AMD stock rallies 12% as Google Cloud taps its EPYC processors for next-generation AI infrastructure - Livemint
AMD shares rose 12.5% on Monday, November 9, reaching an intraday high of $504.80 after announcing its fifth-generation EPYC processors will power Google Cloud's TPU 8tx systems.
The rally came after the Nvidia rival announced that its fifth-generation EPYC server processors would power Alphabet's Google Cloud TPU 8tx systems. The collaboration brings together AMD's EPYC central processing units (CPUs), Infinity Guard security capabilities, and x86 software compatibility with Google's TPU AI acceleration technology.
Confidential AI Computing
The partnership is aimed at helping enterprises deploy sensitive AI workloads, including inference, fine-tuning, and reinforcement learning, in the cloud while maintaining data protection. Under the deployment, AMD said its fifth-generation EPYC processors will manage workload orchestration, data preparation, and high-throughput communication with Google's TPU accelerators.
The system uses AMD's Secure Encrypted Virtualization (SEV) technology alongside Google's TPU architecture to extend confidential computing protection beyond the CPU to connected accelerators. This allows sensitive information, including customer prompts, model weights, and intermediate data, to remain protected while being processed.
Market Expansion
The expanded collaboration builds on AMD's growing role as a host CPU provider across AI infrastructure systems powered by GPUs and custom accelerators. According to AMD, its EPYC processors serve as host CPUs for more than 10 leading hyperscalers and frontier AI labs, supporting systems built around AMD Instinct GPUs, Nvidia GPUs, and custom AI accelerator chips.
The company said its processors help prepare data, coordinate workloads, and manage communication between accelerators and the wider computing infrastructure. AMD and Google Cloud have also collaborated across multiple generations of EPYC processors, supporting a range of general-purpose, compute-intensive, storage-intensive, and accelerated workloads.
The latest deployment extends this relationship into confidential AI computing, as enterprises increasingly look to run advanced reasoning and agentic AI applications while protecting sensitive data and intellectual property.
Related
