Investing.com– Advanced Micro Devices Inc presented a new line of artificial intelligence processors at the Computex 2024 convention on Sunday, only one day after rival NVIDIA Corporation unveiled its next generation of AI chips.
The rush to develop generative artificial intelligence systems has resulted in skyrocketing demand for powerful AI data center chips capable of supporting these complicated applications.
AMD has been competing with Nvidia, which now dominates the lucrative AI chip market, accounting for roughly 80% of it.
Since last year, Nvidia has made it obvious to investors that it intends to shorten its release cycle to once a year, and AMD has now followed.
AMD’s announcement comes just hours after rival and market leader Nvidia debuted its next-generation AI CPUs. The new chip design, nicknamed “Rubin,” arrived while the firm’s current architecture, “Blackwell,” was still in construction and set to launch later in 2024.
AMD also announced the MI350 series of chips, which are slated to be released in 2025 and would be based on new chip architecture.
AMD anticipates the MI350 to outperform the current MI300 family of AI chips in inference, the process of computing generative AI responses, by 35 times.
AMD also introduced the MI400 series, which will be available in 2026 and built on the “Next” architecture.
Investors, who have put billions of dollars into Wall Street’s picks-and-shovels trade, have been looking for longer-term updates from chipmakers to assess the endurance of the surging genAI rally, which has so far showed no signs of slowing.
On Monday, AMD shares remained unchanged, while Nvidia rose more than 3%. AMD’s worth has more than doubled since the beginning of 2023, but the gain pales in compared to Nvidia’s share price, which has increased more than sevenfold in the same time period.
AMD’s Su said in April that the firm expected AI chip sales of around $4 billion in 2024, up $500 million from its previous projection.
At the Computex show, AMD announced that their latest generation of central processor units will be ready in the second half of 2024.