Broadcom, Marvell Set To Capitalize On AI Investments From Microsoft, Alphabet, Meta: JPMorgan

JPMorgan analyst Harlan Sur has offered important insights into current trends in cloud and hyperscale datacenter infrastructure spending. On Sunday, he emphasized the increased focus on AI and accelerated computing.

This trend – as supported by insights from major players like Microsoft Corp MSFT, Alphabet Inc GOOG GOOGL, and Meta Platforms Inc META – points towards a robust multi-year spending cycle for AI infrastructure build-out.

Per their latest earnings reports, these companies are directing their investments towards GPUs, custom AI compute processors (ASICs), and networking infrastructure. This spells good news for chip manufacturers like Broadcom Inc AVGO and Marvell Technology Inc MRVL.

Also Read: Google ‘Copy Pasted What META Did Last Quarter,’ Redditor On Q1 Earnings

Broadcom’s Strategic Collaborations

According to Sur, Meta’s decision to increase its 2024 capex guidance to $37.5 billion, up 33% YoY, underscores its commitment to scaling AI investments. The company’s plans to deploy its own custom silicon chip (MTIA AI chip, co-designed by Broadcom) indicates a potential significant revenue stream for Broadcom.

Similarly, Google’s emphasis on sustaining its technological edge through investments in infrastructure, including the Google TPU (co-designed by Broadcom), presents a promising opportunity for Broadcom.

Marvell’s Crucial Role in AI Infrastructure

Sur notes that Microsoft’s substantial near-term capex spending suggests a continuous strong AI infrastructure build-out. Microsoft also plans to exceed FY24 capex levels, going forward.

Marvell’s involvement in ramping up Microsoft’s next-gen AI processor ASIC (“Maia”) for 2026, positions the company favorably to capitalize on the growing demand for AI infrastructure, the analyst adds.

Custom AI ASIC Programs

Sur highlights that both Broadcom and Marvell, the top two global ASIC chip suppliers, are witnessing a surge in demand for custom AI ASICs. Recent investor events have highlighted strong customer adoptions and pipeline expansions, indicating a sustained demand for custom AI ASICs. This bodes well for Broadcom and Marvell.

Networking Infrastructure

AI deployments necessitate robust high-speed networking capabilities, an area where Broadcom and Marvell excel. Marvell’s 800G PAM4 DSP chipsets and Broadcom’s Tomahawk 5 switching/Jericho 3 routing silicon are witnessing increased adoption for AI interconnectivity in accelerated compute clusters.

Additionally, Marvell’s forthcoming 1.6T pluggable optical transceivers are poised to support critical infrastructure for Nvidia Corp‘s NVDA Blackwell GPU/QuantumX 800 InfiniBand switch and Google’s TPU V6/Broadcom’s Tomahawk 5 switching chipset.

What Sets Broadcom And Marvell Apart As Key Beneficiaries

Sur’s insights underscore the pivotal roles of Broadcom and Marvell in the evolving AI infrastructure landscape. While the broader tech landscape includes other players like Nvidia, Advanced Micro Devices Inc AMD, Arm Holdings Inc ARM, and Intel Corp INTC, who are well-positioned to meet the demand for AI compute; Broadcom and Marvell’s strategic positioning in custom AI ASICs and networking infrastructure sets them apart, he adds.

With increased investments from tech giants in AI infrastructure, Broadcom and Marvell are poised to capitalize on this trend and emerge as key beneficiaries in the AI infrastructure build-out.

Read Next: Microsoft, Alphabet And Meta’s Raised AI Capex Outlook Could Benefit These JPMorgan Stock Picks

Photo: Shutterstock