Project Helix will see Dell and Nvidia combine their hardware and software infrastructure to help enterprises build and manage generative AI models on-premises.
The Facebook parent said that it is working on a new AI-optimized data center design and the second phase of its 16,000 GPU supercomputer for AI research.
Google says its TPU v4 supercomputer is more powerful and efficient than ever, thanks to optical circuit switching technology and architecture, and challenges Nvidia.
As businesses adopt AI, high performance computing and cloud technology, servers gobble up memory chips to the point where there is more DRAM inside servers than in all mobile devices combined, TrendForce reports.