Cantor expects the semiconductor sector to outperform again in 2026 due to a continued buildout of artificial intelligence infrastructure and memory supply constraints.
The Philadelphia Semiconductor Index (SOX) has increased 40% year to date. In comparison, the S&P 500 Index (SP500) and the NASDAQ Composite Index (COMP:IND) rose by 15% and 19%, respectively. Cantor expects this discrepancy to continue.
Cantor’s top picks in AI compute include Nvidia (NVDA) and Broadcom (AVGO); Micron Technology (MU) in DRAM; Western Digital (WDC) and Seagate Technology (STX) in storage; ASML (ASML), Lam Research (LRCX), KLA (KLAC) and MKS (MKSI) in semiconductor capital equipment; and Teradyne (TER) and FormFactor (FORM) on the back end.
“Led by AI and the derivative impact to memory, storage, and networking, we model total Semi revenues of $1T+ in CY26, 4 years ahead of the much discussed McKinsey publication from 2022,” said Cantor analysts, led by C.J. Muse, in an extensive Tuesday report. “Compute is leap-frogging Mobility at the leading-edge, while the robust growth in Accelerated Compute has led to meaningful shortages across TSMC (TSM), COWOS/AP, and Memory/Storage.”
Within analog, Cantor expects Analog Devices (ADI), NXP Semiconductors (NXPI) and Microchip Technology (MCHP) as relative outperformers and Texas Instruments (TXN) and On Semiconductor (ON) as relative underperformers.
Cantor upgraded KLA to Overweight from Neutral and Microchip Technology to Overweight from Neutral. Cantor also increased its price target on KLA to $1,500 from $1,350 and Microchip Technology to $85 from $65.
Other price target changes included Analog Devices to $350 from $300, Applied Materials (AMAT) to $350 from $300, AMD (AMD) to $300 from $350, ASML to $1,533 from $1,356, Lam Research to $210 from $170, Marvell to $100 from $110, Qualcomm (QCOM) to $185 from $170, Teradyne to $240 from $200, Texas Instruments to $190 from $170, and Western Digital to $250 from $200.
“Obviously, there is much debate related to a so-called AI Bubble and Circularity financing conundrum,” Muse said. “But it’s worth highlighting the palpable difference in vision from AI Engineers vs. the more traditional financial markets. Hyperscale players are generating excellent ROICs from AI, tangible productivity gains are arriving, with Agentic AI/Enterprise AI only now on the come. We believe this is the breakthrough technology of our lifetime and will be transformational across all industries supporting an elongated cycle for AI investments. Yes, a handful of players may get left behind, but that doesn’t change the multi-year nature of this buildout.”