’s how to use that data to make better decisions before failures occur.

Watch the Full Session

THE 60-SECOND READ

Water is the utility nobody thinks about until it fails, and much of the infrastructure carrying it in the US is 60 to 70 years old. The panel’s shift in one line: stop reacting to the loudest problem in the system and start prioritizing by consequence. Use the data to decide what to fix now, what to defer, and what failure would actually cost.

What This Conversation Revealed

1. AI helps utilities prioritize what matters most.

Water utilities rarely have the resources to fix everything at once. The panel argued that AI's greatest value lies in helping operators decide which assets require immediate attention, which can safely wait and where intervention will have the greatest impact.

2. Institutional knowledge is becoming critical infrastructure.

For every five workers leaving many utilities, only two are replacing them. As experienced operators retire, decades of practical knowledge risk disappearing with them. Capturing that expertise—and making it usable for future teams—emerged as one of AI's most valuable applications.

3. Better data only matters if people trust it.

Several speakers acknowledged that utilities already collect enormous amounts of operational data. The challenge isn't gathering more information—it's building systems that operators trust enough to use in everyday decision-making and that improve through continuous feedback.

Key Highlights

Cape Town showed what's at stake. The session opened with the story of "Day Zero," when residents faced the prospect of running out of water. It was a reminder that water scarcity is not a future scenario—it is already a lived reality for many communities.
Reactive maintenance has reached its limits. A technician's day often begins with a maintenance plan and ends responding to an unexpected pipe failure. The panel argued that AI can help utilities shift from reacting to emergencies toward anticipating them.
Workforce shortages are reshaping the industry. Utilities are losing experienced operators faster than they can replace them, creating an urgent need to preserve institutional knowledge before it leaves with retiring employees.
Growth is increasing the pressure. Bexar County's population of 2.3 million is expected to grow by another 1.5 million within the next decade, increasing demand on water infrastructure while making smarter planning and resource allocation even more important.
Feedback makes systems smarter. Speakers shared examples of smart meters that flagged leaks or generated inaccurate alerts. Without knowing whether those alerts were correct, AI systems lose the opportunity to improve over time.

By the Numbers

60-70 yrs 0

age of a majority of US water infrastructure

5 → 2 0

workers replaced for every five who leave

0 M

Current population of Bexar County

+ 0 M

Additional residents expected within a decade

In their words

“Utilities are built by families. There are folks with institutional knowledge that isn’t documented anywhere. And they retire.”

Calvin Farr, CEO, Prince William Water

Why it Matters

Water utilities face a challenge that extends well beyond infrastructure. Every day they make decisions about which assets to repair, which investments to prioritize and how to deliver reliable service with limited budgets and shrinking workforces. Those are fundamentally decisions about risk.

This discussion suggested that AI’s greatest contribution isn’t replacing operators or automating utilities. It’s helping experienced teams make better decisions with the information they already have—identifying where intervention will have the greatest impact before small failures become costly crises.

The conversation also reinforced an important lesson for Physical AI more broadly. Sensors, dashboards and predictive models only create value when they become part of operational decision-making. That requires more than technology. It requires trust, continuous feedback and close collaboration with the people who understand the systems best.

Perhaps the session’s biggest insight was this: resilience isn’t built by predicting every failure. It’s built by knowing which ones matter most—and acting before they happen.