In its fiscal fourth quarter ending November 2, Broadcom reported a 74% year-over-year increase in AI semiconductor revenue, reaching $6.5 billion. This growth was driven by strong adoption of its custom ASIC chips among major cloud providers and hyperscale data centers. The company’s ASIC solutions, co-developed with leading AI service operators, offer performance advantages and cost efficiencies that have translated into rapid market share gains against incumbent GPU vendors.

Beyond accelerators, Broadcom’s networking backlog has grown substantially, with more than $10 billion in AI switch orders at the end of fiscal 2025. Overall, the firm reported a combined AI-related backlog of $73 billion to be delivered over the next 18 months, nearly $20 billion of which is for non-accelerator content such as high-speed optical modules and digital signal processors. This dual backlog underscores the company’s ability to capitalize on both compute and interconnect bottlenecks in next-generation AI clusters.

For the first quarter of fiscal 2026, Broadcom forecasts AI semiconductor revenue of $8.2 billion, a year-over-year increase exceeding 100%. This projection implies that AI chips will account for an even larger portion of total sales, reflecting continued strength in custom chip design engagements and expanded production capacity. Management has reiterated plans to invest in additional fabrication partnerships and R&D to meet surging demand through 2027.

Broadcom ended fiscal 2025 with a market capitalization of approximately $1.6 trillion and a gross margin near 65%. The company maintains a disciplined capital allocation strategy, returning nearly 70% of free cash flow to shareholders through dividends and share repurchases. With a dividend yield around 0.7% and ample cash generation to fund ongoing M&A and R&D, investors can expect both stable income and exposure to the high-growth AI infrastructure market.