Data Centres in the Era of AI with Jay Park
Analyse Asia with Bernard Leong
Deep Dive
Why is the Asia-Pacific (APAC) region experiencing massive growth in data center investments?
The APAC region is experiencing massive growth in data center investments due to the increasing demand for AI servers and data processing. According to a recent structural research report, the data center industry will spend $100 billion, with 50% of that growth occurring in APAC. AI servers require data centers to be built closer to users to reduce latency and improve efficiency, driving the need for localized infrastructure.
What challenges do AI servers pose for data center infrastructure?
AI servers pose significant challenges for data center infrastructure due to their high power density and cooling requirements. Traditional air cooling systems are insufficient, necessitating the adoption of liquid cooling technologies. Additionally, the power draw from AI servers can surge unpredictably, requiring data centers to manage peak power capacity efficiently.
How is Digital Edge addressing the environmental impact of data centers?
Digital Edge is addressing the environmental impact of data centers by focusing on reducing water usage and improving energy efficiency. They have implemented innovative cooling technologies like the State Point Liquid Cooling (SPLC) system, which has achieved a Power Usage Effectiveness (PUE) of below 1.2 in hot and humid regions like Manila. Additionally, they are developing hybrid supercapacitors to replace traditional lithium-ion batteries, reducing fire risks and improving energy storage efficiency.
What is the significance of NVIDIA's AI chips in data center engineering?
NVIDIA's AI chips are highly disruptive in data center engineering due to their unprecedented power density, jumping from 10 kW per cabinet to 130 kW. This has forced the industry to rethink cooling solutions, with a shift towards liquid cooling systems. The varying operating temperatures required by different server manufacturers also add complexity to data center design and management.
How does modular data center design compare to traditional construction?
Modular data center design offers faster construction times and reduced material waste, as components are built in factories. However, it lacks flexibility for future upgrades due to transport size limitations and fixed designs. Traditional construction allows for larger, more adaptable buildings, which is crucial as power densities continue to rise.
What role does water usage play in data center sustainability?
Water usage is a critical factor in data center sustainability, particularly in cooling systems and power generation. Digital Edge's SPLC technology reduces water consumption by up to 40%. Additionally, lowering the PUE by just 0.1 in a 100 MW data center can save the equivalent of 1,500 Olympic-sized swimming pools of water annually, highlighting the importance of efficient cooling systems.
What are the key principles for designing efficient data centers?
Key principles for designing efficient data centers include minimizing energy transformation steps, such as reducing AC to DC conversions in UPS systems and avoiding unnecessary energy conversions in cooling systems. Understanding the internal components of data centers, like AI servers, is also crucial for designing infrastructure that can handle increasing power densities and cooling demands.
How is Digital Edge innovating in power management for data centers?
Digital Edge is innovating in power management by developing hybrid supercapacitors that replace traditional lithium-ion batteries, eliminating fire risks and improving energy storage efficiency. They are also working on power shaving systems to manage peak power demands more effectively, reducing the need for customers to purchase excess capacity.
- APAC will account for 50% of the $100 billion data center investment
- AI servers are driving data center growth in APAC
- Data centers are being built closer to users to reduce latency
Shownotes Transcript