Meeting the demands of AI necessitates densification across many layers of the data center—from chips to power, cooling to networks, and other essential equipment. Managing these elements in increasingly confined spaces presents numerous challenges, coupled with the need to maintain performance and drive efficiency. As the underlying infrastructure requires modernization to keep pace with AI workloads, what elements should we prioritize?
Data Center Dialogues
Highlight episodes from past DCD broadcasts. Listen to our best picks on the go
-
-
Episode 21 - Cooling the cutting edge of AI technologies
Effective thermal management and power efficiency are critical to ensuring the performance and longevity of cutting edge AI chips. In this session, we delve into the technical challenges, as well as the way around, for managing these demands.
-
Episode 20 - Data centers in Africa - What's driving investment into the region?
Africa’s data center market is growing at an unprecedented rate, expected to reach $5 billion by 2026. This is being driven by increased adoption of Internet, Cloud computing and digital services.
However, the market remains challenged by a number of factors, including power shortages and droughts. This episode explores the trajectory for growth for data centers in the region, navigating power availability challenges on the continent, and how macroeconomic factors such as exchange rates and interest rates will impact data center growth in Africa.
-
Episode 19 - How to efficiently cool AI
Liquid cooling is essential for high-density data centers to ensure superior thermal management and peak performance. Adapting infrastructure to accommodate these workloads is challenging, especially when introducing liquid into the facility. Efficient cooling solutions are crucial. This episode explores the technologies leveraged to cool AI effectively while allowing businesses to scale with demand.
-
Episode 18 - Keeping up with demand for digital infrastructure - An investor's perspective
The increased demand for data centers over the next few years have been well documented, driven by the digital economy, public appetite for data & generative AI. But how has this translated so far in 2024 from an investment perspective? This panel explores:
- How the demand for digital infrastructure has shifted investment in 2024 so far
- Opportunities from digital infrastructure for investors and how these can be capitalized on
- Risks and challenges posed by market dynamics and macroeconomic headwinds
- How investors can maximize opportunities whilst maintaining diversified portfolios
-
Episode 17 - Navigating energy management in data center operations
This episode delves into the intricate balance between energy management, alternative energy sources including nuclear solutions, and operational efficiency and management within data centers. With antiquated global grids struggling to meet skyrocketing energy demands, alternative energy strategies become essential. High-density AI servers now require 40 KW or more per rack, compared to the previous 10-12 KW for conventional computing, necessitating innovative approaches to supply, energy efficiency, and cost reduction.
-
Episode 16 - Backup power & emissions regulations - Stacking up the industry response
The need for strong backup power and resiliency in the data center sector has never been greater, driven by outdated grids, power shortage & more extreme weather events. At the same time, the regulatory landscape in North America is becoming increasingly complicated as regulators crack down on emissions reporting. This episode will explore:
- How data centers should be responding to meet regulations, whilst safeguarding business continuity for customers
- Planning for back-up power & gensets alongside the ‘unknown unknowns’ of the future regulatory landscape
- How to ‘stack up’ the complications of emissions regulation in an increasingly regulated environment
-
Episode 15 - Bridging the hardware and software gap to optimize HPC
Whilst HPC and AI have become core to modern data center strategy, do we really understand what’s happening when it comes to the workloads? Despite huge technological headway having been made when it comes to computing, a rift still exists between the hardware and software layers. The question remains: how do we bridge this gap to optimize our IT hardware and workloads, to bring efficiencies throughout the tech stack?
-
Episode 14 - How innovative approaches to server architecture are enabling HPC
As we address the needs of modern computing, data centers must densify server environments and drive connectivity to support the processing and transfer of data at accelerating speeds. In this episode expert speakers will provide a blueprint for developing a robust server infrastructure, capable of scaling in line with HPC demands. In particular they address:
- The evolution of server network topologies in the era of AI
- Innovations in transciever technologies
- How modular servers and networking components drive efficiency and rapid expansion
- The role of chiplet technologies in supporting fibre optics
-
Episode 13 - Overcoming the complexities of HPC deployments
High performance computing is solving challenges across numerous sectors, yet deploying and operating advanced technologies presents multiple hurdles for the data center industry. These problems require strategic, practical and innovative approaches to ensure the seamless integration of next generation hardware and workloads, but what is the reality on the ground when it comes to integrating and operating advanced compute? This episode explores:
- The reality of deploying HPC into retrofit versus newbuild environments
- How are HPC and AI being used to drive sustainability?
- The rising demand for power and advanced cooling solutions
- Scalable and modular solutions for high performance computing
-
Episode 12 - Develop a blueprint for your distributed computing strategy
Surging demand for AI, OT and IoT devices is reshaping the technological landscape, putting pressure on data centers to create robust, scalable and efficient distributed computing strategies. In order to guide operators towards developing a resilient and adaptive Edge computing roadmap, expert speakers discuss the core fundamentals required to facilitate its rollout.
-
Episode 11 - The real realities of leveraging AI
AI is already playing a pivotal role in shaping more secure, efficient, and resilient networks. But with this growth, how are industry leaders ensuring we have the infrastructure to scale in line with demand? With so much speculation about the size and scale of impact, hear from experts about the reality on the frontlines of the AI revolution.
-
Episode 10 - Cooling the future: Preparing your data center for the AI surge with direct liquid cooling
As liquid cooling takes the spotlight over air cooled systems, it is important to determine how facilities can best support this transition.
This episode discusses why single-phase direct liquid cooling (DLC) is the best choice to manage heat for AI systems and data centers in general, and learn how to leverage single-phase DLC’s compatibility and scalability when transitioning your data center from air to liquid cooling.
-
Episode 9 - The APAC data center market - Where are the next hotspots in this burgeoning region?
Over the last few years, increased investment in new and exciting markets in the Asia Pacific region has led to many countries, such as Indonesia and Malaysia, being seen as potential new hotspots for the industry.
However, with so much untapped potential yet to be explored in the region, what does the current market tell us about where we could be heading?
-
Episode 8 - Liquid cooling implementation in new and existing data centers
The increasing demands on high-performance computing continues to drive the need for efficient cooling systems. This episode provides an in-depth overview of innovative liquid cooling solutions designed to meet evolving data center requirements.
Listen to this episode to hear how to reduce operational costs, maximize redundancy and increase return on investment by handling heat loads of up to 800kW and beyond. -
Episode 7 - Advancing clean-powered microgrids: Overcoming barriers with advanced nuclear technology
Microgrids powered by advanced nuclear technology offer a promising solution to meet the growing demand for reliable, clean energy. But this small-scale approach to nuclear power can fulfill its potential only if it can overcome three key challenges: cost, build time, and public perception. This episode dives deep into these challenges and provides an overview of the financial, operational, and environmental ramifications of nuclear-powered microgrids.
-
Episode 6 - Case study: How data center organizations can secure green investment
In this episode, DCD hosts Mark Bjorsgaard from Deep Green and Bilal Ahmad from Octopus Energy Generation to discuss OEGen's recent investment in Deep Green and the potential of sustainable financing for data centers.
-
Episode 5 - The ASHRAE update: Setting the record straight on the thermal ranges
While ASHRAE's thermal guidelines have been around for some time, due to their complexity and the nuances of different facilities, they have been difficult to apply. During the first episode of this two-part special, ASHRAE joins the DCD studio to cut through the noise and provide clarity to the thermal ranges, a critical factor to get right in the era of high density computing.
-
Episode 4 - Powering the low carbon transition
With AI and public demand for data expected to skyrocket demand for data centers, and the power and water situation in many parts of the world precarious, data centers continue to draw scrutiny from both outside the industry and within.
In this episode, DCD will be joined by NTT Global Data Center's Senior Director for Power & Energy, Neal Kalita, to discuss some of the most pressing sustainability issues within digital infrastructure. They will discuss:
- How the sector needs to be powered differently to meet carbon reduction targets, and global challenges with PPAs
- Why data centers need to change the way they think about sustainability, and why we're still measuring environmental efficiency wrong
- What the good news is when it comes to technology & sustainability innovation, and how these successes need to be built upon
-
Episode 3 - A generational growth opportunity in data centers – what are investors really looking for?
With the huge expansion in data center investment and merger & acquisition activity over the past year, investors are becoming increasingly conscious of the criteria for what digital infrastructure they are financing, with workloads shifting substantially and criteria such as ESG financing becoming increasingly important. In this episode, DCD will be joined by Morgan Laughlin, Managing Director & Global Head of Data Centers Investments at PGIM Real Estate to discuss:
- Data center outlook - what's driving growth and what are the opportunities?
- The evolution of data centers from a niche to a mainstream investment
- How green investment is evolving in light of the acceleration towards improved data center efficiency & sustainability
-
Episode 2 - Future-proofing the data center - how changing workloads are innovating design in APAC
In this episode Paul Christensen from Macquarie Data Centres, Jason Popiolek from Black & White Engineering, and Matt Gurr from AECOM discuss how data centers should be adapting structural design & engineering to accommodate AI.
They also explore what progress is being made with regards towards moving to a liquid cooling design, and if it is ever possible to retrofit.
-
Episode 1 - How to prioritize sustainability and reliability
In this episode, Adam Wray-Summerson from Clarke Energy, Benjamin Rapp from KOHLER Power and Colin Kelly from Gyrogy analyze the challenges in prioritizing sustainability and reliability in energy management and discuss how to solve them.
They also delve into the best practices and skills needed to safeguard operations, discuss how to boost the potential of green power and to navigate the complexities of energy efficiency and dependability.