AMD unveils new server AI chips at San Francisco event

By Max A. Cherney

SAN FRANCISCO (Reuters) – Advanced Micro Devices said on Thursday it plans to start mass production of a new version of its artificial intelligence chip called the MI325X in the fourth quarter of the year as it seeks to strengthen its AI chips in a market dominated by Nvidia.

AMD announced at an event in San Francisco that it also plans to launch its next-generation MI350 series chips in the second half of 2025. These chips include a larger amount of memory and will also boast a new core architecture that the company said would improve it. significant performance over the previous MI300X and MI250X chips.

Demand for AI processors from big tech firms like Microsoft and Meta Platforms far outstrips supply from Nvidia and AMD, allowing semiconductor companies to sell as much as they can make.

That’s led to a massive rally in chip stocks over the past two years, with AMD shares up about 30% from a recent low in early August. But investors are eager to see new technology improvements from these companies to determine whether they can maintain this momentum.

AMD shares were down 3.3% on Thursday afternoon, while rival Nvidia shares were up 1.7% and Intel shares were down 0.5%.

“There are no new clients announced so far,” Summit Insights research analyst Kinngai Chan said, adding that the stock had gained ahead of the event in anticipation of “something new.”

Santa Clara, California-based AMD said vendors such as Super Micro Computer will begin shipping the MI325X AI chip to customers in the first quarter of 2025. AMD’s design aims to compete with Nvidia’s Blackwell architecture.

The MI325X chip uses the same architecture as the already available MI300X that AMD released last year. The new chip includes a new type of memory that AMD said will speed up AI calculations.

AMD also unveiled several networking chips that help speed data transfer between chips and systems in data centers.

The company has announced the availability of a new version of its server central processing unit (CPU) design. The family of chips previously codenamed Turin includes a version of one of them that is designed to keep graphics processing units (GPUs) fed with data – which will speed up AI processing.

The flagship chip has nearly 200 processing cores and is priced at $14,813. The entire line of processors uses the Zen 5 architecture that offers speed gains of up to 37% for advanced AI data analysis.

In July, AMD raised its AI chip estimate to $4.5 billion for the year from its previous target of $4 billion. Demand for its MI300X chips has grown due to the frenzy around building and deploying generative AI products.

The story continues

This year, analysts expect AMD to report data center revenue of $12.83 billion, according to LSEG estimates. Wall Street expects Nvidia to report data center revenue of $110.36 billion. Data center revenue is a proxy for the AI ​​chips needed to build and run AI applications.

(Reporting by Max Cherney in San Francisco; Additional reporting by Aditya Soni and Arsheeya Bajwa in Bengaluru Editing by Sonali Paul, Peter Henderson and Matthew Lewis)