Cathie Wood Avoids $MU and $SKHY Over Concerns About HBM Cyclicality

Cathie Wood says she avoids memory stocks such as $MU and $SKHY because she still views high-bandwidth memory (HBM) as the most cyclical and commoditized part of the semiconductor stack.
Her broader argument is that inference architectures from $CBRS and $NVDA through Groq could reduce or potentially eliminate some demand for HBM, making the current surge in memory prices look more temporary than structural.
Wood argues that when technology costs “triple or quadruple,” engineers eventually design around the bottleneck, and she believes AI inference is already beginning to do exactly that.