AI buildings cut energy use by up to 22%: Schneider

Schneider Electric has published research finding that AI-enabled buildings can cut whole-building energy use by up to 22% compared with traditional controls. The study examined building scenarios in Australia, India and the US.

The research focused on AI-driven heating, ventilation and air conditioning optimisation delivered through a smart building management system. It found this approach could increase building energy savings by 7.2% to 12.7% beyond conventional digital controls.

At current commercial rates, annual utility savings were estimated at USD $13,600 to USD $49,300 per building. In some scenarios, annual energy savings exceeded 200 MWh.

The carbon emissions avoided through the use of AI were more than 100 times greater than the footprint of the AI system itself, according to the research. It also estimated that buildings could avoid up to 59,869kg of CO2e each year, or about 60 metric tonnes.

Buildings account for about 37% of global energy-related carbon emissions, according to figures cited in the research. The findings come as scrutiny grows over the environmental cost of AI systems and their electricity use.

How it works

The study used building energy modelling validated against real-world pilot deployments, Schneider Electric said. It assessed the effect of AI-enabled HVAC control on energy consumption, carbon emissions and operating costs across different building types and operating conditions.

According to the research, the AI layer sits on top of digital building management systems and links data sources that are often held separately. It then continuously analyses building conditions and automates HVAC adjustments in real time.

Inputs included occupancy patterns, weather forecasts, equipment performance and other operational data. This process can help buildings run more efficiently while reducing the workload for facilities teams, Schneider Electric said.

Smaller buildings

The research also pointed to a potential benefit for small and mid-sized buildings of less than 100,000 square feet. Such sites have often struggled to adopt energy management systems because of cost, complexity and limited in-house expertise, Schneider Electric said.

By automating some of the optimisation work, AI could make those systems more accessible to smaller operators. That could widen adoption beyond large commercial properties with dedicated facilities staff and more established digital infrastructure.

The findings also compared AI-led optimisation with smart building controls used on their own. In that comparison, AI could more than double the energy savings achieved by smart controls alone.

The research covered both cloud and edge deployments of AI. It found each model could produce meaningful energy and carbon reductions, suggesting building operators may have more than one technical route to implement such systems.

Pankaj Sharma, EVP, Software & Services, Schneider Electric, said the shift in building management would depend on how organisations use operational data. "The future of building management will be defined by how effectively organizations connect and contextualize data that was previously trapped in silos. An AI layer on top of existing business systems can turn complexity into intelligence and intelligence into action, reducing emissions, lowering costs and improving performance simultaneously," Sharma said.

He added that rising energy demand would increase the need for systems that address several pressures at once. "As energy demand continues to rise, the ability to deliver these outcomes together, rather than forcing organizations to choose between them, will be critical to achieving both business and sustainability goals," Sharma said.

Sharma said the research suggests AI could broaden access to more advanced energy management in smaller properties as well as larger estates. "AI is putting the power of energy intelligence into the hands of smaller building owners and operators. What once required significant expertise and investment can now be achieved more simply and at greater scale, helping organizations reduce energy waste, lower costs, improve performance, and make smarter decisions with confidence," he said.