The burgeoning landscape of artificial intelligence is creating an unprecedented demand for specialized hardware, a topic thoroughly explored by Joseph Moore, Senior Semiconductor Analyst at Morgan Stanley, during his recent interview on CNBC’s 'The Exchange'. Moore articulated a clear vision: while a "laundry list of concerns" has shadowed the market for the past six months, including issues like Chinese competition and product delays, the underlying reality is that "usage of these AI models is explosive," necessitating a significant surge in hardware investment.
Moore emphasized that while "everyone's going to benefit" from this robust AI environment, Nvidia stands out as the primary recipient of this capital expenditure. "I do think Nvidia is the best beneficiary," he stated, attributing this to their substantial proportion of revenue derived directly from AI applications. This advantage is further amplified by the reinstatement of certain products for the Chinese market, providing an additional tailwind for both Nvidia and AMD on the processor side.
The two companies holding the strongest positions in core AI processors, according to Moore, are Nvidia and Broadcom. However, the AI-driven upside for companies like Broadcom and Marvell is expected to come predominantly from networking, rather than solely from their processor offerings. This distinction highlights the comprehensive nature of AI infrastructure build-out, extending beyond just the compute units to the crucial data transfer and connectivity layers.
Marvell's optical business, in particular, is positioned for strong near-term performance. While acknowledging that the company previously fell short of high expectations, Moore sees considerable opportunity in their optical segment.
"AI strength is exceptional in both supply and demand." The analyst pointed to an acceleration in inference workloads as a key driver of compute needs, underscoring the sustained durability of AI spending well into 2026, bolstered by new product cycles such as Nvidia's Blackwell.
Even in segments like memory and storage, crucial for AI's data hunger, opportunities abound. Moore specifically highlighted High Bandwidth Memory (HBM), a specialized DRAM designed for AI, and noted that companies like Micron are "gaining share" in this competitive space. Despite concerns about eroding premiums, the overall growth in HBM demand is expected to contribute positively to margins.
