AMD has integrated AI directly into the manufacturing process in the design of its next-generation graphics processors. Speaking at the “Winning the AI Race” summit in Washington, CEO Dr. Lisa Su announced that AI is now enabling speed and reliability in GPU development. This approach, which incorporates AI not only into algorithms but also into hardware engineering, highlights AMD’s new strides in chip manufacturing.
AMD Leverages AI in GPU Design
The company has transitioned to a unified architectural design model for both its data center solutions and end-user graphics units. This model also includes the Radeon series for the desktop and console markets.
AMD positions AI not as a technology that will take over the entire process, but as a tool that optimizes complex calculations and reduces the likelihood of errors. This complementary role of AI allows human engineers to maintain control over the design.
Dr. Su emphasized in her speech that while AI has taken over some aspects of chip design, human intelligence still dominates the creative aspects. AMD’s design approach places humans at the center of the engineering process. In this model, AI serves as an accelerator and support element.
The company’s incorporation of AI into the design process aligns with strategies pursued by rival companies like Nvidia. This allows AMD to both increase its competitiveness and significantly shorten its design cycle. The introduction of AI minimizes time loss in product development, resulting in more stable and efficient chips.
AMD’s strategy isn’t limited to hardware. Dr. Lisa Su noted that AI is also driving fundamental changes in science, medicine, and manufacturing, highlighting the potential for societal benefits.
Su stated that AI can improve science, medicine, and manufacturing. AMD’s long-term plans include using AI as a tool to address major challenges facing the world.
This new approach, details of which AMD has not yet released, is expected to directly impact next-generation GPUs featuring the RDNA 5 architecture. These chips, which will be developed with an artificial intelligence-assisted design methodology, are expected to surpass current generations in terms of performance and reliability.
