Nvidia, AMD among group of winners as BofA thinks AI spending has 'staying power' - MSN

Bank of America believes that despite a sharp rise in semiconductor stocks in 2026, the momentum driven by artificial intelligence (AI) spending is likely to continue. Analysts led by Wamsi Mohan high... Read original →

The structural surge in AI infrastructure demand is triggering a deep reconfiguration across the semiconductor stack. The reliance on EUV and advanced packaging at 3nm and below elevates EDA (e.g., Cadence) and interconnects (e.g., Marvell) into core value-chain nodes, forging a ‘compute-connect-design’ triad. Geopolitical friction continues to inflate compliance costs, with U.S.-China tech decoupling forcing redundant capacity builds that extend lead times and capex. Facing NVIDIA’s dominance in AI training, AMD and Broadcom are accelerating custom ASIC strategies to capture inference and edge AI through vertical integration. Over the next 12–24 months, sovereign AI initiatives and industrial digitization will unlock latent demand, driving memory (Micron) and analog (TI, ADI) into new capacity ramps, while equipment makers (Lam, KLA) benefit from foundry expansions in Taiwan, China and mainland China. Current valuations remain grounded in earnings—not speculation—confirming AI spend has durable fundamentals.