The booming trend of AI chips continues! RBC predicts that the scale is expected to exceed $550 billion within three years.

date
10:13 16/01/2026
avatar
GMT Eight
According to the Royal Bank of Canada's capital market forecast, semiconductor revenue from the field of artificial intelligence (AI) applications is expected to increase from $220 billion in 2025 to over $550 billion in 2028.
Royal Bank of Canada Capital Markets (RBC Capital Markets) predicts that semiconductor revenue from artificial intelligence (AI) applications will grow from $220 billion in 2025 to over $550 billion in 2028. Analysts Srini Pajjuri and Grant Li at the bank stated in an investor report, "The current market is tight in supply, with delivery cycles for enterprise orders extending to 18 months, making the industry outlook clearer. Infrastructure bottlenecks may cause some projects to be delayed, but we believe this may actually lengthen and smooth out the expenditure cycle in the AI field. While customized application-specific integrated circuits (ASIC) have made progress recently, considering the rapid pace of AI technology iteration and the longer ASIC design cycle, the dominance of graphics processing units (GPUs) is unlikely to be shaken in the short term." Based on this market background, Financial Institutions, Inc. has covered multiple semiconductor companies for the first time and rated them as "outperforming the market," including NVIDIA Corporation, Micron Technology, Inc., Marvell Technology, Inc., Arm, Astera Labs, ASML Holding NV ADR, Applied Materials, Lam Research Corporation, and Lattice Semiconductor Corporation. At the same time, Royal Bank of Canada gave "industry average" ratings to Broadcom Inc., AMD, Intel Corporation, KLA Corporation, SanDisk, Qualcomm, Skyworks Solutions, and Silicon Laboratories Inc. The report believes that the demand for high bandwidth memory (HBM) will become the core growth engine and is expected to weaken the cyclical fluctuations in the memory market. Pajjuri stated, "AI workloads are moving towards reinforcement learning and distributed reasoning, both of which have high demands for memory performance. The upcoming HBM4 iteration is another significant positive, with an expected average selling price increase of 30-50%. The explosion of generative AI also drives the growth in high-capacity server memory modules (DIMM) and NAND eSSDs. Despite high memory prices potentially putting pressure on demand in the PC and smartphone markets, we expect the memory industry to remain in a state of supply shortage by 2027." In line with the views of many Financial Institutions, Inc., the bank expects strong capital expenditure growth in the wafer fabrication equipment (WFE) sector over the next two years. Pajjuri added, "Furthermore, the implementation of trends such as backside power delivery, advanced packaging, and three-dimensional structures give us reason to believe that the growth rate of the wafer fabrication equipment market will reach at least high single-digit levels in the next two years."