The insatiable hunger of artificial intelligence for data is pushing data centers to their limits. As tech behemoths scramble to expand their infrastructure to accommodate these computationally demanding AI workloads, they’re facing a critical dilemma: how to power these operations sustainably and affordably? This challenge has even spurred companies like Oracle and Microsoft to explore unconventional energy sources, including nuclear power.
Another pressing issue is the heat generated by the powerful hardware driving AI. Liquid cooling has emerged as a cutting-edge solution to keep these systems running smoothly while minimizing energy consumption.
In October 2024 alone, a wave of announcements from major tech firms highlighted the industry’s rapid shift towards liquid cooling solutions.
The Rise of Liquid-Cooled SuperClusters
At its Lenovo Tech World event, the company unveiled its next-generation Neptune liquid cooling system for servers. This sixth-generation Neptune, utilizing an innovative open-loop, direct warm-water cooling system, is being rolled out across Lenovo’s partner ecosystem. It promises to empower organizations to build and deploy accelerated computing for generative AI while slashing data center power consumption by up to 40%.
Giga Computing, a subsidiary of Gigabyte, took center stage at OCP Global Summit 2024, showcasing a direct liquid cooling (DLC) server specifically designed for Nvidia HGX H200 systems. Recognizing that not all data centers are ready for full-fledged liquid cooling, Giga also presented the G593-SD1, featuring a dedicated air cooling chamber for the Nvidia H200 Tensor Core GPU.
Dell’s newly launched Integrated Rack 7000 (IR7000) is a scalable system designed with liquid cooling in mind. This powerhouse is capable of managing future deployments reaching up to 480KW, capturing nearly 100% of the generated heat.
“Today’s data centers are struggling to keep pace with AI’s relentless demands, requiring high-density compute and liquid cooling innovations with modular, flexible, and efficient designs,” said Arthur Lewis, president of Dell’s Infrastructure Solutions Group. “These new systems deliver the performance needed for organizations to stay ahead of the curve in the rapidly evolving AI landscape.”
Supermicro has also entered the arena with its liquid-cooled SuperClusters designed to handle the heavy lifting of AI workloads, powered by the Nvidia Blackwell platform. Supermicro’s liquid-cooling solutions, supported by the Nvidia GB200 NVL72 platform for exascale computing, have begun sampling to select customers, with full-scale production slated for late Q4.
“We’re pioneering the future of sustainable AI computing, and our liquid-cooled AI solutions are rapidly being adopted by some of the world’s most ambitious AI infrastructure projects with over 2,000 liquid-cooled racks shipped since June 2024,” said Charles Liang, president and CEO of Supermicro.
The liquid-cooled SuperClusters feature advanced in-rack or in-row coolant distribution units (CDUs) and custom cold plates for housing two Nvidia GB200 Grace Blackwell Superchips in a compact 1U form factor.
It’s crystal clear that liquid cooling will be at the heart of data center operations as workloads continue to grow. This technology will be crucial for managing the heat and energy demands of the next generation of AI computing. We’re just starting to scratch the surface of its potential impact on efficiency, scalability, and sustainability in the years to come.
More from TechRadar Pro
Interview: The Future of Data Centers in the Age of AI
Editor: Welcome to Time.news, where we discuss the trends shaping our world today. Joining us is Arthur Lewis, the President of Dell’s Infrastructure Solutions Group. With the explosive growth of artificial intelligence, data centers are experiencing unprecedented challenges. Arthur, thank you for joining us.
Arthur Lewis: Thank you for having me! I’m excited to dive into this important topic.
Editor: Let’s start with the scale of the problem. You’ve mentioned that data centers are struggling to keep pace with the AI demands. Can you elaborate on what those demands are?
Arthur Lewis: Absolutely. The appetite for data that AI has is insatiable. The compute power required for training AI models is increasing exponentially. This leads to higher energy consumption and heat generation in data centers. We need solutions that not only keep up with these demands but do so sustainably and affordably.
Editor: You mentioned sustainability. It’s a critical issue today. What are some measures tech companies are taking to power these operations sustainably?
Arthur Lewis: Many companies, including Dell, are exploring unconventional energy sources to power data centers more sustainably. There’s a growing interest in solutions like nuclear power, which can provide a stable and low-carbon energy supply. We’re also investing in more efficient cooling technologies, like liquid cooling, which can cut energy costs significantly.
Editor: Speaking of liquid cooling, it seems to be a game-changer. What benefits does it offer for data centers, especially with the rise of generative AI?
Arthur Lewis: Liquid cooling is indeed a revolutionary approach. For instance, our Integrated Rack 7000 can manage massive heat outputs, and it captures nearly 100% of the heat generated. This not only enhances operational efficiency but also leads to a reduction in power consumption—up to 40%—which is essential as we scale up operations for AI applications.
Editor: That’s impressive! At the Lenovo Tech World event, they unveiled a new liquid cooling system too. How do you see the industry adapting to these innovations?
Arthur Lewis: The industry is responding rapidly. Lenovo’s Neptune system indicates a clear pivot toward more efficient cooling technologies. I anticipate that we’ll see a trend where more companies collaborate in ecosystems to leverage these innovations, allowing for greater flexibility and modularity in data center designs.
Editor: It sounds like a shift towards more adaptive frameworks. But what about data centers that aren’t ready for liquid cooling yet? How can they manage their heat generation?
Arthur Lewis: Great question! Not every data center has the infrastructure for liquid cooling. Companies like Giga Computing are addressing this by offering hybrid solutions—servers that combine liquid cooling with traditional air cooling. This gives data centers a phased approach to adopting advanced cooling systems.
Editor: As this technological evolution continues, what do you envision for the future of data centers?
Arthur Lewis: I envision a future where data centers are not only more efficient but also more sustainable. We’ll likely see more integration between AI and operational technologies, leading to smarter data centers that can optimize power usage and cooling dynamically. Ultimately, this will help us support the demands of AI while minimizing our environmental impact.
Editor: Exciting times ahead indeed! Before we wrap up, what’s one piece of advice you would give to organizations looking to modernize their data centers?
Arthur Lewis: Start by assessing your current infrastructure and understanding your future needs. Embrace innovation, whether it’s through liquid cooling or alternative energy sources. Sustainability should be at the forefront of your strategy to accommodate the ever-growing demands of AI.
Editor: Wise words! Thank you, Arthur, for sharing your insights with us today. It’s clear that as we navigate the challenges of AI, the evolution of data centers will play a crucial role.
Arthur Lewis: Thank you for having me! It was a pleasure to discuss these pivotal developments.