At the Advancing AI event held on October 10, AMD (AMD) officially released its new generation AI chip series, including the fifth-generation EPYC server processor, Ryzen AI PRO processor and Instinct MI325X GPU. This series of chips is scheduled to begin mass production in the fourth quarter of this year and is expected to be available to partners in the first quarter of next year.

These new chips are designed to meet the world's growing demand for AI computing and further enhance AMD's competitiveness in the AI ​​market. Currently, EPYC server processors account for 34% of the market and are expected to continue to grow in the future. Although AMD released a number of AI-related products, its stock price fell 4% after the launch event.

Analysts pointed out that investors have higher expectations for AMD and expect it to make more significant progress in the competition with rival Nvidia (NVDA). The market generally hopes that AMD can more clearly demonstrate how it challenges Nvidia's leadership in the AI ​​chip market and announce new important customer partnerships.

AMD CEO Lisa Su highlighted the company's cooperation with several technology giants at the event, including Meta (META), Google (GOOGL), Oracle (ORCL) and Microsoft (MSFT). Kevin Salvadori, Meta's vice president of infrastructure and engineering, said the company leverages more than 1.5 million AMD EPYC CPUs for generative AI computing, including training its Llama large language model. In addition, Oracle and Google spoke highly of the performance of AMD's Instinct GPU, and Google said that the demand for AI computing power is "endless."

Despite this, analysts believe AMD is still far behind Nvidia in the AI ​​chip market. Nvidia continues to hold a major position in the AI ​​accelerator space with its Blackwell series of GPUs and strong market dominance.

Although the market demand for AI chips is high, supply is limited, causing Nvidia and AMD's production capacity to be almost full. In the short term, it will be difficult for AMD to shake Nvidia's dominant position in data center revenue. Wall Street expects Nvidia's data center revenue to reach $110.36 billion in 2024, while AMD's expected revenue is only $12.83 billion.

The newly released products include a server chip with nearly 200 processing cores priced at $14,813. This series of processors adopts Zen 5 architecture, which can significantly improve AI computing speed, and the performance of some products can be improved by up to 37%.

AMD expects that its high-performance MI350 series will be launched in 2025, which is expected to significantly surpass the MI300X and MI250X in terms of performance and memory capacity, and introduce a new underlying architecture.