Ctg sit23 hub banner.jpg

Datacentre operators grapple with surging capacity demand

Global capacity projected to double to 21ZB by 2027

Datacentre operators grapple with capacity planning challenges amid surging AI demands

Image:
Datacentre operators grapple with capacity planning challenges amid surging AI demands

Report by real estate consultancy JLL sheds light on the mounting pressures facing datacentre operators worldwide as the demand for AI workloads surges

The ongoing evolution (or possibly revolution) of generative AI, is presenting datacentre operators with significant capacity planning challenges. The extent of that challenge is necessitating a reevaluation of infrastructure strategy.

According to a JLL report, global datacentre storage capacities are projected to skyrocket from 10.1 ZB in 2023 to a staggering 21 ZB by 2027. This exponential growth in AI-driven workloads is poised to reshape the landscape of datacentre operations, forcing operators to rethink their approach to building and managing facilities.

Jonathan Kinsey, EMEA lead and global chair for datacentre solutions at JLL, underscored the critical role of strategic site selection in ensuring operational scalability and meeting environmental goals amidst mounting power challenges and the imperative for sustainable energy.

Kinsey highlighted the need for collaboration between real estate professionals and partners to secure adequate future power, especially as existing grid infrastructure may struggle to support the global shift to electrification and digital expansion. This is certainly the case in UK where outdated grid infrastructure is already acting as a bottleneck in the development of a lower carbon economy.

The report underscores the dramatic transformation in datacentre requirements over the past decade, noting that while new datacentre builds once boasted a critical IT load capacity of less than 10MW, recent developments now see developers announcing builds exceeding 100MW. Moreover, the emergence of generative AI poses unique challenges, as AI-specialised datacentres necessitate tailored planning, design, and power allocation strategies.

With AI workloads being inherently energy-intensive, achieving a balance between energy needs and sustainability strategies has become paramount. Datacentre operators are under increasing scrutiny as awareness of the environmental impact of computing grows. Measures such as switching from traditional air-based cooling to liquid-based cooling are already becoming widespread practice.

The report emphasises the imperative for datacentre operators to enhance efficiency and collaborate with local governments to secure sustainable energy sources.

It also highlights the need for substantial investments in electricity grids. In Europe, one-third of the grid infrastructure is over 40 years old, and an estimated €584 billion investment by 2030 is required to meet net zero goals. This level of investment is proving increasingly contentious for governments across Europe as net zero measures are increasingly exploited by right wing populist parties in multiple member states.

Similarly, in the United States, achieving energy transition objectives will demand an estimated $2 trillion investment.

According to JLL report, specialised data centres tailored for training and tuning AI models will emerge over time, offering enhanced efficiency and performance. In the interim, significant opportunities exist to redesign existing datacentres to accommodate the heightened power and cooling demands associated with AI.

"Those which can meet these demands while minimising environmental footprints will stand ready to benefit from meeting the needs of our data-intense digital economy," the report says.

Catriona Shearer, EMEA Data Center Consulting Lead at JLL, remarked, "Managing aging infrastructure with increased demand intensity, alongside new ESG requirements and legislation, provide data centers with a uniquely challenging environment in 2024."

"However, each is an opportunity to drive meaningful change using a combination of data center expertise, platform tools and truly creative innovation - we're set for an extremely exciting year."

You may also like

Meta's 'pay or consent' model likely to be on borrowed time in EU
/news/4339147/metas-pay-consent-model-borrowed-eu

Artificial Intelligence

Is this why release of newest LLM is confined to the US?

Health and safety gone rad: Fogsphere brings AI chops to workplace monitoring
/feature/4338679/heath-safety-gone-rad-fogsphere-brings-ai-chops-workplace-monitoring

Artificial Intelligence

Health and safety gone rad: Fogsphere brings AI chops to workplace monitoring

London-based startup Fogsphere uses standard 4K security cameras - the type that can be purchased for £70 - to monitor workplaces and create immediate alerts as soon as dangerous conditions are spotted.

Tata's UK gigafactory project takes major step forward
/news/4338523/tatas-uk-gigafactory-project-takes-major-step-forward

Components

Tata's UK gigafactory project takes major step forward

Sir Robert McAlpine to build multi-billion-pound factory