•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
In the race to build ever-larger artificial intelligence (AI) systems, connectivity is quietly emerging as the next critical bottleneck after raw compute and the need for high-bandwidth memory. At the recent Computex conference, Nvidia CEO Jensen Huang underscored this shift by publicly endorsing Marvell Technology, calling the chipmaker the next trillion-dollar company.
Marvell designs the connective tissue that allows AI workloads to operate at scale. The company offers Ethernet switches, optical DSPs for high-bandwidth, low-latency links, and custom application-specific integrated circuits (ASICs).
In hyperscale AI data centers, individual GPUs do not operate in isolation. These accelerators exchange enormous volumes of data in real time across tens of thousands of racks and rows. Marvell’s solutions address this data flow by delivering high-speed fabrics, optics, and programmable networking that keep large-scale training clusters and inference deployments running efficiently.
Marvell’s exposure to both critical networking gear and custom silicon positions it as a “picks-and-shovels” player in the AI infrastructure era. By offering a variety of connectivity solutions, the company has multiple avenues for revenue and earnings expansion as AI architectures evolve.
Nvidia’s first-mover advantage in GPUs allowed the company to strike early during the initial phases of the generative AI wave. Nvidia was the first semiconductor stock to enter the trillion-dollar club, propelled by its dominant data center services operation.
Broadcom followed Nvidia’s path to a trillion-dollar valuation, supported by demand for custom AI ASICs and high-speed networking switches that complement GPU clusters. Taiwan Semiconductor Manufacturing later achieved a $1 trillion market cap as investors recognized the company’s role as an indispensable foundry partner, manufacturing the most advanced nodes for the broader chip ecosystem.
More recently, Micron Technology, Samsung, and SK Hynix joined the list of trillion-dollar chip stocks as valuations rerated and investors priced in the potential of an AI memory supercycle.
The sequence highlights how capital has flowed through the AI chip value chain: investors first rewarded the core compute layer that unlocks new capabilities, then extended those premiums to enablers of AI deployment at scale. Marvell is described as occupying such an enabling role across high-performance connectivity.
Networking and interconnect specialists are beginning to receive similar attention as investors increasingly focus on the value of moving data efficiently. The article argues that every incremental GPU added to a cluster increases the perceived importance of connectivity, putting Marvell’s networking leadership and custom silicon momentum in line with the next phase of AI infrastructure buildouts.
If AI capex continues accelerating and Marvell captures a meaningful share of the expanding connectivity and custom chip opportunity, the article states that the company’s earnings power could rise substantially over the next several years.
It also suggests that a combination of strong top-line growth and robust profit margins could support a premium valuation multiple that remains elevated among high-growth semiconductor names—making a trillion-dollar milestone achievable for Marvell.
Coinbase has launched a High Yield USDC vault within its in-app DeFi lending offering, adding a second lending option that provides exposure to a wider range of collateral assets. The product is powered by Morpho infrastructure and uses vault allocations curated by Steakhouse Financial.
