Given AI’s appetite for energy there are growing sustainability concerns around its carbon footprint, but Lithuania’s Exergio believes it can also offer a solution to the problem
Artificial intelligence (AI) continues to develop rapidly, but there are concerns about its environmental footprint, and not without reason.
Data centres are notorious for their voracious appetites for energy to power the vast arrays of servers, and water to keep them cool. Inevitably, this has implications not only for climate change and the environment, but other sectors looking to make use of these resources – be they domestic or industrial.
Yet without data centres to host the powerful computers that make AI possible, the technology is ‘dead in the water’. It’s a conundrum that besets more than just the tech sector looking to make the most of its potential.
According to a recent WTW report, the growth of data centres harms the sustainability goals of technology firms such as Google, Amazon or Microsoft. These centres currently consume 6% of all electricity in the US and are expected to double by 2026.
However, the answer to the AI conundrum may lie within the technology itself. Well, that’s according to Exergio, a Lithuanian company that develops AI-based tools intended to optimise energy use in commercial buildings.
AI doesn’t only waste energy, the company argues, and AI-powered tools can already be used to do the opposite – and cut energy waste.
Donatas Karčiauskas, Exergio’s chief executive, said: “The common narrative suggests that AI’s growth contradicts the urgent need to reduce greenhouse emissions. However, this overlooks the fact that certain AI tools are already helping multiple industries cut energy waste. And not just more affordably, but also faster and more precisely than other existing traditional methods such as rule-based automation, manual energy audits, or reactive maintenance. Big tech has to simply utilise it.”
Optimal consumption
Karčiauskas suggests that the buildings sector can be used as an example. It is, after all, a major energy consumer; responsible for about 40% of global energy usage, and 33% of greenhouse gas emissions.
Inefficiencies in heating, ventilation, and air conditioning (HVAC) systems cause buildings to consume more energy.
Currently, however, existing AI solutions are able to rearrange the concept of how energy can be managed in properties. They analyse real-time data and then optimise HVAC operations, lighting, and other energy-intensive systems.
For example, AI can predict occupancy patterns, and adjust heating and cooling systems accordingly.
“It’s a common misconception that AI just wastes – this technology, in such cases, also ensures that energy is used only where and when it’s needed. Numbers-wise, it depends on the project, but it could reach up to 29%,” Karčiauskas added.
Smart grids and cities
The focus for this AI potential can be expanded from individual buildings to entire urban areas, Exergio believes, citing so-called smart grids as an example.
A smart grid is an electricity supply network that uses digital communications technology to detect and react to local changes in usage. Currently, similar technologies are being used in the development of smart grids and cities to optimise energy consumption.
Indeed, smart grids primarily integrate AI to upgrade energy distribution. A study published in the International Journal of Communications, Network and System Sciences showed that smart grid implementation manages to reduce electricity transmission losses by up to 20%.
After analysis of historical consumption patterns, AI systems can predict fluctuations in energy demand, and enable utilities to adjust power flows dynamically – a process similar to what happens in buildings. Further from grids, AI is also used in weather monitoring, which is then integrated into building systems.
“In the past, we’ve been used to reactive decisions. The industry is shifting towards learning, analysing, and predicting now,” Karčiauskas said.
“In some cities, weather analysis can inform building owners on how to better anticipate energy demand, mitigate extreme weather impacts, and optimise renewable energy use. As AI models become more sophisticated, we’ll see even greater precision in how we balance sustainability with energy efficiency.”
Integrating AI-driven weather prediction into building management systems is claimed to directly improve energy efficiency. This way, AI enables proactive adjustments to HVAC operations.
According to a Nature study, adopting AI technology at scale and combining it with existing energy policies and low-carbon power generation could lead to 90% lower carbon emissions.
Scaling automation
Scaling such technologies would be crucial to make an impact, though. Here, robotics are being used as a technology that could scale AI-based tools and apply them in different environments.
“Even though AI-powered robotics are mostly used for domestic tasks and mundane jobs that are dangerous for humans, in the future they will be trained to achieve sustainability by automating tasks, especially in buildings,” said Karčiauskas.
“Imagine if an HVAC sensor gets faulty – a robot gets notified, moved to a specific location, and gets to fix the problem.”
While AI-powered robotics have already transformed industries like manufacturing and logistics, their role in the building sector’s energy management is still in the early stages, but it is something that Exergio is beginning to explore, as Build in Digital has previously reported.
“We’ve started training a humanoid robot that will be based on the NVIDIA GR00T system,” said Karčiauskas. “This only feels natural. After years of collecting energy consumption data in the buildings industry, and learning how to utilise it, we see how AI-based technologies can help us take it to the next level.”
Read next: Hornsea 3 converter station sets sail from Thailand
Are you a building professional? Sign up for a FREE MEMBERSHIP to upload news stories, post job vacancies, and connect with colleagues on our secure social feed.
