Key Highlights

  • Bank of America argues traditional CPU Demand is resurging as a complement to AI accelerator growth.
  • AMD competes both in server CPUs through its EPYC line and in AI accelerators through its MI series.
  • The dual exposure positions AMD differently from pure-play AI accelerator companies.
  • AMD's stock has benefited from broader enthusiasm for AI-exposed semiconductor names in 2026.

Why CPUs Matter Again

Bank of America's thesis is that as data centers scale up AI accelerator deployment, they also need more traditional CPU capacity to manage, orchestrate, and support those AI workloads, reversing years of investor focus shifting almost entirely toward GPUs and specialized AI chips. This view positions companies with strong server CPU franchises, including AMD's EPYC line, to benefit from a less-discussed but necessary layer of the AI infrastructure buildout.

AMD's Dual Exposure

Unlike pure-play AI accelerator companies, AMD competes across both server CPUs and AI accelerators through its MI series of chips, giving it exposure to multiple layers of Data Center spending. This dual positioning could make AMD's growth less dependent on any single product category succeeding, though it also means AMD must execute well across two highly competitive markets simultaneously.

Is AMD the Next AI Winner

Whether AMD becomes a top-tier AI winner depends on whether its MI series accelerators can meaningfully close the gap with Nvidia's dominant AI chip franchise, while its CPU Business continues benefiting from the renewed demand BofA describes. Success in both categories simultaneously would be a meaningfully different outcome than AMD being viewed primarily as a distant number two to Nvidia in AI chips alone.

Where Things Stand

AMD has a credible path to being a broader AI infrastructure winner given its dual exposure to both resurging CPU demand and growing AI accelerator sales, though it still needs to prove its accelerator line can meaningfully compete with Nvidia's dominant position.