Investor fears about vendor financing in the U.S. semiconductor industry are overstated, according to Bank of America Securities analyst Vivek Arya.

The fundamental drivers of AI growth will be data center expansion and access to power, rather than financing gimmicks, he added.

OpenAI's $100 billion deal with Nvidia (NVDA  ) and its agreement with Advanced Micro Devices (AMD  ) for up to 10% of shares have fueled concerns about a potential return to "circular" financing reminiscent of the dot-com bubble, the analyst noted.

However, he emphasized that these arrangements likely account for only 5-10% of the $1.2 trillion in annual AI-related capital expenditures expected by 2030.

Nvidia, AMD, and Broadcom (AVGO  ) are among Arya's top semiconductor picks, alongside peers in the equipment, memory, and optics sectors.

While OpenAI plans to deploy 250 gigawatts (GW) of compute power by 2033, Nvidia's 10 GW and AMD's 6 GW commitments represent only about 6% of that total. And OpenAI is just one of several major AI ecosystems, which include the four U.S. hyperscalers, Arya adds.

Tesla's (TSLA  ) xAI, for example, and numerous sovereign and neocloud initiatives in the Middle East and Asia involve little or no vendor financing. Even if the projected $1.2 trillion annual AI spend materializes, it would translate to about $5.2 trillion in cumulative investment from 2026 to 2030, he argues. That dwarfs Nvidia's and AMD's combined $200 billion in vendor-linked commitments.

Arya also framed Nvidia's $100 billion OpenAI investment as a mutually beneficial structure, not a subsidy. For every $10 billion Nvidia commits, OpenAI must first secure data center space, reliable gigawatt-scale power, and $50-$60 billion in infrastructure capital. Nvidia then supplies the GPUs, generating $30-$40 billion in sales and potentially earning an equity stake in OpenAI as well.

Price Actions: AMD stock was trading higher by 3.46% to $210.40 at last check Tuesday. NVDA was down 0.33% and AVGO was down 2.58%.