Advanced Micro Devices, Inc. (NASDAQ: AMD) has announced plans to increase production of its new artificial intelligence chip, aiming to compete with market leader NVIDIA Corporation (NASDAQ: NVDA). The company, led by CEO Lisa Su, revealed the details at an event held in San Francisco.
The upcoming MI300X chip from AMD will feature an impressive 192 gigabytes of memory, surpassing the capacity of any current Nvidia chip. This significant memory capacity is a crucial performance metric that determines how well the chip can handle large AI systems like ChatGPT. According to Su, customers can expect sample chips to be available in the third quarter, with full production scaling up by the end of the year.
Su emphasized the importance of AI in driving future silicon consumption, stating, “There’s no question that AI will be the key driver of silicon consumption for the foreseeable future.”
In addition, AMD plans to introduce a computer system that combines eight MI300X chips to directly compete with Nvidia’s offerings.
The company also confirmed that it has commenced shipping high volumes of its “Bergamo” general-purpose central processor chip to major companies, including Meta Platforms. Meta, the parent company of Facebook, has adopted the Bergamo chip, which caters to a different sector of AMD’s data center business focused on cloud computing providers and other large chip buyers.
Currently, Nvidia dominates the AI computing market, holding between 80% to 95% of the market share, according to analysts. In the previous month, Nvidia’s market capitalization briefly surpassed $1 trillion, following news of increased chip supplies to meet the growing demand.
In response to AMD’s news, NVIDIA’s shares rose by 3.3% on Tuesday, adding to their already remarkable 170% year-to-date surge. Conversely, AMD’s stock experienced a 4% decline following the announcements but still reached a 16-month high. Overall, AMD’s stock has doubled in value since the beginning of this year.