Water intelligence platform

The pipes are telling you
something. We translate.

Amarine brings AI-native intelligence to municipal water infrastructure — detecting leaks before they surface, predicting contamination events, and optimising treatment in real time.

30%
avg. water lost to leaks
$7.6B
annual NRW cost, US alone
93%
AI detection accuracy
1:2
water main break frequency (min)

The grid that
keeps cities alive
is running blind.

Water utilities manage tens of thousands of kilometres of buried pipe, but most of that network is monitored with tools designed before the internet existed. Operators make decisions from lagging data, paper logs, and guesswork.

$2B+
lost annually to non-revenue water in the US — treated, pumped, and chemically processed water that generates zero revenue for the utility.
01 / LEAK DETECTION
Leaks stay hidden for months
Traditional walking surveys catch less than half of active leaks. Most are discovered when they surface — by which point thousands of litres have been lost and road damage is already underway.
02 / CONTAMINATION
No advance warning of quality events
Contamination events are reactive. Utilities sample periodically; the window between a pressure drop, backflow, or chemical intrusion and a public health alert is days — not hours.
03 / TREATMENT
Treatment plants running on fixed schedules
Chemical dosing, filtration cycles, and energy consumption are tuned quarterly at best. Dynamic demand, seasonal inflow, and variable raw water quality are handled manually.
04 / FLOODING
Early warning comes too late
Stormwater overflow events and combined sewer flooding cause millions in infrastructure damage each year. Most systems lack the sensor density and predictive models to act before the threshold is breached.

One intelligence layer.
Four critical systems.

Amarine integrates with your existing SCADA, AMI meters, and IoT sensor arrays — adding a predictive AI layer that reasons across the entire network in real time.

Built specifically for municipal utilities. No rip-and-replace. No 18-month deployment timelines. Amarine is deployable alongside legacy infrastructure and is designed to meet the regulatory, audit, and procurement requirements water authorities operate under.

SCADA Integration AMI / Smart Meter IoT Sensor Network DMA Analysis REST API On-Premise Available
Leak Intelligence
LeakPulse
Continuous acoustic and pressure analysis across district metered areas. Detects leaks within 1–3 days of onset — not months. Pinpoints location to within metres, eliminating speculative excavation.
Water Quality
AquaGuard
Multi-parameter quality monitoring paired with contamination propagation modelling. Detects anomalies hours before they manifest at the customer tap, with automated regulatory alert routing.
Treatment Optimisation
FlowMind
Adaptive treatment plant intelligence. Dynamically optimises chemical dosing, filtration cycles, and energy consumption based on real-time inflow quality, demand forecasts, and regulatory thresholds.
Flood Early Warning
StormSense
Hydrological AI combining weather feeds, catchment topography, sewer capacity, and historical event data. Issues flood warnings 6–18 hours ahead, with actionable diversion and response protocols.

How Amarine
sees your network

STEP 01
Sensor Ingestion
Pressure loggers, acoustic sensors, AMI meters, SCADA feeds, and weather APIs stream into a unified data layer.
STEP 02
Network Modelling
Hydraulic and topological models build a live digital twin of the distribution network.
STEP 03
Anomaly Detection
Machine learning models identify deviation from expected baselines, scored by probability and severity.
STEP 04
Predictive Reasoning
Events are classified, geolocated, and cross-referenced against maintenance history, pipe age, and material.
STEP 05
Operator Action
Prioritised work orders are issued directly to field crews with mapped locations and remediation guidance.
93%
Leak detection accuracy
50%
NRW reduction potential
18hr
Flood early warning window
<8wk
Typical deployment timeline

Built to fit the
world utilities
actually live in.

01
No-displacement integration
Amarine connects alongside existing SCADA, GIS, and billing systems. We add intelligence without requiring utilities to retire legacy infrastructure — capital programs that stall deals are not part of our model.
02
Regulatory-grade audit trails
Every detection, alert, and recommendation is fully logged with timestamps, confidence scores, and supporting data. Built to satisfy EPA, DWI, and Ofwat audit requirements from day one.
03
Sector-specific AI, not generic ML
Models are trained on water network data — acoustic leak signatures, hydraulic pressure profiles, treatment chemistry behaviour. Not repurposed industrial IoT platforms fitted with a water skin.
04
Sovereign deployment options
For utilities with data sovereignty requirements, Amarine deploys on-premise or in a dedicated private cloud environment. Network telemetry never leaves the utility's control perimeter.
Why now

Water utilities are facing a convergence of pressure: ageing infrastructure, climate volatility, tightening regulation, and stretched maintenance budgets. The technology to address all four exists — what's been missing is a platform that makes it deployable inside procurement-constrained public institutions.

Market context

Global water utility software is a $4B+ market growing at 9% annually. The average utility still runs on systems from the 1990s. Less than 15% of municipal networks have deployed any form of continuous acoustic monitoring.

Our edge

We go below the surface where traditional survey crews stop. Acoustic ML models trained on two million+ signatures. Real-time DMA minimum night flow analysis. And a field-operations layer that turns detections into crew-ready work orders within minutes.

AS
Aakarshan Saxena
Founder & CEO · Amarine
"Water infrastructure is the most critical system most people never think about — until the day it fails."

Aakarshan Saxena founded Amarine with a conviction that the water sector's technology gap is not an unsolvable problem — it is an unsolved one. Having studied the technical and institutional dynamics of water utilities across South Asia, Europe, and North America, he identified a consistent pattern: the data exists, the sensors exist, but the intelligence layer that turns network signals into operator decisions has been absent.

Amarine is the system Aakarshan set out to build: one that meets utilities where they are — on 30-year procurement cycles, with 1990s-era SCADA systems and teams stretched across thousands of miles of pipe — and makes AI-grade water intelligence deployable within months, not years.

His focus is on the intersection of machine learning and physical infrastructure, with particular depth in acoustic signal processing for subsurface anomaly detection and hydraulic network modelling for large-scale distribution systems. He believes the hardest part of modernising critical infrastructure is not the technology — it is building the institutional trust to use it.

Hydraulic Engineering Acoustic ML Municipal Utilities IoT Infrastructure B2B SaaS

Let's talk about
your network.

We work directly with water utilities, municipal authorities, and infrastructure investors. Tell us about your network and we'll configure a pilot proposal within two weeks.

General Enquiries
Pilot Programme
Investment & Partnerships