Advanced Micro Devices (AMD.O) has introduced its latest artificial intelligence processors on Monday, outlining its strategy to develop AI chips over the next two years to challenge Nvidia’s leadership (NVDA.O).
During the Computex technology trade show in Taipei, AMD CEO Lisa Su presented the MI325X accelerator, slated for release in the fourth quarter of 2024. The escalating demand for advanced chips, crucial for AI data centers supporting intricate applications, underscores the competitive landscape in the race to develop generative artificial intelligence programs.
AMD is positioning itself to compete with Nvidia, which currently dominates the lucrative AI semiconductor market, holding approximately 80% market share. Following Nvidia’s lead, AMD has committed to a shorter release cycle, with an annual cadence of new products.
Su emphasized AI as the company’s top priority, leveraging internal development capabilities to meet market demands for innovation. The move towards an annual release cycle reflects the necessity for newer products and capabilities, ensuring a competitive portfolio.
In addition to the MI325X, AMD announced the forthcoming MI350 series, slated for release in 2025, featuring a new chip architecture. The MI350 is projected to deliver a 35-fold improvement in inference performance compared to the existing MI300 series of AI chips. Looking further ahead, AMD revealed plans for the MI400 series, scheduled for launch in 2026, based on the “Next” architecture.
Meanwhile, Nvidia’s CEO Jensen Huang unveiled plans for Rubin, the company’s next-generation AI chip platform, slated for release in 2026, encompassing GPUs, CPUs, and networking chips.
Investors, closely monitoring developments in the chip industry, seek long-term updates to evaluate the sustainability of the booming genAI rally, which continues to show robust growth.
