Predictive modelling drive sustainability wins 

It is undoubtedly true that AI promises great potential in the journey towards sustainability. The International Energy Association (IEA) estimates that building energy consumption can be cut by up to 20% with predictive monitoring. AI’s exemplary efficiency is particularly noticeable in large-scale smart building systems. At 45 Broadway in New York City, the AI-powered smart HVAC system produced a 15.8% energy saving, cutting around 37 tonnes of CO2. Numbers from the logistics industry tell a similar story, forecasting AI-driven emission reductions of about 30% globally. 

These demonstrable benefits of AI strengthen its momentum in the sustainability sector, projected to be valued at over $100 billion by 2035. 

Hidden footprints 

Yet the environmental impact of AI – at least in its current iteration – is not solely positive. Large language models (LLMs) such as the likes of Gemini and ChatGPT are particularly resource intensive in water and energy. A report by Goldman Sachs forecasts that global power demand from data centres will increase by 165% from 2023 to the end of the decade. 

What exacerbates the situation is the distribution of these environmental impacts. At a time when water supply in the global south faces severe challenges, many data centres are opting for these water-stressed regions due to lower energy costs and the availability of land. In parts of mid-west America, South America, and south-east Asia, precious water sources are being extracted, and sometimes even diverted from local communities by new data centre projects. 

In the keynote presentation at the Smart Workplace Summit at ISE 2025, consultant Claudia Finer spoke eloquently about AII’s sustainability dilemma. Here’s a one-minute clip: 

You can watch the whole video here. 

Industry response 

The IT industry is certainly not blind to the environmental and human cost of AI. In the past year, AI providers have been scrambling to improve overall efficiency. Google Gemini AI now cites 0.24Wh per text prompt, a 33-fold reduction on last year. However, experts caution that in an age of skyrocketing demands, such efficiency gains may be seen as misunderstood or overstated. The IMF still believes that in 2025-30, AI could generate CO2 equivalent to the entire emissions of Italy over the same period. 

Looking ahead 

AI’s contribution to sustainable development cannot be dismissed, but neither can its cost. Striking a balance is therefore the main challenge for organisations that are looking to, or in the process of, adopting AI into their daily workflow: while AI can be used to deliver optimisation, its use must be accompanied with transparency and scrutiny to its long-term impacts. Without safeguards and a continued technological push towards sustainability, AI risks perpetuating the very inequalities and environmental strains it is being used to tackle. 

Stay informed!

AI and sustainability are two of the megatrends that are revolutionising the world of AV and systems integration.

ISE is the world-renowned annual tech show for the AV and systems integration industry, taking place in Barcelona, 3-6 February 2026. For more updates on AI, and to discover more about ISE 2026 as details are released, sign up for updates.

Don't miss out – join our community today and keep up to speed with all the latest industry trends.