Investors Warn on Memory Stock Boom-Bust Risks
CNBC reported Monday that seasoned fund managers are cautioning investors against assuming the memory stocks boom-bust cycle has been permanently broken, even as artificial intelligence spending drives extraordinary share price gains across the sector.
AI Tailwinds Fuel Enormous Memory Stock Gains
The memory industry entered a sustained growth phase after ChatGPT’s December 2022 launch. That event unleashed surging demand for high-bandwidth memory chips. South Korean giants Samsung and SK Hynix have climbed 114% and 186% year-to-date, respectively. US-listed Micron Technology and SanDisk have posted comparable advances of 141% and 156% in 2026.
Some executives argue AI has permanently restructured the industry. They say a structural supply shortage could keep chip prices elevated for years ahead.
“A Pretty Dreadful Industry” Over the Long Run
Not everyone is convinced. William de Gale, portfolio manager at BlueBox Asset Management, told CNBC the memory sector has historically produced “enormous ups and downs.” He warned that every time investors declare the memory cycle dead, a sharp downturn tends to follow. His view frames AI optimism as potentially the latest chapter in a recurring pattern of misplaced confidence.
Jon Cunliffe, head of investment office at wealth manager JM Finn, told CNBC that production capacity could rise significantly over the next three years. That supply growth could ease current shortages, particularly if AI adoption settles into a more moderate expansion pace. He flagged that current valuations assume persistently high prices and exceptional corporate discipline, conditions that rarely hold indefinitely. He also noted the sector has attracted heavy momentum-driven flows, leaving it exposed to a sudden reversal.
Background: A Sector Defined by Cyclicality
Memory chips have historically been plagued by mismatches between fixed supply and volatile demand. That dynamic has repeatedly produced sharp price collapses after periods of scarcity. Andrew Lapping, chief investment officer at Ranmore Fund Management, argued that a sector with historically average capital returns cannot sustainably trade as though superior returns are guaranteed. His assessment suggests current pricing carries meaningful downside risk.
New Technology Adds Another Layer of Uncertainty
A fresh threat emerged in late March when Google unveiled TurboQuant, a compression technique the company claims could reduce the memory requirements for large language models by as much as six times. Deutsche Bank analysts wrote that investors should expect continued AI-driven disruption of this kind, though they cautioned it remains unclear whether TurboQuant will produce a lasting structural shift in memory demand.
Meanwhile, Samsung and SK Hynix together account for more than half of South Korea’s benchmark Kospi index, meaning any sector-wide correction would ripple directly through the country’s broader equity market.
Read Next: Samsung Earnings Preview: Can Memory Margins Hold Through 2026?
