Predictive Maintenance That Prevents Costly HVAC Failures
Wattnest reads the data a building already collects and lists every fault in plain language, weeks early, each with a dollar figure. Built for HVAC service companies. No hardware, no training period.
The Faults Buildings Hide
On one university campus building we analyzed, 20 real equipment faults were confirmed in a single season. The building's own alarms caught zero of them. The typical fault sat invisible for five months. Alarms are built for emergencies, not for slow decay.
- LBNL research estimates roughly 40% of commercial air handlers run with at least one active fault
- Undetected faults waste 15 to 30% of HVAC energy before anyone sees a symptom
- Quarterly PM visits leave months-long blind gaps between inspections
- Emergency repairs carry rush labor, expedited parts, and downtime that planned fixes avoid
The Technology Behind Early Detection
Three pillars that catch equipment problems early and keep the alert queue quiet enough to trust.
Findings From Week One
Nothing to install and no training period. Connect the data and the first findings arrive the same week, with false alarms held under one alert per unit per week.
Gets Sharper Over Time
Every building runs differently, so Wattnest learns each building's normal and improves with every week of data. Tested on 1,500+ real fault events from operating buildings.
Software Only
No hardware, no controllers, no downtime. Wattnest reads the data your building systems already record (Tridium, Niagara, SkySpark) through exports or read-only access.
Real Detection Results from Real Buildings
Wattnest is tested on real building data: 1,500+ labeled fault events from operating buildings, plus a season of data from one university campus building where 20 real faults were confirmed. The building's alarms caught none of them. Wattnest caught 19.
Tested on real operating buildings and expert-labeled fault records
Built for Enterprise HVAC Infrastructure
Data Sources
Building data systems (Tridium, Niagara, SkySpark), trend exports, REST API
Ingestion
15-minute or hourly trend data; a few months of history is enough to start
Deployment
Cloud-hosted, read-only access. No control writes, no site hardware
Models
Detection that works from week one and learns each building over time
Analytics
Plain-language evidence, degradation tracking, dollars-per-day fault pricing
Security
TLS 1.3, encrypted at rest, read-only access
Comprehensive HVAC Fault Detection
Purpose-built detection for the equipment that matters most in commercial buildings. Tested on real buildings, not lab benchmarks.
Air Handling Units
Our strongest coverage: fans, circulation pumps, temperature sensors, dampers, and economizers. Tested on 1,500+ real fault events from operating air handlers.
Chiller Plants
Centrifugal, screw, and absorption chillers with their pumps and loops. On a three-year university chiller plant record, Wattnest surfaced problems a median of 24 days before the platform already watching them.
VAVs, Pumps & Plant Equipment
Coverage from day one for VAV boxes, hydronic loops, and plant equipment: stuck dampers, simultaneous heating and cooling, and flow problems. Coverage grows as your data accumulates.
Built for HVAC Service Companies
Give your team a short, dollar-ranked fault list and give every customer a client-ready view under your brand. Serve the customers a truck-roll model can't reach profitably.
Predictive Maintenance for Every Facility
Data Centers
Continuous watch over critical cooling. Flag CRAC and chiller problems early, before they threaten uptime targets or tenant SLAs.
Commercial Buildings
Catch problems between service visits, before they affect tenant comfort or trigger SLA penalties. And stop paying the energy cost of invisible faults.
HVAC Service Providers
Add data-driven monitoring to your service offering. Reach the customers a truck-roll model can't serve profitably, and show every account the contract's value in dollars.
Send Us Three Months of Building Data
We'll send back every fault we find, each with a dollar figure. One building is enough to start.