Bank of America Names AMD (AMD) and Nvidia (NVDA) as Premier AI Chip Investments

Table of Contents Bank of America has increased its server CPU market projection to $125 billion by 2030, representing a significant upgrade from its previous $110 billion estimate, while highlighting AMD and Nvidia as its preferred semiconductor investments to capitalize on this expansion. NVIDIA Corporation, NVDA The revised forecast reflects the emergence of agentic AI — intelligent systems capable of task planning, information retrieval, memory management, and executing complex multi-stage operations. According to BofA analyst Vivek Arya, who holds the 94th position among over 12,000 analysts monitored by TipRanks, these advanced AI architectures require substantially more CPU resources than earlier iterations. The financial institution anticipates the server CPU sector will grow at a 31% compound annual rate from 2026 through 2030, building from an approximately $43 billion foundation. Throughout the initial AI revolution, graphics processing units dominated computational workloads, particularly for training large language models. However, as artificial intelligence architectures grow increasingly sophisticated, central processing units have assumed critical responsibilities in task coordination, state management, and interfacing with database and retrieval infrastructure. Bank of America characterizes CPUs as the orchestration backbone of AI inference operations, positioning them as the essential coordination layer that enables agentic AI systems to function effectively. The firm emphasizes this development doesn’t diminish GPU requirements. Rather, it signals an overall expansion of data center infrastructure needs. Newly deployed CPU-dedicated server configurations are anticipated to address workloads that GPU-based systems cannot economically support. Bank of America maintains Buy recommendations on both companies, assigning AMD a $500 price objective. AMD is projected to capture approximately 28% of server CPU market value by 2030, with anticipated growth across both hyperscale cloud and traditional enterprise deployments. Nvidia receives favorable ratings for its comprehensive approach to integrating CPUs, GPUs, networking infrastructure, and memory solutions into unified AI platforms. Nvidia’s forthcoming Vera CPU, scheduled for release alongside its Vera Rubin platform during the latter half of 2026, is anticipated to operate in nearly equal proportion with GPUs in next-generation AI server architectures. Custom ARM-architecture processors, encompassing AWS Graviton, Google Axion, and Microsoft Cobalt, are forecast as the market’s fastest-expanding category, climbing from approximately 15% of current server CPU market value to roughly 37% by decade’s end. Intel faces projected market share erosion in both cloud infrastructure and enterprise segments. Nvidia’s latest quarterly results reinforce the CPU demand trajectory. The company posted Q1 Fiscal 2027 adjusted earnings of $1.87 per share, surpassing Wall Street’s consensus estimate of $1.75. Revenue reached $81.6 billion, marking an 85% year-over-year increase. Despite robust growth expectations, server CPUs are still forecast to constitute merely 5 to 6% of aggregate data center capital expenditure by 2030. AI accelerators are projected to maintain their position as the predominant component within a broader AI data center ecosystem that BofA values at $1.7 trillion.