Powering AI innovation while preserving sustainability in data centres

By Jad Jebara, CEO of Hyperview.

The rise of artificial intelligence is transforming how we work and live. From medical diagnosis to financial forecasting, AI is bringing automation, efficiency, and advanced analytics to a vast array of industries. However, this progress comes at a serious cost – the immense energy consumption required to power these advanced processes. According to a recent study, generative AI may require up to 33x more energy than task-specific software. 

Across industries we can expect to see this energy consumption increase as the UK AI market increases. Indeed, according to the International Trade Administration, the UK AI market is worth more than £16.8 billion and is expected to grow to £801.6 billion by 2035. It's imperative data centre operators begin to identify and implement key strategies to support navigating the growing AI boom.  

By strategically addressing these factors, data centres can facilitate AI innovation while also ensuring sustainable and efficient operations. The first step operators should take is identifying where data centre energy is used the most – but how? 

Data driven insights for targeted efficiency gains 

Traditionally, tackling energy waste in data centres has been a guessing game. However, modern data centre management innovations that have emerged in recent years offer a more precise approach. These platforms continuously monitor resource utilisation, power consumption, and even environmental conditions within the facility. This real-time data empowers operators to identify inefficiencies, such as overtaxed cooling systems.  

With this knowledge in hand, operators can strategically target the necessary areas for improvement - leading to significant energy savings while driving sustainability efforts. Another beneficial use case includes analysing data so operators can identify hotspots within the data centre and optimise cooling systems to focus on areas that require it. Similarly, detailed power usage effectiveness (PUE) reports can highlight inefficient servers that can be replaced with more energy-efficient models.  

By acting on accurate, real-time data analytics, data centre operators can make targeted decisions – and investments - that maximise energy efficiency while facilitating the increased power demands of AI. 

Optimising resource utilisation 

The power required to train and run AI models can be substantial. Traditional data centres often operate with underutilised resources, using energy unnecessarily. To achieve optimal resource utilisation, data centres need to develop and implement a comprehensive strategy based on data. This includes capacity planning, real-time monitoring, asset management, environmental control, and real-time data analytics. 

By deploying these strategies, data centres can significantly reduce idle processing power, translating into lower energy consumption. This optimisation not only benefits the environment, but also frees up resources for more demanding AI workloads, boosting efficiency and cost effectiveness. 

Embracing sustainable power sources  

When addressing long-term sustainability efforts, data centres need to think ahead to making the switch to renewable energy sources. This means the data centre will need to work with companies that generate electricity from renewable sources, or invest in on-site renewable energy generation technologies, like solar panels. These approaches can significantly reduce reliance on energy derived from fossil fuels and minimise the environmental impact of increased AI operations. 

Furthermore, exploring alternative cooling solutions is another important step data centre operators can take. Traditional air conditioning systems are energy intensive. Implementing liquid cooling technologies, which transfer heat more efficiently, can lead to substantial energy savings. Additionally, utilising innovative cooling techniques, such as outside air economiser for free cooling during colder periods, can further decrease reliance on energy-hungry cooling systems.  

A sustainable future with AI 

The rise of artificial intelligence is undeniable, and data centres will undoubtedly play a critical role in its advancement. However, data centres must ensure this progress doesn't come at the cost of environmental damage.  By leveraging data insights, embracing renewable energy sources, and implementing innovative cooling solutions, data centres can push help push the AI revolution to the next level, while remaining sustainable. 

The time to act is now. Data centre operators must prioritise energy efficiency as they plan for the future by investing in data centre management platforms, embracing continuous monitoring, and collaborating with technology partners who offer sustainable solutions. By taking these steps, operators can future-proof their data centres - contributing to a greener tomorrow while ensuring the power of AI is harnessed responsibly. 

By David de Santiago, Group AI & Digital Services Director at OCS.
By Krishna Sai, Senior VP of Technology and Engineering.
By Danny Lopez, CEO of Glasswall.
By Oz Olivo, VP, Product Management at Inrupt.
By Jason Beckett, Head of Technical Sales, Hitachi Vantara.
By Thomas Kiessling, CTO Siemens Smart Infrastructure & Gerhard Kress, SVP Xcelerator Portfolio...
By Dael Williamson, Chief Technology Officer EMEA at Databricks.