This blog is co-authored by Marlee Morgan.
In the ever-evolving landscape of information technology, artificial intelligence (AI) stands at the forefront, ushering in a new era of data center design and operation. At Dell Technologies World 2024, Mindy Cancila, VP of Corporate Strategy at Dell Technologies, Wes Cummins, CEO and Chairman of Applied Digital, and Jeremy Williford, VP of Strategic Sales at NVIDIA, explored the transformative power of generative AI and the need to redefine data center architectures, push the boundaries of power density and champion sustainable computing.
Key Takeaways
- The future of data centers
- The critical role of sustainable computing
- The importance of understanding data structures is key to leveraging AI effectively
- Partnerships and open ecosystems are crucial for the advancement of AI technologies
Defining the Future of Data Centers
Cancila kicked off the discussion by asking Cummins about the concept of next-generation data centers. “This is all about compute…moving from air cooled to liquid cooled, but really, high power density is what we’re building, with large amounts of power,” explained Cummins. Cancila then shifted gears and asked Williford why we need to think about our next-generation centers differently. He emphasized the critical need to “look out over the landscape of how we do generative AI, as it requires a whole new data center implementation.” This includes a shift from traditional data center networking to data centers that support the high bandwidth and low latency required by north-south traffic for high data throughput at scale and east-west traffic providing the connectivity between GPU servers to boost performance for AI training and inference inside clusters.
Sustainable Computing as a Core Focus
The topic of sustainable computing began with Cancila addressing its intersection with AI within data centers. “You can’t have a generative AI conversation without talking about sustainability,” Cancila noted, emphasizing the dual focus on technological advancement and sustainable computing. Cummins, when asked about approaches to defining ESG targets, shared, “We worked with great organizations like NVIDIA and were able to take best practices to drive towards the most-efficient building and infrastructure we can build.” This reflects a strategic approach to power and performance and the structural impact of data center operations. “Sustainability can’t be an afterthought in the conversation about generative AI,” Mindy Cancila pointed out. She stressed the need for future data centers to be operated as efficiently and with as minimal an environmental footprint as possible, a sentiment echoed by her fellow panelists. Williford added, “People need to understand what their data structures look like.” He advocated for starting with a workshop to grasp the fundamentals of data and its application in a sustainable, AI-driven environment.
The Role of Data Structures in AI
Williford underscored the importance of understanding data structures in the context of AI. He praised Dell for providing the necessary expertise through its consulting services. “They’re one of the best in the market to get customers from where they are to develop their own use cases.” Cancila asked about setting efficiency targets. “There are a lot of facets that go into optimizing power so that we’re getting more productivity out,” Williford explained, highlighting the various elements that contribute to a data center’s efficiency.
A data structure is a specialized format for organizing, processing, retrieving, and storing data. It defines how data is arranged in memory and how operations like insertion, deletion and traversal are performed. Data structures are essential in AI for efficient data management, algorithm implementation, performance optimization, memory management, parallel processing and pattern recognition.
Powering the AI Revolution
As the conversation shifted toward the future of AI, Williford explained the computing paradigm shift from a traditional instruction-driven model to one that is intention-driven with generative AI. Cancila asserted that partnerships and open ecosystems are the key to unlocking a new level of productivity. “I think that’s where we really unlock a whole other level of productivity,” she said, recognizing the potential for innovation through collaboration. Cummins agreed, “You have to think about the importance of open ecosystems and the partnerships like Dell’s forging.” He underscored the critical role of collaboration in advancing data center technology.
The panelists provided a compelling vision of the future of data centers in an AI-driven world. The discussions highlighted the need for a new perspective on data center design, focusing on sustainable computing, understanding data structures and embracing open ecosystems for collaboration. As we move forward, these principles will be foundational in realizing the full potential of AI within the data center.
Learn how you can accelerate your innovations with AI.
Click the video below to watch the full conversation.
