AMD has set forth an ambitious vision for AI-driven data centers at its recent financial analyst’s day, as articulated by CEO Lisa Su. The technology firm anticipates the total addressable AI market to exceed $1 trillion by 2030, a monumental increase from its previous forecast of $500 billion by 2028. The relentless demand for AI solutions could potentially boost AMD’s revenue growth to 35% annually over the next three to five years.
Su revealed expectations of a 60% increase in data center revenue from the current $16 billion by 2025, asserting that the AI data center market will also surge to $1 trillion in the next five years. This encompasses all hardware including GPUs, CPUs, and networking equipment.
AMD has landed significant data center contracts recently, including a major deal with OpenAI and a plan to deliver 50,000 chips to Oracle. Additionally, the company is working on two advanced supercomputers for Oak Ridge National Labs.
Focusing on its EPYC server CPUs, senior vice president Dan McNamara noted that AMD initiated a strategic vision three years ago aimed at disrupting the server market. With the launch of two generations of EPYC products and a growing partner ecosystem, AMD claims a dominant 40% share of the server CPU market. This success is not merely coincidental, as it coincides with challenges faced by rival Intel.
McNamara highlighted the increasing adoption of EPYC among Fortune 500 companies, boasting a threefold increase this year. The EPYC processors are designed to enhance performance for substantial enterprise workloads, facilitating a seamless transition to hybrid multi-cloud environments.
To support its growth plans, AMD will rollout new CPUs and GPUs driven by the Zen 6 architecture in 2026. The Venice generation of EPYC server processors aims to leverage advanced design technologies and improved throughput capabilities.
The new Zen 6 CPUs, constructed via a 2-nanometer process, promise backward compatibility with existing AM5 motherboards. Enhanced AI capabilities are expected to be part of this generational leap, although finer details are yet to be disclosed.
AMD also unveiled its plans for the Instinct MI400 series of GPU accelerators, scheduled for release in 2026. This series will focus on doubling compute performance over its predecessor, the MI350, helping to accelerate AI training and high-performance computing tasks.
In addition, AMD has teased the upcoming Instinct MI500 series, expected to launch in 2027, which is currently in advanced design phases. As AMD gears up for a future filled with innovations in AI-driven technology, its competitive edge in the market appears increasingly formidable.
