
Share
AMD's new MI325X AI accelerator, unveiled at Computex 2024, aims to challenge NVIDIA's dominance with cutting-edge performance and innovative design tailored for demanding AI tasks.
At the Computex technology trade show in Taipei, Advanced Micro Devices (AMD) CEO Lisa Su announced the company's latest artificial intelligence (AI) processor, the MI325X accelerator. This new chip is set to launch in the fourth quarter of 2024 and marks AMD’s aggressive push to challenge industry leader NVIDIA in the lucrative AI semiconductor market.
New Architecture and Performance: The MI325X is designed to handle the complex workloads required by modern AI applications. It leverages advanced manufacturing processes and optimized architecture to deliver significant performance improvements over previous generations.
Annual Release Cycle: AMD has adopted an annual release cycle for its AI chips, aligning with NVIDIA's strategy. This rapid development pace is driven by the fast-evolving nature of AI technology and the high demand from data centers.
The surge in generative AI applications has created unprecedented demand for advanced AI chips. These chips are crucial for training large-scale models and running complex inference tasks, which require significant computational power.

"AI is clearly our number one priority as a company, and we have really harnessed all of the development capability within the company to do that," said AMD CEO Lisa Su. She emphasized that the annual cadence is essential to meet market demands for newer products and capabilities.
The MI325X accelerator is designed with several key technical advancements:
AMD’s unveiling of the MI325X accelerator at Computex 2024 is a significant step in its strategy to challenge NVIDIA's dominance in the AI chip market. By adopting an annual release cycle and focusing on advanced technical features, AMD aims to provide data centers and AI developers with powerful and efficient hardware solutions.
Tags
Original Sources
About the author
Kai built ML infrastructure at a Bay Area startup before developing an obsession with transformer architectures and inference optimisation that eventually pulled him out of product work entirely. A stint at a compute research lab sharpened his instinct for what actually matters in a model release versus what is marketing. He writes from the inside — from the perspective of someone who has debugged the systems he is describing at three in the morning. He is allergic to hype and instinctively drawn to the unglamorous plumbing questions that everyone else skips over.
More from The Engineer →This Week's Edition
4 June 2024
88 articles
Related Articles

OpenEvidence Targets Hospitals to Expand Its AI Chatbot for Doctors
Products & Applications · 3 min

OpenEvidence Launches Voice AI to Enhance Physician Workflow
Products & Applications · 3 min

Doximity Accelerates AI Investment in 2026, Targeting Multibillion-Dollar Market
Products & Applications · 3 min
Related Articles

OpenEvidence Targets Hospitals to Expand Its AI Chatbot for Doctors
Products & Applications · 3 min

OpenEvidence Launches Voice AI to Enhance Physician Workflow
Products & Applications · 3 min

Doximity Accelerates AI Investment in 2026, Targeting Multibillion-Dollar Market
Products & Applications · 3 min
More Stories