Water utilities are generally reactive to emergencies such as main failures or sewer overflows - a “break-and-fix” model unable to keep up with dilapidated infrastructure and more frequent severe weather, said Ryan Jeffrey, vice president of sustainability at the gener8tor venture capital firm.
A gener8tor-backed program focused on AI and machine learning is helping utilities and municipalities solve water challenges before they become costly emergencies. Launched in August, the gener8tor Great Lakes Innovation Accelerator welcomed five technology commercialization companies focused on data-driven infrastructure management.
Invitees’ AI-powered solutions span predictive analytics, smart monitoring and treatment optimization. Real-time sensor networks and analytics platforms offered by participants can flag pressure drops in a pipe, or harness algorithms to forecast storm surges before they occur.
Rather than replacing human engineers, this technology consolidates fragmented data into real-time insights for utilities. With the compounding pressures now facing utilities, water managers have little choice but to change tack, said Jeffrey.
“A few years ago, the water sector was about hardware, sensors and treatment methods,” Jeffrey said. “Data systems are more usable now, and AI can turn those sensors into predictive support and insight. There’s so much more momentum in this space.”
Supported by the Department of Commerce and the National Oceanic and Atmospheric Administration, the accelerator provides start-ups with funding, mentorship and access to corporate partners and investors.
Drawing from across the Great Lakes region, the 2026 gener8tor cohort includes:
*CRIO Systems: A sensor network utilizing predictive analytics for enhanced visibility and monitoring.
*Connected Things: An automated data platform designed to guide complex utility decisions.
*Mira Intel: A drone inspection and predictive analytics platform to monitor structural health for ports, seawalls and waterfront infrastructure.
*NatureDots: A sensor-free AI digital twin built for utilities, aquaculture operators and conservation agencies.
*NeuralX: An artificial life simulation engine for the monitoring of fish health, biomass and feed optimization.
Storm watch
CRIO Systems is the developer of DrainWatch, which combines machine learning and low-cost sensing to give engineers insight into their sewer networks. Considering emergency repairs are much more costly than scheduled maintenance, it’s incumbent upon utilities to maintain visibility across their sewer networks, noted company co-founder Beth Martini.
The startup represents a pivot for Martini, who transitioned to industrial design from hospitality during the COVID pandemic. Joined by civil engineer Michael Sansone and Trevin Cox, an aerospace professional, Martini helped develop a “sparse sensing” model that utilizes carefully placed sensors rather than a dense sensor array.
This system is both cost-effective and highly reliable for the mid-sized and under resourced communities that are the company’s focus, said Martini.
“We’re not putting sensors in every manhole, where we need to monitor the data densely,” Martini said. “With DrainWatch, we are limiting the sensoring and showing conditions network-wide. You don’t need enterprise platforms where utilities have to invest millions of dollars.”
Deploying 15 to 20 DrainWatch sensors across a two-to-four-square-mile area typically costs around $40,000 – in comparison, the large-scale sensor networks purchased by many cities go for upwards of six figures. Through DrainWatch, even underserved municipalities can get ahead of shifting climate patterns and rapid spikes in stormwater runoff, Martini said.
The water tech entrepreneur need only reference CRIO Systems’ Greater Chicago operational hub, which was hit by multiple 500- and 1,000-year rain events in summer 2025. Historical data is no longer sufficient with history itself changing so rapidly, she added.
“Storm frequencies and intensities are exceeding even the most conservative projections, so those trends no longer describe our present reality,” Martini said. “DrainWatch provides real-time data immediately, and that begins to build a dataset that is current and rapidly updating. This new data can be compared against historical data and a more accurate representation of network capacity becomes available.”
Staying ahead of the game
Brainchild of Kate Myers, Ed Devinney, Ray Williams, Connected Things similarly targets small-to-medium-sized utilities that lack access to modern digital tools. By tapping into digital water meters and connected field sensors, the startup’s platform unifies fragmented data to deliver immediate answers and speed up operational decision-making.
“There’s an AI agent on top that’s tuned into water-specific modeling,” Williams said in an interview with Water Daily. “I can ask the AI agent a question, and it will give me insights I couldn’t get from a static dashboard.”
Via the tech accelerator, Connected Things is collaborating with smart water meter manufacturers to streamline data extraction, ideally enabling utilities to detect leaks early and schedule targeted repairs.
Knowing how much water has been consumed can remove critical conservation barriers while ensuring water loss does not get passed down to the consumer, said Williams.
“If you’re losing trillions of gallons a year, that’s also a cost to the consumer,” he said. “Some utilities will say that’s no big deal, but to us, water is a resource we don’t want to see wasted. With AI generating these insights, we have a provenance of data to collect. As we roll out our AI, we want to make sure it’s accurate, and that it comes with trust.”
Outdated sewer systems combining stormwater with wastewater make Great Lakes municipalities especially susceptible to climate change impacts, said Jeffrey of gener8tor. The innovations emerging from his accelerator can provide an operational baseline in an environment undergoing rapid change, he said.
“Better stormwater and sewer monitoring can be an early-warning system before (a sewer network) hits capacity,” Jeffrey said. “Small utilities that don’t have a big staff are able to get this predictive insight.”
