Microsoft Corp. is in talks with Broadcom Inc. to take over custom chip design work currently handled by Marvell Technology Inc., according to report in The Information.
The potential partnership would have Broadcom design specialized chips for Microsoft’s Azure cloud services and artificial intelligence (AI) infrastructure, heralding a major realignment in the semiconductor industry.
The discussions reflect Microsoft’s broader strategy to diversify its supply chain and reduce dependence on single vendors, mirroring similar approaches by rivals Amazon.com Inc. and Google. Sources familiar with the matter told The Information that while Marvell has been collaborating with Microsoft on application-specific integrated circuits (ASICs) for networking and data center acceleration, the tech giant is now evaluating Broadcom’s expertise in high-performance chip design.
Broadcom has established itself as a preferred partner for major technology companies developing custom ASICs. The company recently announced a collaboration with OpenAI to co-develop AI chips for generative processing needs, a deal that boosted Broadcom’s stock and underscored investor confidence in its position within the AI boom.
For Microsoft, the potential switch appears driven by performance requirements and the need for more comprehensive solutions. While Marvell has been instrumental in providing chips for Microsoft’s networking equipment, Broadcom offers a broader portfolio including Ethernet switching and AI accelerators. Industry analysts note that Broadcom’s custom chip division has experienced substantial growth, with AI-related product revenues projected to surge significantly.
The move fits within Microsoft’s aggressive push into proprietary hardware. The company has already unveiled its own chips, including the Maia AI accelerator and Cobalt CPU, designed to power AI training and inference tasks without relying exclusively on external suppliers like NVIDIA Corp. A partnership with Broadcom could accelerate these efforts, enabling faster iteration of designs that integrate seamlessly with Microsoft’s software ecosystem.
Marvell’s stock showed volatility following the report, briefly dipping before recovering to trade modestly higher. The news comes just days after Marvell tumbled on a lackluster earnings report, then surged when CEO Matt Murphy outlined a $10 billion revenue target for the upcoming fiscal year, exceeding analyst expectations. Microsoft represents Marvell’s second-largest hyperscaler client after Amazon.
Analysts note the semiconductor industry is witnessing a fundamental shift as hyperscale cloud providers increasingly turn to custom solutions optimized for specific workloads rather than general-purpose chips. This trend has intensified with the explosive growth of AI applications, which demand unprecedented computational power and efficiency.
Microsoft’s exploration of new chip partnerships also includes a 2024 deal with Intel Corp. for custom chip manufacturing, demonstrating the company’s commitment to casting a wider net and potentially mitigating geopolitical risks in chip manufacturing.
