U.S. chipmakers including NVIDIA Corp (NASDAQ:NVDA), Broadcom Inc (NASDAQ:AVGO), Advanced Micro Devices Inc (NASDAQ:AMD), Marvell Technology Inc (NASDAQ:MRVL), and Credo Technology Holding Ltd. (NASDAQ:CRDO) are set to benefit from a massive wave of artificial intelligence infrastructure spending projected to top $1.2 trillion by 2030.
NVDA is in positive territory. See what is happening here.
Fueled by hyperscalers, sovereign AI projects and enterprise adoption, the unprecedented surge in capital expenditures could potentially triple between 2025 and 2030, cementing semiconductors and equipment makers like Lam Research Corp (NASDAQ:LRCX) and KLA Corp (NASDAQ:KLAC).
BofA Securities analyst Vivek Arya projects U.S. semiconductor companies will benefit heavily from an unprecedented surge in artificial intelligence infrastructure spending.
Also Read: Nvidia And Fujitsu Team Up To Build AI Systems For The Future
The analyst expects annual AI-related capital expenditures to nearly triple from 2025 to 2030, reaching more than $1.2 trillion, constrained only by the ability to scale data center buildings and power supplies.
He identifies four core drivers of this spending:
Upgrading traditional infrastructure to accelerated computing.
Defending the moats of major tech platforms in search, social media, e-commerce and cloud services.
Sovereign governments building national AI infrastructure.
Steady enterprise adoption of AI productivity tools.
Arya stresses that chipmakers such as Nvidia, Broadcom, AMD, Marvell and Credo, alongside semiconductor equipment vendors like Lam Research and KLA, stand best positioned to capture the gains.
The analyst argues that, unlike past investment cycles, this wave of spending is more durable because it is funded by robust cash flows from hyperscalers and sovereign buyers.
The top five U.S. cloud providers, he notes, are set to allocate about 25% of their 2025 sales to capital spending, supported by operating cash flow margins above 30%. At the same time, end users do not need to upgrade devices, meaning AI services like ChatGPT can scale instantly without costly consumer hardware refresh cycles, Arya noted.
Updating his cloud spending tracker, the analyst now sees capital expenditures of $443 billion in 2025 and $528 billion in 2026, reflecting year-over-year growth of 58% and 19%.
The analyst projects spending could climb to more than $1.2 trillion by 2030 if current capex intensity levels hold and U.S. hyperscalers contribute 60% to 70% of the total. While periods of digestion may occur, Arya considers the growth assumptions reasonable given the demand trajectory.
The analyst also addresses investor concerns around vendor financing. Nvidia’s $100 billion commitment to OpenAI, Arya explains, should be viewed less as “free GPUs” and more as an offensive investment to expand its ecosystem.
To access that capital, OpenAI must secure up to 10 gigawatts of power and invest an additional $300 billion to $500 billion in equipment, while also ceding $100 billion in equity, he said.
Arya sees this deal as a catalyst, pushing other hyperscalers to accelerate their competitive AI deployments.
The analyst draws parallels with the 4G cycle, when global telecom operators invested $1.3 trillion between 2010 and 2020 while end users spent $3.6 trillion on smartphones. In contrast, the current AI cycle requires no new consumer devices, which makes adoption faster and cheaper.
Nvidia remains Arya’s top AI pick, backed by its moat in accelerators, networking hardware and ecosystem, supported by upcoming catalysts like the Vera Rubin platform and its role as OpenAI’s preferred partner.
NVDA Price Action: Nvidia stock was trading higher by 0.15% to $189.17 at publication on Friday.
Read Next:
Broadcom Rides Microsoft, Meta, and Apple’s Billions in AI Chip Deals to Cement Its Role in the Tech Boom
Photo: Shutterstock