In 2024, U.S. commercial buildings spent over $241 billion on energy. Heating and cooling account for 21% of commercial energy use, over $50 billion. While low occupancy rates and hybrid work lead to waste in unused commercial buildings, the problem is compounded by a significant inefficiency problem: information technology (IT) systems that track digital access and occupancy don’t communicate with operational technology (OT) systems that manage HVAC, lighting, and other equipment.
This IT-OT divide has existed for decades. IT systems manage data, networks, and enterprise applications. OT systems handle the physical side, heating and cooling systems, elevators, security, and life-safety equipment. These two worlds operate on different protocols and answer to different stakeholders, yet the data needed to make autonomous decisions to reduce costs, save energy, and optimize building efficiency exists. Until now, IT and OT have lived on opposite ends of the spectrum, but with Operational AI, they are combined into a single view.
Mind-Numbing Numbers
About one-third of the energy used in U.S. residential and commercial buildings is wasted each year, costing roughly $150 billion annually, according to the U.S. Department of Energy.
Globally, the scale of inefficiency is far larger. While no single statistic captures worldwide electricity waste in unoccupied commercial buildings, analysis of the U.S. Energy Information Administration (EIA) and the European Council for an Energy Efficient Economy (ECEEE) indicates that commercial buildings consumed an estimated 1.35 trillion kilowatt-hours of electricity unnecessarily during nights and weekends in 2025.
That amount is roughly half of the European Union’s annual electricity generation and enough to power about 129 million average U.S. homes for a year. Assuming a conservative global average business electricity price of 10 cents per kilowatt-hour, wasted electricity in commercial buildings globally amounts to an estimated $135 billion annually.
Pandemic Exposure
The COVID-19 pandemic exposed the problem worldwide. One might expect that as the pandemic emptied office buildings, energy use would plummet. Not so. Multiple studies across Europe, Asia, and North America found that energy use in commercial buildings dropped by only 10% to 30%, despite occupancy reductions of 70% to 90%.
Electricity usage in the Empire State Building fell by less than one-third during the pandemic, despite it being virtually unoccupied. UK offices experienced a 10% decline in footfall, while electricity costs remained near 80% of baseline levels. ECEEE studies uncovered similar patterns in Japan and China, with HVAC and miscellaneous electrical loads continuing to operate at near-normal levels.
Airport terminals are some of the most energy-hungry public buildings. Even when terminals sat empty for months, airports kept paying almost the same energy bills. Various studies indicate that terminals use two to three times as much energy per square foot as standard commercial buildings. Heating, ventilation, and air conditioning systems alone account for 40% to 60% of that consumption. During the pandemic, passenger traffic plunged to just 5% to 10% of normal levels, and energy systems, designed for peak occupancy, lacked integrated data to scale back.
Operational AI For the Built World Offers the Solution
Operational AI is finally connecting the dots between IT and OT by leveraging digital twins that tap into millions of real-time telemetry points to provide unmatched insights into how to operate in the built world. These virtual models of physical buildings pull together data from sensors, IoT devices, building management systems, and occupancy trackers. Operational AI gives facilities managers deep visibility into their facilities, helping them identify and mitigate issues, predict when new ones will arise, and optimize systems for greater sustainability. Building operators can see everything in one place, augmented with insights previously unavailable, and act as necessary.
Imagine a typical office building on a Friday afternoon. Data from IT systems shows that only 30% of employees are on-site. Room booking systems confirm that most meeting rooms are empty. Yet physical and operational systems behave as if the building were full. An integrated digital twin allows the building to respond to real conditions rather than static schedules.
Organizations adopting operational AI, from universities to hospitals to retailers, are already seeing millions of dollars in annual savings and meaningful drops in energy use. Fortune 500 companies are saving millions by integrating scattered building data into dynamic digital twins.
The economic case becomes even stronger given today’s commercial real estate landscape. U.S. office vacancy rates hit a record 19.6% in late 2023, the highest since tracking began in 1979. In some markets, such as San Francisco, vacancy has reached 34%. Low-occupancy buildings continue to consume substantial energy.
Operational AI also enables predictive maintenance. It shifts building management from reactive to proactive. A circulation pump showing unusual vibration patterns might indicate a problem. Addressing it now costs a fraction of what an emergency replacement would and avoids the subsequent damage that often accompanies sudden pump failures.
Operational AI addresses the issue of knowledge capture and transfer. The facilities management industry faces an acute labor shortage, driven by an aging workforce (the average age of facilities managers is 50, and 20% of the workforce is over 65) and the retirement of experienced facilities technicians. Institutional knowledge is being rapidly lost. Digital twin platforms capture this knowledge and make it accessible to new technicians. Predictive maintenance algorithms identify potential failures before they occur, guided by patterns that experienced workers once detected. The frontline facilities workforce is finally getting the digital tools that transformed office work two decades ago.
Problem Solved
The problem isn’t empty buildings; it’s operating them as if they were full. Lack of IT-OT integration and systems observability keeps energy usage high even when occupancy drops to zero, costing billions of dollars annually worldwide. Buildings are more intelligent than their systems let them be. Operational AI finally provides the visibility and autonomous control to run them that way.
