Virtual Submetering: Equipment-Level Energy Data Without the Hardware
How AI creates accurate equipment-level energy meters from data you already have — replacing $500-2,000 per-point hardware installations with software.
The Submetering Gap
Every energy manager knows the problem: you can't optimize what you can't measure. But installing physical submeters on every piece of major equipment costs $500 to $2,000 per point, requires electrical shutdowns, and takes months to deploy across a portfolio.
The result? Most buildings have a main meter and maybe a handful of submeters. The rest is guesswork.
What Virtual Submetering Does
Virtual submetering uses AI to disaggregate your main meter data into equipment-level consumption — no hardware required. By analyzing the electrical signatures, operating schedules, and BAS data you already collect, the AI can determine how much energy each major system consumes at any given time.
How It Works
The process relies on three data sources most buildings already have:
- **Main meter data**: The total building electrical consumption at 15-minute or higher resolution
- **BAS trend data**: Equipment status, setpoints, valve positions, fan speeds — the operational data your building automation system already logs
- **Weather data**: Outdoor conditions that drive HVAC loads
The AI learns the relationship between equipment operation and energy consumption. When AHU-3 ramps up its fan from 60% to 80% speed while the cooling valve opens to 70%, the model knows what that costs in kWh. Multiply across every piece of equipment, and you have a complete picture.
Accuracy
Virtual submeters typically achieve 90-95% accuracy compared to physical hardware — sufficient for identifying savings opportunities, tracking performance, and benchmarking equipment.
For most applications, this accuracy is more than enough. You don't need $2,000 hardware accuracy to know that your chiller is consuming 30% more energy than it should.
Real-World Applications
Tenant Billing Allocate energy costs to individual tenants based on actual HVAC consumption in their zones, not just square footage. Fair billing improves tenant relationships and creates incentives for conservation.
Equipment Benchmarking Compare similar equipment across your portfolio. If Building A's chiller plant runs at 0.65 kW/ton while Building B's identical plant runs at 0.85 kW/ton, you know where to focus your optimization efforts.
Savings Verification When you implement an energy conservation measure, virtual submetering provides the before-and-after data to verify actual savings — critical for utility incentive programs and ESG reporting.
The Economics
Consider a 500,000 sq ft commercial office building with 40 major pieces of HVAC equipment:
- **Physical submetering**: 40 points x $1,000 average = $40,000 hardware + $15,000 installation + 3 months of downtime coordination
- **Virtual submetering**: $0 hardware, deployed in the software you already run, operational in days
The question isn't whether virtual submetering is good enough. It's why you'd install hardware when you don't have to.