|
Chipmaker Qualcomm and Amazon will jointly develop multiple generations of customised chips for Amazon Web Services' (AWS) growing artificial intelligence data centre infrastructure, with the companies focusing on chips used for AI inference.
According to a Qualcomm press release issued on Tuesday, the multi-generation collaboration will support large-scale AI data centres and will also cover high-speed optical connectivity used within Amazon's data centre networks. "As AI demand accelerates, data center infrastructure will require advances in both computing and connectivity to deliver greater performance with more efficiency," Qualcomm President and CEO Cristiano Amon said. He added, "Qualcomm is pleased to work with AWS on customized silicon and connectivity solutions, bringing decades of leadership in advanced processing and power-efficient compute, to deliver breakthrough performance and enable the next generation of AI infrastructure." Qualcomm said AI workloads are driving higher demand for computing, storage, networking, memory bandwidth and energy-efficient data centre infrastructure. Apart from customised silicon, Qualcomm Technologies and Amazon will work on high-performance optical connectivity solutions to handle increasing data and bandwidth requirements in AI data centres. The companies are working on optical connectivity solutions extending up to 1.6T as well as future-generation technologies, the release said. The collaboration will use Qualcomm's SerDes and optical digital signal processing technologies. AWS Vice President Prasad Kalyanaraman said the partnership would focus on improving the performance and cost of AI infrastructure. "By working together on customized silicon and advanced connectivity, we're delivering more performant, efficient, and cost-effective infrastructure for our customers," Kalyanaraman said. Qualcomm also plans to increase its use of AWS' AI infrastructure, including Amazon Bedrock, for electronic design automation workloads. The company said the move is aimed at reducing the time required for chip design cycles. The announcement expands Qualcomm's push beyond its traditional mobile-chip business into infrastructure used for large-scale AI computing. The Qualcomm release did not disclose financial terms of the collaboration or provide a timeline for the rollout of the customised chips. (ANI)
|
