Over the last few years, Nvidia (NVDA +1.70%) transformed itself from a niche graphics chip designer into the undisputed king of artificial intelligence (AI) compute. Currently, the company's market capitalization is $5.3 trillion -- making Nvidia the most valuable company in the world.
Growth investors are constantly hunting for the next AI candidate capable of delivering Nvidia-style returns. Given shares have already tripled so far this year, some eyes are turning to Marvell Technology (MRVL +9.84%) as the next Nvidia.
Can Marvell ride the AI infrastructure boom to a similar height as Nvidia? Let's find out.
What does Marvell do, and why is it important for AI infrastructure?
At its core, Marvell specializes in custom silicon chips known as application-specific integrated circuits (ASICs), or XPUs. The company also sells a comprehensive suite of interconnect technologies, including Ethernet switches and optical digital signal processors (DSPs) for high-bandwidth connectivity.
In an era where training massive AI models requires linking tens of thousands of GPU clusters across server racks, Marvell's switches and photonic fabric technologies tackle the bottlenecks plaguing data movement, latency, and power efficiency that legacy architectures struggle with.
With hyperscalers like Microsoft, Amazon, Alphabet, and Meta Platforms pouring hundreds of billions into AI capital expenditures (capex), Marvell's infrastructure portfolio positions the company as a critical enabler of AI factories.
As connectivity and custom silicon continue to capture a growing share of the multitrillion-dollar AI infrastructure opportunity, Marvell's focus on the full data center chip stack -- beyond GPUs -- makes the company indispensable.
Image source: Nvidia.
Marvell has the support of both Nvidia and AMD
What lends credibility to Marvell's ambitions in custom chips is the explicit backing from two semiconductor heavyweights, Nvidia and Advanced Micro Devices. Both companies have strategic investments in Marvell, potentially signaling strong confidence in its technology. I think affirmations from chip peers of this magnitude de-risk the idea that Marvell can ride the secular tailwinds fueling AI infrastructure build-outs.
NASDAQ: MRVL
Key Data Points
Nvidia's $2 billion investment in Marvell features a partnership on NVLink Fusion and silicon photonics, integrating Marvell's products directly into an existing, dominant AI ecosystem. Recently, Nvidia CEO Jensen Huang publicly praised Marvell's role in AI data centers -- going as far as to call the company the next trillion-dollar AI chip stock.
Will Marvell become the next Nvidia?
With a market capitalization of around $275 billion amid the stock's recent surge, Marvell now trades at a premium multiple on both trailing revenue and earnings.
In reality, Marvell's absolute scale is just a fraction of Nvidia's. Replicating Nvidia's dominance would require Marvell to capture broader layers of the overall AI compute stack. This is unlikely given the competitive landscape in the chip realm and Marvell's specific focus on enabling rather than leading AI training workloads. Against this backdrop, labeling Marvell as the "next Nvidia" may prove hyperbolic.
Nevertheless, I think Marvell could realistically become a trillion-dollar stock as AI hyperscalers expand their data center footprints and demand more custom silicon. Over the next few years, Marvell is poised to emerge as a foundational player within AI chip stocks -- just not the singular architect of next-generation data centers.
