Tech Consortium Announces AI Accelerator Standards

The Ultra Accelerator Link Consortium, comprising major tech firms like AMD, Google, and Microsoft, aims to establish an open standard for AI accelerator links, intentionally excluding Nvidia. This move comes as Nvidia's revenue from proprietary networking solutions grows significantly, prompting competitors to seek alternatives that promote market competition and lower prices. Meanwhile, Sakura Internet Inc. plans to purchase 10,000 Nvidia GPUs annually for its upcoming data center, anticipating high demand upon its launch in three years. Cisco Systems has unveiled new data center appliances optimized for AI, featuring Nvidia GPUs and advanced networking capabilities. F5 Networks introduced BIG-IP Next for Kubernetes, leveraging Nvidia BlueField-3 DPUs to enhance AI application delivery and security. Lastly, Nvidia announced enterprise reference architectures to aid businesses in developing scalable AI infrastructures, emphasizing the growing need for efficient AI deployment solutions.