AMD and Micron Eye $2 Trillion Club: Top Analyst Sees AI Powering the Next Chip Giants Continued Rise

The artificial intelligence boom has already propelled a select group of semiconductor companies into the trillion-dollar club. Now, two chipmakers could be headed toward an even more exclusive milestone.

Advanced Micro Devices and Micron Technology could eventually reach market capitalizations of $2 trillion, according to Melius Research analyst Ben Reitzes, who argues that the next phase of AI computing could dramatically increase the importance of both CPUs and memory.

In a note to clients seen by CNBC, Reitzes laid out separate paths for AMD and Micron to double their valuations after both companies crossed the $1 trillion threshold in 2026. For AMD, the opportunity centers on the traditional central processing unit.

Much of the AI boom has revolved around GPUs, the graphics processors that turned Nvidia into the dominant supplier of chips used to train and run advanced AI models. But Reitzes believes the rise of agentic AI, systems designed to independently perform tasks and make decisions, is changing the architecture of AI infrastructure.

Those systems may require more CPUs relative to GPUs, potentially expanding AMD's opportunity beyond the GPU competition that has dominated the industry's AI narrative. There is a catch. To justify another trillion dollars in market value, AMD would have to show substantially greater earnings potential.

"To achieve its next trillion, it will likely need to demonstrate that EPS power of $50+ is possible, meaning its 2030 target of > $20 in EPS needs to go a lot higher," Reitzes wrote. AMD has been aggressively expanding its AI ambitions as it tries to challenge Nvidia across hardware and software. On Wednesday, Reuters reported that AMD is acquiring AI startup World Labs in an $8.2 billion all-stock deal, part of a broader push into AI and robotics.

Micron's route to $2 trillion looks different. The Boise, Idaho-based company specializes in memory and storage chips, including high-bandwidth memory, or HBM, which has become increasingly important in AI data centers.

HBM can improve AI computing performance while reducing latency and power requirements, making memory a critical component of increasingly powerful AI accelerators. Reitzes argues that greater use of memory can deliver significant computational efficiency gains as AI systems grow more complex. But Micron carries baggage that AMD does not: the memory industry's long history of brutal boom-and-bust cycles.

"The 'memory' of memory's past is holding it down," Reitzes wrote. He suggested the company could eventually generate enough cash to repurchase roughly 25% of its market capitalization in 2027 and 2028, adding, "not even Jensen can do that," referring to Nvidia CEO Jensen Huang.

Michael Burry, the investor known for betting against the U.S. housing market before the 2007-2009 financial crisis, recently argued that the memory industry's current shortages and extraordinary pricing power will not last.

Burry expects expanding production, including additional capacity from Chinese manufacturers, to eventually catch up with demand. "Over the next two years this shortage will blow off as production catches up, and memory will have a down cycle again," Burry wrote in a September 22 Substack post.