BofA Securities' analysts say the global buildout of AI infrastructure is entering a more durable phase, with annual capital investment forecast to nearly triple to $1.4 trillion by 2030. The research team expects a sustained level of "capex intensity" in the range of 25% to 30% as major cloud providers and sovereign entities adapt existing IT environments for AI workloads.
Within this framework, the bank ranks AI compute as the highest priority, followed by semicaps, networking, and memory. The analysts identified a set of large-cap semiconductor and related technology names they view as well positioned for the transition. Below is a company-by-company look at the stocks BofA highlights and the facts the analysts emphasize.
Nvidia
Nvidia occupies a central role in BofA's AI infrastructure outlook as the leading supplier of AI accelerators. The firm notes that Nvidia continues to benefit from investor confidence tied to its forthcoming Blackwell chip architecture and the strong demand backdrop for its current-generation processors, particularly from large cloud and technology customers such as Meta and Microsoft.
The analysts point to Nvidia's market valuation, which exceeds $4.3 trillion, as a reflection of that confidence. They also observe recent share-price volatility - including a drop to a six-week low followed by stabilization near the upper end of the stock's 52-week range - while demand indicators remain robust.
Broadcom
BofA frames Broadcom as a major beneficiary of the move toward custom AI silicon and the expanding need for high-speed data center networking. The analysts highlight a recently disclosed agreement with OpenAI for custom chip development, a contract the bank says could potentially generate billions of dollars in revenue through 2029.
Broadcom's stock has shown mixed options market sentiment and has experienced a modest pullback over the past month, according to the analysts. The research note also cites Broadcom's record fiscal 2024 revenue of $51 billion, a milestone reached after the integration of VMware into the company's business.
Advanced Micro Devices
Advanced Micro Devices is identified as a meaningful alternative for customers seeking to diversify their AI hardware supply. BofA points to the doubling of AMD's data center revenue in late 2024 and early 2025, which the bank attributes to the rapid ramp of the company's Instinct MI300 accelerators.
Investors, the analysts say, are watching AMD to see how effectively it can capture spillover demand as supply of leading GPUs from competitors remains constrained. That dynamic has helped push AMD's shares to trade near all-time highs hit in late 2025, the research note states.
Micron
In the memory subsector, BofA highlights Micron as a strategic way to play the High Bandwidth Memory (HBM) market that feeds advanced AI accelerators. The analysts acknowledge some concerns around "peak margin" for Micron, but note the stock currently trades at the low end of its historical price-to-earnings range at a cited price of $377.76.
Micron is described as aggressively expanding production capacity, with plans to invest roughly $25 billion in fiscal 2026. The research note adds that the company's HBM supply is already reported to be sold out through that year, underscoring tightness in that segment.
Across these names, BofA's coverage underscores a thematic investment case centered on rising AI compute demand, the need for specialized silicon and networking, and the critical role of memory bandwidth. The analysts' sector prioritization and company calls reflect their view of where capex tied to AI workloads will concentrate as hyperscalers and governments upgrade infrastructure.