Covering Disruptive Technology Powering Business in The Digital Age

Written By: Jon Curry, Vice President, Operations, Asia Pacific, Digital Realty

Jon Curry, Vice President, Operations, Asia Pacific, Digital Realty

Amid all the discussion about the rapid uptake of Artificial Intelligence (AI) services, we are omitting a critical question: How do we align the technology’s growth with Asia Pacific businesses’ sustainability and efficiency requirements? It’s akin to asking, “How can we have ‘Green AI’?”

With adoption set to surge, the energy needed to train and run large AI models is likely to have an impact on the environment footprint of businesses in this region. In fact, AI is estimated to consume up to 3.5% of the world’s electricity by 2030.

A key challenge for businesses over the next decade will be balancing the complexity and intensity of AI workloads with the need for environmental sustainability.

The Fundamental Role of Optimised Infrastructure

One thing is certain: AI isn’t going anywhere.

IDC forecasts that Asia Pacific spending on AI, including software, services, and hardware for AI-centric systems, will grow to $78.4 billion in 2027. Much of this growth stems from enterprises’ hunger for High-Performance Computing (HPC), as they develop innovative applications and services for internal use and customer-related needs.

Yet this growth also sparks environmental concerns, particularly relating to increased energy demands for advanced cooling systems in data centres to accommodate higher-density deployments.

Data centres are fundamental to AI adoption, providing the underpinning digital infrastructure for a range of innovative applications across numerous industries.

AI has not caught the data centre industry by surprise. In fact, dialogue around the growing demand for AI and data analytics, and the associated environmental implications, has been ongoing for nearly a decade.

Innovations have since been introduced to address the sustainability impact of the technology. For example, advanced cooling systems, like Direct Liquid Cooling, which circulates a liquid coolant directly to heat-generating components in servers to efficiently dissipate heat, offer higher cooling efficiency.

Coupled with improved infrastructure design to support high power densities, data centres can now provide High-Density Colocation, the optimisation of space for computing equipment across data centres, to reduce overall resources and carbon footprint while handling the demand of high-intensity AI workloads.

KakaoBank, a leading mobile bank in Korea, established an AI lab at Digital Realty’s ICN10 data centre in Seoul using this approach. The bank is leveraging next-generation data centre infrastructure, including optimised cooling, layout, and connectivity options to accelerate AI-powered service enhancements, develop personalised content, and fuel research and development for new financial services.

Modularity Is Key to Efficient Design

Modular design is another key component of efficient AI infrastructure, which allows racks and servers to be removed or added without interrupting operations.

As AI applications diversify and increase in complexity, modularity enables efficient infrastructure scaling, eliminating the need for costly and time-consuming overhauls. This ensures data centres can readily adapt to escalating AI needs while maintaining peak performance without breaking the bank or experiencing extended downtime.

“One thing is certain: AI isn’t going anywhere.”

The benefits of modularity also extend beyond scaling. Rapid time-to-market is crucial for AI advancements. Modular designs, with pre-manufactured components, expedite deployment, facilitating rapid AI technology adoption and bolstering market competitiveness. Cooling in modular data centres is also naturally more efficient.

For this reason, their design makes it easier to incorporate advanced cooling technologies. This allows for more precise control and adaptive temperature management based on real-time IT equipment heat loads.

Construction plays a part in this equation, too, including designing buildings free from reliance on water and fossil fuels. Alongside energy reuse systems and district heating networks, which redistribute excess heat to nearby communities, these initiatives ensure that infrastructure not only meets businesses’ technological needs but also benefits the communities they are located in.

Creating More Efficient Infrastructure with Green AI

Alongside discussions about the energy efficiency of green AI deployments, we should not forget AI’s potential to create a positive sustainability impact.

Green AI can enhance efficiency and reduce environmental impact within data centres. For example, data centres can use green AI to process facility-level data to identify anomalies and suggest optimisations to operational settings to support data-driven decision-making.

“A key challenge for businesses over the next decade will be balancing the complexity and intensity of AI workloads with the need for environmental sustainability.”

Software company Ekkosense uses cutting-edge AI, immersive 3D visualisations, and thermal optimisation analytics at its Cloud House data centre in London to pinpoint and rectify airflow and cooling inefficiencies. This has delivered savings and greater efficiencies, with a 20% reduction in cooling energy consumption within a year.

These diverse applications may only be the tip of the iceberg in the role green AI can play in driving the sustainability of IT infrastructure over the next decade.

Fundamental Shift in Digital Architecture Needed

While green AI offers environmental benefits, its computational demands create sustainability challenges. Data centres, constantly powering and cooling AI hardware, require a rethink beyond mere technology upgrades. A fundamental shift in digital architecture is needed.

Prioritising efficient, sustainable, and modular infrastructure will be key to enabling a future in which technological innovation and environmental sustainability can coexist harmoniously.

(0)(0)