A 14% decline is painful, but by itself it does not invalidate the memory supercycle. Memory stocks are among the most cyclical and sentiment-driven names, so sharp corrections after strong rallies are common.

The key questions are whether:

HBM and enterprise SSD demand remain strong.

Customer inventory stays healthy rather than building excessively.

Pricing for DRAM and NAND remains firm over the next few quarters.

If those fundamentals remain intact, this looks more like a valuation reset than the end of the cycle. If, however, hyperscalers begin cutting AI infrastructure spending or memory pricing weakens materially, then the thesis would deserve reassessment.

Rather than trying to call the exact bottom, I would prefer averaging in gradually over several tranches. That captures potential upside if the sell-off is driven mainly by positioning, while limiting risk if the correction deepens. The next earnings reports and management guidance will be far more important than a couple of volatile trading sessions.