Challenging NVIDIA! AMD releases next-generation AI chips and server products, claiming superior performance over competitors

According to Zhitong Finance APP, on Thursday, AMD (AMD.US) released a series of new products aimed at data centers, including next-generation server processors, AI accelerators, and server architectures, claiming that the performance of several products will surpass that of its competitor NVIDIA (NVDA.US), further competing for the rapidly growing artificial intelligence (AI) computing power market.

At a product launch event held in San Francisco, AMD introduced the new generation Epyc server processors, MI400 series AI accelerators, and a brand new Helios server rack design. The company stated that the new products are significant upgrades to the existing product line, offering performance advantages over competitors.

AMD is demonstrating increasing confidence in narrowing the gap with NVIDIA in the AI infrastructure field. In recent years, NVIDIA has dominated global AI data center construction with its GPU products, generating hundreds of billions of dollars in revenue. AMD believes that as enterprises' demand for AI computing power continues to grow, both companies are accelerating the launch of more powerful next-generation chips.

However, the capital market reacted relatively cautiously to this launch event. AMD's stock price fell about 3.5% during Thursday's trading session. Prior to the event, the company's stock price had already more than doubled this year, indicating that the market had high expectations for the new products.

AMD expects that by 2030, the global AI accelerator market will grow to approximately $1.4 trillion, while the traditional data center CPU market will reach $220 billion. Overall, the company anticipates that by 2030, its total addressable market (TAM) will reach about $2 trillion.

At the launch event, executives from several technology companies showed support for AMD's new products, including leading AI model developers OpenAI and Anthropic.

Among them, AMD highlighted the new Helios server rack. This solution features an open design that can be produced and sold by server manufacturers, with a single rack capable of housing up to 72 MI400 series AI accelerators. The MI455 model has achieved significant improvements in memory capacity, inference speed, and operating costs compared to the previous generation.

Sachin Katti, head of network infrastructure at OpenAI, stated that as the demand for AI infrastructure continues to grow rapidly, the company plans to deploy Helios servers on a large scale.

He said, "If you haven't heard enough, I need more computing power, and I need to get more computing power as soon as possible."

Meanwhile, AMD also launched the new generation Epyc server processors based on the Venice architecture. The company stated that this product will continue to lead in performance over NVIDIA's corresponding Vera server CPU, which NVIDIA had previously claimed would help the company further enhance its competitiveness in data centers.

Additionally, AMD announced a partnership with AI chip company Cerebras Systems (CBRS.US). The two parties will jointly launch an AI infrastructure solution that integrates Helios servers and Cerebras systems, targeting enterprise customers with extremely high requirements for inference speed According to the cooperation plan, AMD chips will be responsible for processing and interpreting user requests, while Cerebras' AI chips, equipped with ultra-large capacity storage, will quickly generate inference results. Both parties stated that the plan will first be deployed to Cerebras' own data center in the fourth quarter, and then opened to other AI companies.

Boosted by the cooperation news, Cerebras' stock price rose 4.86% on Thursday.

Cerebras CEO Andrew Feldman stated that the launch of related products by both parties will occur earlier than NVIDIA's similar solutions after integrating the recently acquired assets of Groq. He expressed confidence in being the first to market, as the company has been developing this joint product for months.

Feldman said, "We will be the first to enter the market, and we are confident about it. This product is designed for customers who place a high value on response speed."