Pani puts a brain
on the world's
water plants.

Here is a fact about water treatment that the water industry would prefer you find boring: a desalination plant is, at heart, an enormous and very expensive way to lose energy. Seawater goes in, drinking water comes out, and in between, high-pressure pumps shove that water through delicate membranes at a cost measured in megawatt-hours. The membranes foul. The pumps overwork. Hundreds of sensors report on all of it, all day, to almost nobody in particular. Somewhere in that river of numbers is the answer to "how do we run this for less money and less carbon?" - and historically, nobody had the time to read it.

Pani - legally Pani Energy Inc., founded in 2017 by Devesh Bharadwaj and based in Victoria, British Columbia - exists to read those numbers. Its product is not a pump or a membrane or a chemical. It is software. The platform, called Pani Zed, ingests the sensor chaos from a plant, builds a digital twin of the facility, and then does the thing a very tired operator cannot: it runs the math continuously and says, in effect, turn this dial, and you will save this much energy without breaking anything.

The word "pani" means "water" in Hindi and several South Asian languages, which is either the most on-the-nose company name in cleantech or a small act of honesty. The pitch is unusually literal. There is no metaphor to unpack. The company makes water plants work better, and it charges for that as a subscription.

What makes this interesting - and what makes Pani's whole thesis go - is a claim that sounds like it should be false: you can get meaningful efficiency gains from a water plant without touching the plant. No new pumps. No retrofit. No capital expenditure. No shutdown. Just a login and a set of recommendations. If you have ever watched an industrial company weigh a nine-figure equipment upgrade, you understand why "keep your infrastructure, add a brain" is a compelling sentence.

There is a temptation, in climate tech, to build the dramatic thing - the new membrane, the novel chemistry, the machine that promises to change everything if only someone will spend a fortune installing it. Pani went the other direction, and the direction is telling. It assumed the plants already exist, the infrastructure is already sunk, and the fastest path to lower emissions is not new steel but better decisions inside the steel already there.

This turns efficiency and sustainability into the same problem, which is a genuinely useful reframing. In a desalination plant, energy is the dominant operating cost and, if that energy comes from fossil sources, the dominant source of emissions. Every kilowatt-hour Zed shaves off is money the operator keeps and carbon the atmosphere never sees. You do not have to choose between the green pitch and the CFO pitch. They are the same spreadsheet.

The customers Pani has named are not experiments. They are serious industrial water players - Aquatech International, VA Tech WABAG, SAWACO Water Desalination, Aditya Birla Group, Abunayyan Holding, Murugappa Water Technology & Solutions - operating real plants in places where water is genuinely scarce. When SAWACO put Pani's decision support on its SOJECO plant near Jeddah, it was not a demo. It was a facility that helps supply a coastal city, quietly getting smarter.

The recognition followed the deployments. Pani has appeared on the Global Cleantech 100, been flagged by Global Water Intelligence as a disruptive technology, and its founder made Forbes 30 Under 30 for social impact. None of that treats water plants. But it does suggest that the boring, essential frontier of AI - the one aimed at infrastructure rather than inboxes - is a real place, and Pani got there early.