Victor Peng’s departure comes as AMD expands its roadmap for Instinct processors, looking to challenge Nvidia’s dominance in AI chips for data centers.
AMD said on Monday that President Victor Peng, who has been leading the chip designer’s entire AI strategy, will step down next month.
Peng (pictured) joined AMD in 2022 when the Santa Clara, California-based company acquired Xilinx, the programmable chip designer that Peng led for four years as president and CEO. Prior to his 14-year tenure at Xilinx, Peng joined AMD through the acquisition of his previous GPU designer ATI Technologies, where he served as director of silicon engineering.
[Related: Forrest Norrod On How AMD Is Fighting Nvidia With ‘Significant’ AI Investments]
The company said Peng will remain with AMD’s management team in an advisory role until his retirement on August 30 and will assist with the leadership transition.
With Peng’s departure, AMD will hand over responsibility for its Instinct data center accelerator chip business to Xilinx veteran Vamsi Boppana, who serves as senior vice president of the company’s artificial intelligence group, leading AI software and ecosystem development as well as the AI hardware roadmap.
AMD announced that Boppana and Salil Raje, head of AMD’s Adaptive and Embedded Computing Group, which formerly operated as Xilinx, will now report to the company’s CEO, Lisa Su.
“Peng played a key role in successfully integrating and expanding our embedded business and leading our cross-enterprise AI strategy,” Su, who also serves as AMD’s chairman, said in a statement.
“Under his leadership, AMD has become the industry leader. [No. 1] “On behalf of AMD’s board of directors, executive leadership team and the thousands of employees who worked alongside him as a provider of FPGA and adaptive computing solutions, I thank Victor for his outstanding leadership and wish him the best in his retirement,” she said in a statement.
Peng’s retirement comes as AMD expands its Instinct processor roadmap, which includes the recently launched MI300 and the MI325 due to launch later this year, as it seeks to challenge Nvidia’s dominance in AI chips for data centers.
