The rapid growth and application of artificial intelligence (AI) and Machine Learning (ML) is shaping the design and operation of data centres, says Ben Selier, Vice President: Secure Power, Anglophone Africa of Schneider Electric.

The training requirements associated with AI are driving new chip and server technologies and the need for extreme rack power densities.

The distinction between training and inference is critical when designing AI systems. Training workloads are used to train AI models like large language models (LLMs). These workloads require massive amounts of data fed to specialised servers with processors known as accelerators.

Inference models might be deployed on edge devices or cloud servers, depending on the application’s needs, and take the previously trained AI model into production to predict the output of new queries (inputs).

This demand for high-performance computing has led to increased requirements for powerful servers, GPUs (Graphics Processing Units) and other specialised hardware within data centres to support workloads.

The rise of edge computing, where computing resources are brought closer to the location where data is generated, is driven in part by AI applications.

Edge data centres are deployed to reduce latency and enhance the performance of AI applications in scenarios where real-time processing is critical like monitoring the movements of hospital patients to keep them safe.

The servers supporting these AI applications use advanced AI chipsets known as AI accelerators. These chipsets play a crucial role in enhancing the performance of AI applications across various domains.

Energy consumption and demand for power

Modern data centres use high-density servers and equipment that demand more power for processing. This leads to concentrated energy usage in smaller spaces, increasing the overall energy footprint, which could raise concerns about the amount of energy required to process AI operations.

As dependence on AI grows, new technologies and techniques will be deployed to make AI faster, more accurate and efficient.

One of these techniques is moving compressed models in the form of edge AI data centres (strategically located closer to the sources of data generation). This will enable businesses to match the application to the model and optimise performance and energy use.

The rapid growth in data traffic

The emergence of autonomous AI agents and decision-making programmes holds the potential to revolutionise various aspects of business operations. These intelligent programmes perform tasks independently, adapting and learning from its environment.

With AI-driven automation, data centre employees can automate routine tasks, reduce manual workloads and enhance overall efficiency. Responsibilities such as server system maintenance or system monitoring can be handled by these intelligence programmes.

AI and data centre evolution

Advancements bring efficiency and innovation, but also pose challenges related to energy consumption, and power and cooling systems. To meet these evolving needs, the data centre industry needs to adapt, embracing scalable and flexible infrastructure design to support intensive AI workloads.

Strategies could include deploying high-efficiency and high-capacity power systems and liquid cooling systems, 48U wide enclosures, upgrading hardware, and data centre infrastructure management.