Skip to content
‹ Back to IndustriesEnergy

AI now accounts for more than 90% of Gradiant's core water business

Gradiant now draws more than 90% of its core business from AI data centers, chip fabs and their power plants. Three emerging technologies now sit in a unit called Entropy. Because backlog already covers most of a $150 million 2027 earnings forecast, AI developers in dry regions should treat water infrastructure as an early site-planning decision.

This story was produced through MarketScale. See how Energy teams put it to work with Customer Stories & Case Studies.

By MarketScale Newsroom · GradiantEntropyData Center Water TreatmentSemiconductor Fab Water
Share

Key takeaways

01

When more than 90% of a water treatment company's core business comes from data centers, chip fabs and their power plants, water belongs on the same site checklist as the electrical interconnect.

02

With most of its projected 2027 operating-business EBITDA already covered by backlog, the numbers argue for starting water infrastructure planning early for 2028-and-later sites—alongside power, not after the electrical interconnect is secured.

Free workspace

Turn your Energy expertise into content.

Record interviews, organize footage, and write with AI on a free trial of the MarketScale platform for qualifying companies. No demo required, no credit card.

Try it Free

More than 90% of Gradiant's core operating business now comes from customers building AI infrastructure: hyperscale data centers, semiconductor fabs, and the power and industrial plants that keep them running. On November 19, the Boston-based water treatment company reorganized around that fact, Data Center Frontier reported. It is focusing the operating business on water for the AI economy and moving three earlier-stage technologies into a separate platform called Entropy.

Co-founder and CEO Prakash Govindan summed it up in the announcement: "AI runs on chips, and chips run on water." For anyone siting a data center or running a fab, the useful question is narrower. What does it mean when a specialist supplier in this niche already does most of its work for AI customers, and how much room does it have for the next project?

Most of 2027 is already booked

Gradiant shared figures that are unusually specific for a reorganization announcement. Its existing order backlog already covers most of its projected 2027 earnings for the operating business. The earnings measure throughout is EBITDA, earnings before interest, taxes, depreciation and amortization, so these are operating figures before financing costs and equipment write-downs.

Yahoo Finance, in coverage summarized by Dealroom, added a growth figure that helps explain how the company got to this point. The same summary describes Gradiant as a water-tech unicorn.

Gradiant's operating business by the numbers

About $150 million
Forecast 2027 EBITDA for the AI-focused operating business
Above 20%
Targeted 2027 EBITDA margin
About 35%
Share of the 2028 earnings forecast already covered by backlog
3x
EBITDA growth from 2024 to 2026

Data Center Frontier; Yahoo Finance via Dealroom.co

Put together, the grid gives a sense of scale. Simple arithmetic on the two headline figures puts implied 2027 revenue below $750 million, since a margin above 20% on $150 million of EBITDA requires revenue under that line. Earnings that tripled in two years and a backlog reaching well into 2028 point to a business that has been taking on AI projects quickly and has committed much of its near-term delivery capacity.

Development teams scheduling sites for 2028 and later should take note of this. A supplier with most of next year's earnings booked and about a third of the following year's in hand may still have plenty of capacity. Gradiant has said nothing to suggest otherwise. But the numbers argue for starting the water infrastructure conversation early in site planning, alongside power, rather than after the electrical interconnect is secured. That matters most for teams with projects in water-stressed regions, where the treatment design can decide whether a site works at all.

A photo caption on an earlier Data Center Frontier podcast page offers one clue to how the company packages faster jobs. It shows a mobile, turnkey Gradiant treatment unit installed at a large integrated refining and petrochemical complex in Malaysia. Gradiant delivered the 3.6-megaliter-per-day system in 21 days, and it produces demineralized water used for power generation. The unit runs on Gradiant's Selective Contaminant Extraction (SCE) and RO Infinity (ROI) technologies, which allowed the facility to get operations back up quickly while still meeting strict water-quality requirements.

A refinery is not an AI campus, and a mobile unit is not a permanent campus plant, so those 21 days say little about hyperscale schedules. What the example does show is a packaged format for power-generation water, the same category of work Gradiant counts toward its AI-related figure when the plant serves a data center or fab. For teams whose campus depends on a dedicated power plant, the water for that plant belongs in the same early conversation.

West Texas as the working template

The reorganization follows a contract Gradiant disclosed in September to provide integrated water supply and wastewater treatment for a hyperscale AI data center campus in West Texas. The design is zero-surface-discharge, meant to make data center development workable in a region with limited water.

One contract covers both getting water into the campus and dealing with what comes out, instead of a single treatment system bolted onto someone else's plan. Data Center Frontier noted that as AI construction moves into markets with water availability challenges, treatment, recycling and wastewater management are becoming bigger factors in data center site selection and operations.

Consultancy Roland Berger, in its water management overview, frames the problem as a chain: accessing, treating, delivering, disposing of and replenishing water, each step growing more complex in an era of increasing water stress. Gradiant's West Texas scope spans several links in that chain at once, from supply through treatment to discharge. For a site team, that could mean fewer handoffs between whoever brings water in and whoever answers for what leaves.

Zero surface discharge is built into the design of the West Texas project, with the aim of making data center development possible in a region where water is constrained. That makes the treatment and reuse system a core part of the campus rather than an add-on, and it shapes how a dry-region site gets permitted and operated.

When more than 90% of a water treatment company's core business comes from data centers, chip fabs and the power plants behind them, water belongs on the same site checklist as the electrical interconnect. The West Texas project suggests a water-constrained location doesn't have to be ruled out, provided the supply and discharge plan is designed in from the start. For operators weighing candidate sites, that plan could become one more line in the comparison.

Not every answer to that line runs through a treatment contract. Data Center Frontier has also covered a partnership between Emergence Water and Nimbus Advanced Process Cooling Systems that pairs atmospheric water generation with highly water-efficient adiabatic cooling, aimed at helping large AI and hyperscale facilities balance cooling performance, energy efficiency and water consumption. That approach works on how much water the building uses and where it comes from. Gradiant's West Texas contract works on campus-scale supply and discharge.

Set side by side, the two suggest a practical way to split the scoping work for a dry site. First, what does the cooling design actually need? Second, who handles the water going in and coming out?

Entropy takes the long-horizon bets

Entropy, led by CEO Anurag Bajpayee, stays inside the Gradiant group. Each of its three businesses gets a dedicated team and its own development resources. alkaLi recovers lithium and other critical minerals from brines and wastewater. ForeverGone focuses on destroying PFAS contaminants. Turing applies artificial intelligence to running water treatment plants. According to the Yahoo Finance coverage, other early-stage technologies may join the platform later.

For current customers, Gradiant says nothing changes: the same contacts, the same project teams, the same contracts. The split gives the business that delivers infrastructure projects its own focus and separates it from the businesses still being developed. Turing is the one closest to a facility operator's daily work, because it targets plant operations instead of a new treatment chemistry.

The 90% figure has limits worth keeping in mind. It describes one company's customer mix, and its definition of AI-related is broad, covering the power and industrial facilities that serve data centers and fabs as well as the facilities themselves. It shows where Gradiant is putting its attention. It doesn't measure how much water the AI sector uses, and the three Entropy businesses are described as emerging, so their value to data center buyers is still ahead of them.

Before shortlisting a water partner for an AI campus, ask three things: whether their delivery schedule has room for your energization date, whether the scope covers both supply and wastewater discharge, and which entity in the group holds your contract.

The 2028 figure to track

The number to watch is that 35%. If later updates show backlog coverage of 2028 climbing well ahead of the calendar, demand for AI water infrastructure is filling the outer years of Gradiant's order book, and the case for early procurement gets stronger.

The other signal is what shows up next inside Entropy. A new technology joining alkaLi, ForeverGone and Turing would show which water problems Gradiant expects its AI customers to bring next.

Featured companies

Your experts belong here

Every story in MarketScale Energy starts with a company putting its field engineers, operations leads, and project developers on the record. Buyers are already reading this topic. The only question is whose experts they find.

Developers and operators shortlist on credibility, and your engineers give your sales team something real to send.

Book DemoSee how it works15 minutes, straight to a calendar.

About the author

MarketScale Newsroom
MarketScale NewsroomEditorial Team, MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

B2B Weekly

The week in Energy, and sixteen other industries, every Monday.

Ten stories, one-line takes, five minutes. Free.

Energy: are you visible to AI?

Before they reach out, Energy buyers ask AI engines which vendors to trust. Explore how your experts, customers, and partners can become useful content for buyers and AI search.

Free Trial

You just read one Energy expert. Your company is full of them.

This article was produced through MarketScale. The same platform turns your field engineers, operations leads, and project developers into the articles, video, and social content Energy buyers are searching for. Start a free trial and see it with your own people. For qualifying companies, no credit card, no demo required.

NPS +73 · 1,000+ creators · 38+ countries

What your free trial includes

Hands-on access to the MarketScale platform
Media requests to your crowd, remote recording, AI writing tools
No demo required. No credit card.
For qualifying companies. Company confirmation required.

More Energy Insights

PSE&G cut 38.15 MW in a single one-hour demand response event

PSE&G cut 38.15 MW in a single one-hour demand response event

PSE&G closed its second demand response season with nearly 45,000 New Jersey customers enrolled, up from about 23,000 in 2025. One summer event cut 38.15 megawatts for an hour. PSE&G says the experience will help inform its peak demand work as New Jersey advances its virtual power plant program.

  • 01Averaged across ~45,000 enrollments, the 38.15 MW one-hour event implies under 1 kW per enrolled customer (a blended estimate, not per-participant performance).
  • 02Enrollment grew from about 23,000 customers and just over 30,000 devices to about 45,000 customers and 60,000 devices, which suggests most of the growth came from new households.
  • 03For operators, the question is whether demand reductions keep pace with enrollment as the pool grows, which could shape how much a virtual power plant can rely on it.

Nov 21, 2026

Why WEF’s emerging energy techs deliver more in combination

Why WEF’s emerging energy techs deliver more in combination

The World Economic Forum's Top 10 Emerging Technologies for Planetary Health: Energy Edition, released at COP31, names ten clean-power technologies for 2030. They deliver most when deployed together. Two of the three nearest deployment store energy, and the third is grid software, so factory energy buyers weighing heat batteries feel it first.

  • 01Two of the three technologies closest to deployment on WEF's 2026 energy list store energy, and the third is the software that decides when they charge.
  • 02Battery fire behavior is becoming a siting question: EnerVenue says its nickel-hydrogen system won a multi‑MWh oilfield tender where lithium‑ion couldn’t be used due to fire risk.

Nov 20, 2026

Two-thirds of Tantalus gateway buyers are upgrading systems they already own

Two-thirds of Tantalus gateway buyers are upgrading systems they already own

Tantalus Systems held third-quarter 2026 adjusted revenue flat at $14.3 million after meter shortages and a utility budget deferral pushed shipments out. Budget-bound utilities are building on what they own. About two-thirds of the 92 utilities ordering its TRUSense Gateway are existing customers upgrading installed systems.

  • 01About two-thirds of the 92 utilities ordering the TRUSense Gateway already run a Tantalus network, suggesting many co-ops and municipal utilities facing finite annual budgets are prioritizing upgrades to existing systems and sequencing new deployments as they add analytics features over time.
  • 02Meter supply is now a scheduling risk for AMI projects: one major meter partner's line was down for about six weeks, another cut factory allocations, and memory lead times are stretching at contract manufacturers.
  • 03Utilities that paid IEEPA tariff charges on grid hardware have a fair question for every vendor: Tantalus turned its refunds into about $1.4 million in credits toward eligible customers' future purchases.

Nov 20, 2026

Explore More Energy Insights

Read more expert perspectives from across Energy.

Browse Energy Hub

About the Expert

MarketScale Newsroom
MarketScale Newsroom

Editorial Team

MarketScale

The MarketScale Newsroom reports on the companies, technologies, and trends shaping 16 B2B industries. It turns primary sources and expert commentary into clear, useful coverage for the people doing the work.

For B2B teams

Your experts could be publishing here

Stories like this one run on content MarketScale captures from real practitioners. See how your team's expertise becomes coverage in Energy and beyond.

Book a Demo

Or call us. No forms required. We pick up. 214-945-2512