Story

Server CPU Market to Hit $170 Billion by 2030 Amid NVIDIA-AMD AI Battle, BofA Says

ENTHMSVIIDZHZH-TWJAKOHI
Jul 22, 20262 min read
Server CPU Market to Hit $170 Billion by 2030 Amid NVIDIA-AMD AI Battle, BofA Says

Summary

Bank of America projects the server CPU market will quadruple by 2030, driven by a battle between NVIDIA's new latency-focused Vera CPU and AMD's throughput-oriented EPYC line for dominance in AI workloads.

Text size
Background

Bank of America forecasts the server CPU market will quadruple to a $170 billion total addressable market (TAM) by 2030, up from current levels. The projection comes as NVIDIA's entry into the space with its Vera CPU architecture sets up a direct confrontation with AMD over the dominant performance metrics for artificial intelligence workloads.

The Battle for AI Workloads

In a research note, BofA Securities frames the competition as a pivotal debate for the current AI infrastructure cycle, maintaining a Buy rating and a $350 price target on NVIDIA (NASDAQ: NVDA). The firm sees the market's direction hinging on a key philosophical question about how to measure and monetize advanced AI.

"The key question for investors is whether agentic AI is primarily constrained by time-to-complete an agent or number-of-agents-per-rack," analyst Vivek Arya wrote. This pits NVIDIA's focus on latency against AMD's (NASDAQ: AMD) established emphasis on concurrency and throughput.

Competing Architectures

The two chipmakers are pursuing fundamentally different design strategies to address the demands of AI. BofA highlights the competing frameworks:

  • NVIDIA's Vera: Built on an ARM-based architecture, Vera features a monolithic die with 88 custom Olympus cores. It is positioned as the core of an integrated, co-designed system that includes the Rubin GPU and other networking components, arguing that value is created across the entire platform. Its framework is rooted in latency, per-thread progress, and GPU utilization.
  • AMD's EPYC: Leverages a proven x86 chiplet design. BofA notes that, based on AMD's own estimates, its EPYC chips offer a significant throughput advantage. Its framework is rooted in concurrency, throughput, and service density.
Sample IUX Markets – In-articleAd

According to the report, AMD estimates its EPYC 9965 (Turin) chip delivers approximately 2.4x the rack-level throughput of NVIDIA's Vera baseline, with the next-generation EPYC 6 (Venice) projected to extend that advantage to roughly 3.3x.

Market Implications and Outlook

BofA suggests the upcoming AMD "AI 2026 Day" this Thursday will be a critical event, focused less on direct benchmark comparisons and more on establishing which performance metric becomes the industry standard for justifying future investment.

A secondary battleground is the underlying instruction set. The bank notes that the software incumbency of the x86 architecture used by AMD and Intel could become an important differentiator as AI expands into traditional enterprise workflows, posing a longer-term risk for NVIDIA's ARM-based approach.

Ultimately, BofA's bull case for NVIDIA rests on the success of its integrated system strategy. The firm believes the response from hyperscalers and enterprise buyers to these competing frameworks will shape capital allocation in a server CPU market it expects to grow dramatically over the next four years.

Read next

More on Stocks
Back to latest news

LATEST