The first wave of the AI boom crowned Nvidia Corp. (NASDAQ:NVDA) and memory leader SK Hynix Inc. (NASDAQ:SKHY) as Wall Street’s biggest winners. But as AI infrastructure spending broadens, the next opportunities may lie deeper in the supply chain.
GraniteShares CEO Will Rhind says investors looking beyond GPUs and high-bandwidth memory should focus on three companies sitting at the heart of AI’s next growth phase: Broadcom Inc. (NASDAQ:AVGO), Marvell Technology Inc. (NASDAQ:MRVL) and Taiwan Semiconductor Manufacturing Co. (NYSE:TSM)
“Follow the bottlenecks,” Rhind told Benzinga in an exclusive email interview.
According to Rhind, the biggest investment opportunities are shifting from AI processors themselves to the technologies that connect, manufacture and support them.
Networking And Custom Silicon Take Center Stage
After GPUs and memory, Rhind sees networking infrastructure as one of the next critical constraints for AI data centers.
“After the GPU and the memory, the next winners are the companies solving how you connect and power all of it,” he said.
That points directly to Broadcom and Marvell, whose networking chips, custom silicon and high-speed interconnect technologies have become increasingly important as hyperscalers race to build larger AI clusters.
Both companies have already benefited from growing demand for custom AI accelerators and networking equipment, but Rhind believes the structural tailwinds remain intact as AI deployments continue to scale.
TSMC Sits Beneath The Entire AI Supply Chain
Rhind also highlighted TSMC, which manufactures advanced chips for Nvidia, Broadcom, Apple Inc. (NASDAQ:AAPL), Advanced Micro Devices, Inc. (NASDAQ:AMD) and many of the semiconductor industry’s largest names.
Rather than betting on a single AI chip designer, TSMC offers exposure to the broader expansion of AI silicon production. “It’s underneath everyone,” Rhind said.
As more companies develop custom AI processors, TSMC stands to benefit regardless of which chip designer ultimately captures the largest share of the market.
The Next AI Constraint May Not Be Chips
Rhind also argued investors should look beyond semiconductors altogether.
“The one people still sleep on is power,” he said. “Data centers are running into electricity limits, so the electrical and cooling infrastructure companies are turning into an AI trade in their own right.”
His comments echo a growing theme across the AI industry: while Nvidia’s GPUs remain essential, future growth increasingly depends on everything surrounding them—from networking and advanced chip manufacturing to electricity and cooling infrastructure.
For investors, that could mark the next chapter of the AI trade. As spending spreads across the broader ecosystem, companies enabling AI infrastructure may emerge as the market’s next generation of winners after Nvidia and SK Hynix.
Imagen: Shutterstock
Recent Comments