(don't) Waste Water! | Water Tech to Solve the World

On April 30, 2026, CRH plc (NYSE: CRH) announced a $700 million agreement to acquire Axius Water from KKR and XPV Water Partners - instantly making the Irish-American building materials giant one of the largest water infrastructure players in North America. If CRH acquires Axius, what's next?

Most investors missed it, so I thought I'd decode it!
Here are the 10 things you need to see.

So you got it, CRH acquires Axius. Here's the deal in 30 seconds:
• Buyer: CRH plc - a $77B market cap building materials company most water investors don't track
• Target: Axius Water - a wastewater treatment platform built by KKR and XPV Water Partners through a multi-year roll-up
• Price: $700M (reported)
• Sellers: KKR (Global Impact Fund) and XPV Water Partners
• Strategic logic: bolts wastewater treatment onto CRH's existing water infrastructure portfolio (Hydro International, Oldcastle Infrastructure)

What's actually new here - and why it matters for water investors:
00:00 - On your Bingo Card?
02:30 #1: The Cast
07:35 #2: The Substance
11:40 #3: The Number
16:23 #4: The Timing
20:54 #5: The Macro Reality
24:40 #6: The Credibility
28:26 #7: Inside the Platform
32:28 #8: The Synergy Story
35:17 #9: The Risks
39:58 #10: The Sector Signal

📰 Sources & further reading:
• CRH Q1 2026 earnings call transcript: https://www.crh.com/investors/results-presentations/
• KKR Global Impact Fund: https://www.kkr.com/approach/sustainability/sustainable-investing/trends
• XPV Water Partners portfolio: https://xpvwaterpartners.com/companies
• Axius Water: https://www.axiuswater.com/
• My newsletter on Water Infrastructure: https://www.linkedin.com/pulse/us-water-bill-4x-what-everyone-thinks-antoine-walter-3rqnc/

🎙️ If we don't know each other, Hi, I'm Antoine Walter! I help newcomer investors decode the water tech sector. Each week, I break down a water deal (like here, when CRH acquires Axius), technology, or company through an investor-forward lens. No acronyms, no jargon (hopefully) just the signals that matter.

🔔 Subscribe for weekly water investment briefs: youtube.com/channel/UCsMC1BYAun2JtMT177jTGOw/
🎧 Podcast version: https://feed.ausha.co/br23DCZ1GnG3
💼 LinkedIn: https://www.linkedin.com/in/antoinewalter1

⚠️ Disclaimer: This video is for informational and educational purposes only. Nothing here constitutes investment advice. Always do your own research.

#CRH #AxiusWater #KKR #XPVWaterPartners #WaterInvesting #WaterMA #WaterStocks #WaterInfrastructure #WastewaterTreatment #WaterTech #PrivateEquity #ImpactInvesting #ESG #WaterSector

Hosted on Ausha. See ausha.co/privacy-policy for more information.

Everyone says AI is drinking the planet dry, right? Well, the numbers say your morning coffee uses 29,600× more water than a ChatGPT prompt!

In this episode, I sit down with Alex Passini to pressure-test the dominant narrative around AI's water footprint - and what we found completely flips the story the media is telling.

Here's what most "AI water crisis" headlines miss:
→ A single ChatGPT prompt uses roughly 16 milliliters of water. One cup of coffee uses ~140 liters when you count the beans. That's a 29,600× gap.
→ Of the water a hyperscale data center consumes, ~75% isn't used by AI at all — it's used upstream for the energy that powers it. Blame the grid, not the GPU.
→ One banana = ~6,250 ChatGPT prompts. One almond = ~12 prompts. The water-per-prompt math is rounding-error territory next to your lunch.
→ Florida already reuses 800 million gallons/day of treated wastewater. Data centers aren't the threat - they're an accelerant for the water reuse capex the sector has been waiting twenty years for.

If you're an investor trying to figure out whether AI's water story is a real thesis or a media artifact, this episode gives you the framework (and the numbers) to decide.

🧠 What you'll learn:
- Why "data center water consumption" and "AI water consumption" are not the same thing
- The actual per-prompt footprint of generative AI (modeled, not estimated)
- How water reuse technology turns the alarmist narrative into an investment opportunity
- Where the real bottleneck sits - and which sub-sectors stand to benefit

🎙️ Guest: Alex Passini (https://www.linkedin.com/in/alexpassini/)
Alex is Vice President of Business Development at CSA Group
CSA Group is hiring: https://csagroup.com/careers/
(I promised Alex in the podcast to put this in the show notes!)

📰 The newsletter behind the show: https://www.linkedin.com/newsletters/don-t-waste-water-6884833968848474112/
🎧 Podcast feed: https://feed.ausha.co/br23DCZ1GnG3
💼 LinkedIn: https://www.linkedin.com/in/antoinewalter1/
🔔 Subscribe if you want the unfiltered, numerate take on water sector investing - no jargon, no acronym soup, just the asymmetric bets worth pricing.

Hosted on Ausha. See ausha.co/privacy-policy for more information.

Every analyst, every think tank, every consulting deck has an opinion on AI's water footprint (and overall on AI and Water)

But nobody bothered to ask the people actually watching the videos, posting the comments, and shaping the narrative. You'd need to be mad to do that, right?

So I read 2,540 of them. 😅

Across 9 of the most-watched YouTube videos on AI water consumption, and overall on AI and Water.
Here's what I found:
1️⃣ The public is right about geography. Boardman, Oregon, shows up at 80 comments per minute for a reason.
2️⃣ The public is right about the executives. The -0.20 sentiment around Sam Altman's "one-fifteenth of a teaspoon" line is actually a signal.
3️⃣ The public is wrong about closed-loop cooling. The most-liked comment on the entire dataset? Hank Green publicly correcting himself. 13,000 likes for "I was wrong."
4️⃣ And the public is missing the bigger story almost entirely — the one that's about to constrain AI before water ever does.

This is the third and final chapter of a trilogy on AI, water, and the trillion-dollar infrastructure thesis hiding behind the headlines - basically on AI and Water:
📍 Part 1 - Data Center Consumption DOESN'T Matter... But Discharge Does! → [https://www.youtube.com/watch?v=1o-cdb8xzZ0
📍 Part 2 - The 2027 Deadline That's About to Reprice Water Companies → https://www.youtube.com/watch?v=1PKHUo6q5gw
📍 Part 3 - You're watching it 😅

⏱️ Chapters
00:00 — Nobody asked. So I did.
01:44 — The dataset: 2,540 comments, 9 videos, 1 spreadsheet
03:58 — What the public got wrong (closed-loop cooling and the misleading-true problem)
07:07 — The most-liked comment on AI water — and why it matters
10:53 — What everyone is missing

🔬 Methodology 2,540 individual comments analyzed 9 source videos (the most-viewed English-language content on AI water consumption as of) Sentiment scoring + theme clustering done manually, then cross-checked

🏷️ Topics covered AI water consumption · AI water usage · data center cooling · AI environmental impact · closed-loop cooling · AI energy consumption · water sector investing · sustainability · generative AI infrastructure · AI data center water · Sam Altman water · Boardman Oregon data centers

Hosted on Ausha. See ausha.co/privacy-policy for more information.

Who Buys the First Data Center Water Company? And When Does the Repricing Start?

Data center water treatment is a $1.1 billion market growing nearly 15% per year, with 60% of spending recurring - generating an "infinite money glitch" for the righty designed water tech companies. So, strategic buyers, private equity sponsors, and VC-backed platforms are racing to consolidate water tech expertise. Meanwhile, hyperscalers spending $50 billion a year on infrastructure have made zero water acquisitions... for how long?

🌶️ KEY SPICES 🌶️
💵 A $660M recurring-revenue pocket nobody priced as SaaS - 60% of data center water spend is OPEX, not capex (Sustainability at its best)
🏗️ AEA Investors' CRB Water playbook: a $13.2M Missouri distributor turned national data center water platform in 24 months through five add-on acquisitions
📈 Evoqua's 17x EBITDA exit to Xylem in 2023 as the template - and AEA ran that play too
🌐 Microsoft's water use efficiency (WUE, liters per kWh of compute) sits at 0.2, roughly 5x better than Google's 0.96 and 6.5x better than Apple's 1.3
🏢 Amazon's Kiva Systems DNA applied to water: write the blueprint internally, then acquire the builder when the capability becomes competitively load-bearing
⏱️ A 2027 zero-water pledge deadline creating a hard two-year window for Microsoft 🎯 The short list of acquisition targets: Gradiant, CRB Water, Infinite Cooling, Uravu Labs

🥜 IN A NUTSHELL 🥜
Why is data center water treatment the fastest-growing industrial water vertical? At $1.1B in 2024 and above 13% annual growth (roughly double any other industrial water segment) demand tracks the $50B/year hyperscaler capex directly.
Who's actually buying so far? Ecolab (~$7B combined on CoolIT Systems and Ovivo's electronics division), Xylem (Vacom Systems at $42M), Kemira (Water Engineering Inc. at $150M, 2.5x revenue), EQT Infrastructure (Seven Seas Water at roughly $1B), and AEA Investors (Chemtron RiverBend plus five add-ons, now CRB Water).
What makes the hyperscalers different? They spend $50B/year on data centers and run extensive water partnerships and VC positions, but have not acquired a single water company. Yet?
Why does this matter for investors? When the first hyperscaler moves, water companies with data center exposure reprice from the 7.5x EBITDA sector median to 15-20x+ strategic-platform multiples.
Who are the likely targets? Gradiant ($330M raised, 23 products, nine acquisitions, Counter-Flow RO at 99% recovery), CRB Water (PE exit setup), Aquatech or Saltworks Technologies (in the "right brain" approach), Infinite Cooling, Uravu Labs or AirJoule Technologies in the "left brain" one.

#️⃣ Mentioned Links #️⃣
- The PE Water Platforms Nobody's Talking About Yet - https://www.youtube.com/watch?v=i1o78cgd5Hs
- AEA Investors - aeainvestors.com
- CRB Water - https://www.crbwater.com/
- My conversation with Prakash Govindan (Gradiant) - https://youtu.be/Hpk_gPjm_1s?si=d_I3gYuYLYudBmQK

Hosted on Ausha. See ausha.co/privacy-policy for more information.

What Is the Real Water Problem Behind AI Data Centers - and Where Is the $1.3 Billion Opportunity?

AI Water seems to concern everybody, while data center water treatment is the fastest-growing industrial water vertical in the world - $1.35 billion in 2025, growing at 13.3% per year. But the media is chasing the wrong story. The real problem isn’t consumption volume — it’s the concentrated industrial wastewater that cooling towers produce. This episode maps the opportunity most investors are missing. A

🌶️ KEY SPICES 🌶️
💧 American lawns consume 49 times more water than every AI data center in the US combined - so why is the media only alarmed about servers?
🏭 Cooling towers evaporate pure water and discharge “blowdown” - industrial wastewater concentrated at 3-5x intake levels that almost nobody covers
💰 Ecolab acquired CoolIT Systems for $4.75 billion at 25x revenue, assembling a $10.5 billion integrated stack from resin to chip to managed service
📉 Zero hyperscalers (Google, Microsoft, Amazon, Meta) have acquired a single water treatment company despite spending $50B/year on data center construction
📊 The EU has capped data center water usage at 0.4 L/kWh in water-stressed areas — the installed base runs at 2.3-3.8, creating a 6-10x compliance gap

🥜 IN A NUTSHELL 🥜
Why do American lawns matter in a video about AI water? Turfgrass consumes roughly 12 billion cubic meters per year, 49 times more than all US data centers combined — context that reframes the media’s breathless AI-water coverage.
What is “blowdown” and why should investors care? Evaporative cooling towers discharge concentrated wastewater at 3-5x intake levels containing salts, silica, and treatment chemicals — this industrial waste stream scales linearly with every hyperscale facility under construction, creating a $1.35 billion treatment market growing at 13.3% per year.
Why did Ecolab pay $4.75 billion for CoolIT at 25x revenue? Liquid cooling eliminates evaporation but creates a new treatment market for closed-loop fluid chemistry, and Ecolab has spent $10.5 billion since 2021 assembling the only complete stack from ion exchange resin to GPU cold plate to managed service contract for AI Water
Why haven’t hyperscalers acquired a water company? Google, Microsoft, Amazon, and Meta have vertically integrated into chips, power, and fiber — but zero water acquisitions appear in the global M&A record, even as institutional capital poured $13.5 billion into water deals in 2025 alone.

#️⃣ Mentioned Links #️⃣
📌 Arthur D. Little: “AI’s Hidden Dependencies” report: https://www.adlittle.com/sites/default/files/2026-01/BLUE_SHIFT_AI_hidden_dependencies.pdf
📌 GWI's data center rubric: https://www.globalwaterintel.com/industries/data-centres
📌 Ecolab CoolIT acquisition announcement: https://tinyurl.com/fa34yeht
📌 The Ecolab/Ovivo deal, decoded: https://www.youtube.com/watch?v=i1o78cgd5Hs
📌 My deep dive on H2O Innovation (and what we learn about platform building): https://www.youtube.com/watch?v=LOvJK8BnOJo

Hosted on Ausha. See ausha.co/privacy-policy for more information.