John Khazraee
CEO & Founder
15+ years building data center infrastructure. Engineer, MBA. U.S. Veteran.
Predictive liquid-risk intelligence
Predict liquid failures before they become outages.
AI that learns how every cooling loop should behave and warns you 24–72 hours before it fails.
Techstars Alabama EnergyTech · Fall 2026 class
SXSW 2026 · First Place Innovative World Technology Winner
TechCrunch Disrupt 2025 · First Place Industry Award Winner
NVIDIA Inception member
Accelerator
Techstars Alabama EnergyTech, Fall 2026 — 1 of 57 companies selected from 16,000+ applicants.
Read more →#1
Water damage is the most frequent data center insurance claim.
Allianz Commercial, 2026~24%
of data center loss costs are liquid-related.
Swiss Re sigma, 20262×
Liquid damage is more likely than fire to hit a business.
FM, 2026
The stakes · Paris, April 2023
One leaking pipe took a cloud region offline.
A cooling-water leak flooded a battery room and started a fire. 58% of the zone’s servers went down, and the region was offline for more than a day.
- 4:46 pm
Cooling-water pipe starts leaking
- 7:14 pm
Smoke detected
- 8:08 pm
Battery room flooded and on fire
- 11:04 pm
Building shut down
Six hours from leak to shutdown. Every hour of warning before 4:46 pm is time to act.
Times in US/Pacific, from the operator’s public incident report and contemporary trade-press coverage.
The status quo
Today’s tools only react once liquid is already out.
- 1
Normal
operation - 2
Abnormal
pattern - 3
Leak
forming - 4
Alarm
- 5
Damage &
downtime
MayimFlow
24–72 hours of warning. Fix it in a planned window.
Today’s tools
Leak cables, spot sensors, alarms and shutoff valves find liquid after it escapes.
The solution
An Apple Watch for pipes.
MayimFlow learns each liquid system’s normal rhythm and flags the abnormal one.
1
Sense
Clip-on sensors or the BMS / SCADA data you already collect.
2
Predict
AI scores failure risk for every loop in real time, cross-checking signals to keep false alarms down.
3
Act
Ranked alerts and repair tickets arrive before failure, on the dashboard or a phone.
Hardware-agnostic
Built for liquid cooling
Read-only to start
2–4 week deployment
The platform
One screen from sensor signal to repair ticket.
See which loop is drifting, how much time is left and what to do next.
- A live risk score for every cooling loop
- Ranked alerts with a predicted time to failure
- Repair tickets your team can act on
- Works with BMS, SCADA and DCIM data
Six signals, one risk score
The vital signs of a liquid system.
Like a smartwatch for your pipes: MayimFlow reads their vital signs to spot trouble early.
Like Pulse
Vibration
Pump and valve health, cavitation and loosening fittings.
Like Blood pressure
Pressure
Drops and swings that point to restrictions or leaks.
Like Cardiac output
Flow
Ultrasonic, clamp-on measurement of what is actually moving.
Like Cholesterol
Conductivity
Dissolved solids and contamination in the loop.
Like Blood acidity
pH
Coolant chemistry that drives corrosion.
Like Body scan
Thermal imaging
Hot spots and wet spots the other sensors can’t see.
The AI layer
From an odd reading to the exact fitting.
- Find abnormalities. Spot the pattern that breaks a loop’s normal rhythm.
- Locate the leak. Trace the signal back through the pipe network to the likely component.
- Visual inspection. Read thermal and camera images for wet spots and hot spots.
- Guided action. Recommend the next step and open the repair ticket.
Integrations & security
Works with the systems you already run.
Start read-only on the data you already have. Add clip-on sensors only where there are gaps.
Security questions? [email protected]
- PLCs
- Schneider Electric, Siemens, Allen-Bradley, Beckhoff, Opto 22, Phoenix Contact, WAGO
- Systems
- BMS, SCADA and DCIM via API
- Weather
- OpenWeather, Tomorrow.io, Open-Meteo
- Read-only
- Nothing changes setpoints, opens valves or stops pumps unless your team decides it should.
- You own it
- Facility data is encrypted in transit and at rest, never sold or shared, exported or deleted on request.
Where we work
AI data centers first. Then every liquid system.
Liquid cooling turns every rack row into a plumbing system.
AI data centers
CDUs, cold plates, immersion tanks and facility water loops, watched around the clock.
Utilities
Pumps, mains and plant water systems.
Hospitals & buildings
Risers, chillers and plumbing in occupied spaces.
Manufacturing
Process water and cooling for production lines.
Backed by
Techstars Alabama EnergyTech, Fall 2026.
MayimFlow was selected for the Techstars Alabama EnergyTech Accelerator — one of eight energy companies in the Fall 2026 class, and 1 of 57 companies chosen from more than 16,000 applicants worldwide.
Fall 2026 · Alabama EnergyTech
The program puts MayimFlow next to the utilities, grid operators and energy-infrastructure mentors who run the systems we predict failures in — and in front of the capital that scales a liquid-risk platform from pilot sites to fleet-wide deployment.
Through the accelerator network, Techstars mentors and the wider Techstars portfolio, every pilot gets the operating experience of people who have run critical water and cooling plant at scale.
1 of 57companies in the Fall 2026 Techstars class
16,000+applicants worldwide
8energy companies in the cohort
Techstars '26
Alabama EnergyTech
NVIDIA Inception
Nasdaq Milestone Makers
Why we’re building this
We’ve seen leaks firsthand. Then we saw one move a data center.
Visiting a new AI data center, we learned the previous one had flooded from a water-cooling leak, and the team had to move into a new facility.
15+ years building data center infrastructure at Microsoft, Oracle and IBM. As AI racks go liquid-cooled, every rack becomes a plumbing problem.
MayimFlow exists so operators see the warning signs before the floor gets wet.
John KhazraeeCEO & Founder
The people
Meet the team
Jeremy Jason
Chief Marketing Officer
20+ years in technology sales and marketing.
Anahita Sadaf
Business Development
UC Berkeley alumna.
Our scientific advisor
Dr. Butrus (Pierre) Khuri-Yakub
Professor Emeritus of Electrical Engineering, Stanford University
Inventor of the CMUT (capacitive micromachined ultrasonic transducer) and co-inventor of a clamp-on ultrasonic flow meter, with 107 patents.
His CMUT technology has been commercialized by
Our team’s experience
Recognition
Recognition that opened doors.
Fall 2026TechstarsAlabama EnergyTech Accelerator · 1 of 57 from 16,000+ applicants
Mar 2026SXSW PitchFirst Place Innovative World Technology Winner
Oct 2025TechCrunch DisruptFirst Place Industry Award Winner
Mar 2026NVIDIA InceptionProgram member
- Oct 2025
TechCrunch Disrupt
First Place Industry Award Winner
- Mar 2026
SXSW Pitch
First Place Innovative World Technology Winner
- Mar 2026
Financial Times
Featured in SXSW coverage
- Mar 2026
NVIDIA Inception
Accepted into the program
- Apr 2026
eMerge Americas
Top 100 startup
- Jun 2026
Summerfest TechAI
Top 5 Global finalist and featured speaker
- Summer 2026
Nasdaq Entrepreneurial Center
Milestone Makers graduate
- Aug 2026
Plug and Play
AI for Physical Industries
- Fall 2026
Techstars
1 of 57 companies selected from 16,000+ applicants; Alabama EnergyTech Accelerator
Selected by Techstars: 1 of 57 companies from 16,000+ applicants
MayimFlow joined the Techstars Alabama EnergyTech Accelerator, one of eight energy companies in the Fall 2026 class.
What the program means for us
MayimFlow joins NVIDIA Inception
Access to NVIDIA’s GPU platform, developer tools and network.
FAQ
Common questions
What does MayimFlow do?
MayimFlow predicts liquid failures such as leaks, failing seals and pump problems in AI data center cooling loops and critical infrastructure. It learns each system’s normal pressure, flow, vibration and water-chemistry patterns and flags abnormal behavior with 24–72 hours of warning, so teams can repair in a planned window.
How is MayimFlow different from leak detection cables and spot sensors?
Leak detection cables, spot sensors and shutoff valves react after liquid has already escaped. MayimFlow looks for the early signs that come before a leak, such as pressure drops, flow changes and rising vibration, so the problem can be fixed before there is water on the floor.
What sensors or data does MayimFlow need?
MayimFlow is hardware-agnostic. It can start with the BMS, SCADA and DCIM data you already collect, then add clip-on sensors for vibration, pressure, ultrasonic flow, conductivity, pH or thermal imaging where there are gaps.
Does MayimFlow control our cooling equipment?
No. MayimFlow starts read-only, so nothing in your control systems changes until your team decides it should.
How long does a deployment take?
A typical deployment takes 2–4 weeks from kickoff to live monitoring.
Which liquid-cooling systems does MayimFlow support?
MayimFlow is built for liquid cooling, including coolant distribution units (CDUs), direct-to-chip cold plates, immersion cooling and facility water loops.
Is MayimFlow only for data centers?
AI data centers come first. The same pressure, flow and vibration physics apply to utilities, hospitals, commercial buildings and manufacturing sites, which MayimFlow also supports.
Who backs MayimFlow?
MayimFlow is part of the Techstars Alabama EnergyTech Accelerator, Fall 2026 class, and a member of NVIDIA Inception.
Every liquid system, predicted.
See what MayimFlow would watch in your facility. We’ll walk through the dashboard and map your loops — backed by the Techstars Alabama EnergyTech Accelerator, Fall 2026.