Energy & Sustainability
IoT Energy Monitoring for Arizona Data Centers: Metering, PUE and Demand in a Changing Rate Landscape
Published by IOT Arizona Research & Editorial Team

- Why Energy Monitoring Matters More in Arizona Now
- Building a Metering Hierarchy
- Measuring PUE Honestly
- Demand, Not Just Consumption
- Cooling Energy in the Desert
- Batteries, Generators and Grid Events
- Workload Flexibility
- Stranded Capacity: The Hidden Cost
- Heat Waves as Stress Tests
- Reporting That Builds Trust
- Tenant Billing and Accountability
- A Practical Roadmap
- Security for Energy Systems
- Looking Ahead
For Arizona data centers, energy monitoring is moving from a sustainability exercise to a financial and regulatory necessity. The APS rate case, whose hearing ran from May 18 to July 7, 2026, featured debate over changes to the utility's Extra High Load Factor rate, which applies to large users such as data centers and manufacturers. An administrative law judge's recommended order is expected by late November, the Commission faces a December 31 deadline to vote and new rates could take effect in early 2027. APS proposed more than a 45 percent increase for these extra-large energy users. SRP, for its part, already applies minimum billing requirements to new large accounts forecasting 20 megawatts or more.
Add a statewide political debate over data center growth, and the message for operators is clear: every kilowatt and every kilowatt-hour needs to be measured, understood and, where possible, managed. This guide focuses specifically on energy monitoring, complementing our broader article on IoT for Arizona data centers.
Why Energy Monitoring Matters More in Arizona Now
- Rates are changing. Large-load tariffs at APS and SRP are being revised or applied in new ways. Accurate interval data helps operators model costs, verify bills and negotiate service agreements.
- Grid peaks are rising. APS and SRP set new peak demand records in July 2026. SRP told regulators in April 2026 that it serves 59 large load customers with nearly 7,000 megawatts of combined load.
- Public scrutiny is intense. Debate over data center energy use is now a statewide political issue. Verified data supports credible public reporting.
- Heat drives cooling energy. Summer afternoons push cooling loads, and therefore total facility demand, to their highest levels exactly when the grid is most strained.
Building a Metering Hierarchy
Good energy monitoring starts with meters at the right points. A well-designed hierarchy lets operators trace energy from the utility connection down to individual racks.
| Level | What Is Metered | What It Tells You |
|---|---|---|
| Utility service | Main incoming feeds | Total facility consumption and peak demand for billing |
| Switchgear and distribution | Feeders to IT, mechanical and other loads | How energy splits between computing and support systems |
| UPS input and output | Power entering and leaving uninterruptible power supplies | UPS efficiency and losses |
| PDU and busway | Power to rows and zones | Capacity use and balancing across the floor |
| Rack and outlet | Power to individual racks or devices | Customer billing, capacity planning and stranded capacity |
| Mechanical systems | Chillers, pumps, cooling towers, air handlers, CRAHs | Cooling efficiency and seasonal behavior |
| Backup systems | Generators, batteries, fuel and test runs | Readiness and contribution during grid events |
Measuring PUE Honestly
Power usage effectiveness compares total facility energy with energy delivered to IT equipment. A lower number means less overhead for cooling, power conversion and other support systems. But PUE can be measured at different points, such as UPS output or rack inputs, and over different time periods, so headline numbers are not always comparable.
For Arizona facilities, two practices make PUE more meaningful:
- Measure continuously, not annually. Track PUE by month and hour. Summer afternoon PUE often differs sharply from winter night PUE, and the summer number is what matters for peak demand and grid impact.
- Document the method. State exactly where IT energy is measured and how often. Consistency matters more than an impressive-looking figure.
Demand, Not Just Consumption
Large-load rates typically include demand charges, which depend on the highest power draw during billing periods, alongside energy charges. Reported demand charges on APS's Extra High Load Factor tariff have been in the range of roughly $13 to $19 per kilowatt per month, though actual charges depend on the tariff outcome and service agreements. Peak demand monitoring at short intervals helps operators understand which systems drive those peaks, typically cooling during hot afternoons, and whether controls, thermal storage or batteries could reduce them.
Cooling Energy in the Desert
Cooling is where Arizona's climate shows up most clearly in the energy data. Economizer hours, when outside air can help cool the facility, shrink in summer. Chillers and dry coolers work hardest when outdoor temperatures exceed 110 degrees. Detailed monitoring of compressor power, chilled water temperatures, condenser approach and fan energy reveals opportunities such as raising supply air temperatures within equipment limits, optimizing chiller sequencing and catching fouled coils before efficiency drops. Dust storms can quickly degrade coil and filter performance, so monitoring pressure drops and energy together helps schedule cleaning.
Facilities moving to zero-water or closed-loop cooling to address water concerns often trade water use for higher energy use in summer. Monitoring both together, rather than optimizing one in isolation, gives a truer picture of environmental performance.
Batteries, Generators and Grid Events
Backup systems are increasingly part of energy strategy, not just resilience. Tucson's new data center rules call for generators and batteries for backup power, and utilities are exploring how large customers can help during grid emergencies. Monitoring battery state of charge, temperature and health, along with generator readiness and fuel, lets operators know exactly how much flexibility they have when a utility requests load reduction. Batteries also need careful thermal management in Arizona heat, since high temperatures accelerate aging.
Workload Flexibility
Some computing tasks, such as batch processing and certain AI training jobs, can be scheduled for cooler hours or shifted between sites. Combining IT workload data with facility energy data makes it possible to see how much load could move away from summer afternoon peaks without affecting customers. This is an emerging practice, and contracts with tenants or customers must allow it, but it is one of the few tools that reduces peak demand without new infrastructure.
Stranded Capacity: The Hidden Cost
Many data centers reserve more power than they actually use. Racks are provisioned for peak loads that rarely occur, and safety margins stack up at each level of the power chain. Under large-load tariffs with minimum billing, such as SRP's approach for accounts of 20 megawatts or more, paying for reserved capacity that sits unused becomes a direct cost. Rack-level and PDU-level monitoring reveals where capacity is stranded, so operators can safely place more load within existing reservations or right-size future requests. In a constrained grid, using reserved capacity efficiently is both a financial and a public interest benefit.
Heat Waves as Stress Tests
Every Arizona summer provides a natural experiment. Comparing energy data from the hottest week with a mild week shows how much cooling energy rises, how close systems come to capacity limits and how PUE shifts. Operators can use these comparisons to plan upgrades, set realistic expectations with customers and prepare demand response strategies before the next summer. Recording what happened during each heat wave, including any operational changes, builds an institutional memory that improves planning year after year.
Reporting That Builds Trust
As energy use becomes a public issue in Arizona, operators may benefit from consistent reporting to utilities, regulators, customers and communities. Useful metrics include monthly energy use, peak demand by season, PUE by month, share of renewable or contracted clean energy, participation in demand response and actions taken during heat waves. Verified, consistent data is more persuasive than broad claims.
Tenant Billing and Accountability
Colocation operators increasingly pass energy costs to tenants based on metered use. Accurate rack-level and circuit-level data supports fair billing, encourages tenants to manage their own consumption and reduces disputes. As large-load rates rise, tenants will want more detail about how their charges are calculated. Clear, auditable metering becomes part of the customer experience.
A Practical Roadmap
- Audit existing meters. Identify gaps in the hierarchy, especially for mechanical systems and backup equipment.
- Standardize data collection. Bring meter data into one platform at consistent intervals, using open protocols such as Modbus, BACnet and SNMP.
- Model your tariff. Use interval data to understand how current and proposed large-load rates affect costs.
- Target summer peaks. Analyze what drives demand on the hottest afternoons and test strategies to reduce it.
- Integrate backup and workload data. Know your flexibility before a grid event.
- Report consistently. Publish or share standard metrics on a regular schedule.
Revisit the roadmap after every summer. Each season reveals new patterns in demand, cooling performance and backup readiness that should shape the following year's priorities.
Security for Energy Systems
Power monitoring and control systems are operational technology. Segment them from tenant and corporate networks, control vendor access and patch on a schedule. A compromised power management system could cause outages or hide problems. Our IoT security guide outlines a full program.
Looking Ahead
The APS rate case decision, possible legislative action in 2027 and continued growth in AI computing will keep energy at the center of Arizona's data center conversation. Facilities with detailed, trustworthy energy data will be best placed to manage costs, work with utilities and respond to public questions. Smaller edge facilities face the same need at smaller scale, as discussed in our article on edge computing and Arizona data centers.
Our guide to cutting peak demand covers demand charges for other commercial buildings. Find energy monitoring partners in the IoT Arizona company directory, including energy and sustainability providers in Goodyear. Browse more of our Energy and Sustainability coverage.
Frequently asked questions
Related articles
Energy & Sustainability
Energy & Sustainability
How Arizona Solar Projects Are Becoming Connected Energy Systems
Energy & Sustainability
IoT Building Energy Management in Arizona: Running Efficiency as a Continuous Program, Not a Rebate Project
Energy & Sustainability
