Broadcom’s AI Trajectory: Rosenblatt Projects $100B Revenue Milestone by 2027

Broadcom Inc. (AVGO) is increasingly being viewed as the primary alternative to Nvidia in the generative AI hardware space, a sentiment solidified by Rosenblatt Securities' latest analyst note. By projecting that Broadcom’s AI-related revenue could exceed $100 billion by 2027, Rosenblatt is signaling a massive shift in the semiconductor landscape. This forecast is not merely a reflection of general market enthusiasm but a calculated bet on Broadcom’s dominance in two critical sub-sectors: custom accelerators (ASICs) and high-end networking silicon. As hyperscale data center operators like Google, Meta, and Microsoft look to optimize their AI workloads, the demand for tailored silicon that offers better performance-per-watt than general-purpose GPUs has skyrocketed, placing Broadcom in a unique position of strength.

The core of Broadcom’s AI thesis lies in its partnership with major cloud service providers. Unlike Nvidia, which sells off-the-shelf H100 and B200 GPUs, Broadcom co-designs custom AI chips with its customers. This 'custom silicon' approach allows hyperscalers to integrate their own intellectual property directly into the hardware, resulting in significant efficiency gains for specific AI models. Rosenblatt’s $500 price target reflects the belief that this segment is still in the early innings of a multi-year growth cycle. As these custom chips move from pilot phases to mass production, Broadcom’s revenue mix is expected to shift dramatically toward high-margin AI components, potentially decoupling its stock performance from the broader, more cyclical semiconductor market.

“

By projecting that Broadcom’s AI-related revenue could exceed $100 billion by 2027, Rosenblatt is signaling a massive shift in the semiconductor landscape.

Beyond custom chips, Broadcom’s networking business remains a formidable moat. AI clusters require massive bandwidth to synchronize thousands of processors, and Broadcom’s Tomahawk and Jericho switching silicon are the industry standards for Ethernet-based AI fabrics. While Nvidia promotes its proprietary InfiniBand technology, the broader industry trend is leaning toward open Ethernet standards to avoid vendor lock-in. Broadcom is the primary beneficiary of this 'Ethernet-for-AI' movement. The company’s ability to provide the full connectivity stack—from physical layer transceivers to high-capacity switches—ensures that it captures value at every node of the AI data center. This dual-threat capability in both compute (ASICs) and connectivity (Networking) is what differentiates Broadcom from competitors like Marvell Technology.