What is Applied Digital's focus in data center infrastructure?

Applied Digital focuses on building high-power density data centers specifically designed to handle the unique demands of AI workloads, such as those required for training and deploying AI models.

How does Applied Digital address latency in AI model training?

Their data centers are engineered to efficiently handle high latency levels during the AI model training phase, a process that requires significant computational power and extensive data management.

Why is low latency important in AI inference?

In the inference phase, trained AI models are deployed for real-time interactions, which require environments with low latency to ensure optimal performance and user experience.

Applied Digital Redefines Modern Computing...

Applied Digital Redefines Modern Computing with Unmatched Hosting Power for NVIDIA H100 Servers

Applied Digital is a leader in data center infrastructure, focusing on hosting power for NVIDIA H100 servers. The company designs data centers with the capability to handle the demands of AI model training and real-time inference, accommodating higher latencies for training but ensuring low latency for deployment.

This story was produced through MarketScale. See how Engineering & Construction teams put it to work with Partner & Channel Enablement.

Promoted content from Applied Digital on MarketScale.

By Software And Technology · January 25, 2024, 12:20 PM UTCAiAi-first Business PodcastApplied DigitalData Center

Key takeaways

Applied Digital is reshaping data center infrastructure with a focus on hosting NVIDIA H100 servers.

Their facilities are designed to manage heavy-duty AI model workloads, considering latency needs.

Wes Cummings discusses the unique challenges of AI training and inference in a podcast episode.

As the digital landscape continues to evolve at a rapid pace, Applied Digital emerges as a key player in reshaping data center infrastructure. In a recent episode of the AI-First Business Podcast, the spotlight was on Applied Digital’s pioneering approach to meeting the burgeoning demands of modern computing.

Host Tina Yazdi welcomed Wes Cummings, CEO and Chairman of Applied Digital, to discuss the revolutionary high-power density data centers that are setting new standards in the industry. The conversation delved into the unique challenges and solutions in hosting multiple Nvidia H100 servers, a task beyond the capabilities of conventional data centers. Cummings highlighted the company’s focus on building facilities specifically designed for heavy-duty workloads like AI training, which require extensive computational power and can tolerate higher latency.

Applied Digital’s facilities are engineered to efficiently manage these tasks, even with higher latency levels.
— Wes Cummings, CEO and Chairman at Applied Digital

Cummings explained the technical nuances of AI model training, a process central to technologies like ChatGPT. He emphasized the need for data centers capable of handling the immense power requirements and data transfers involved in training AI models. Unlike traditional data centers, Applied Digital’s facilities are engineered to efficiently manage these tasks, even with higher latency levels. This design is particularly crucial for the training phase of AI, where large amounts of data are processed over extended periods.

Furthermore, Cummings discussed the role of these data centers in the inference phase of AI, where trained models are deployed in lower latency environments for real-time interactions. He stressed that while training can tolerate higher latencies, the inference phase demands environments with lower latency for optimal performance.