boom and – Memory-chip shares have leapt on the AI boom, but portfolio managers warn the market’s excitement is running ahead of the industry’s history of sharp boom-and-bust swings—especially as new compression tech could change how much memory big models need.
Outside SK Hynix’s Bundang office in Seongnam, flags fluttered on Jan. 26, 2024—an image that now feels like a snapshot of a market riding a wave. In the years since ChatGPT was launched in December 2022. investors have piled into memory-related stocks after the AI boom ignited massive demand for high-bandwidth memory. or HBM.
The rush has been dramatic. Samsung and SK Hynix—among the biggest producers of HBM chips—have seen their stock prices jump 114% and 186% higher year-to-date, respectively. In the U.S., Micron Technology has advanced 141%, and SanDisk has gained 156% in 2026.
But as the gains compound, market watchers are urging a hard look at what memory markets have always been: cyclical. One of the core beliefs behind the current bull run is that AI has “shaken off” the sector’s past pattern—demand swinging sharply while supply stays relatively fixed.
Executives have argued that AI has changed the rules and that a structural supply shortage could keep prices elevated for years. Yet that confidence is exactly what some investors say they keep running into danger—because memory has a habit of turning fast.
William de Gale, portfolio manager at BlueBox Asset Management, described the industry’s swings bluntly. He said the memory sector tends to have “enormous ups and downs. ” adding that “in the long run it’s a pretty dreadful industry.” He also warned that every time people argue the memory cycle is over—every time they treat it like a stable. long-term value story—something happens “just before it all goes horribly wrong.”.
New compression tech brings the fear into focus
The debate is no longer abstract. Alphabet’s Google. on March 24. unveiled TurboQuant—a new compression method it says could reduce the amount of memory required to run large language models by six times. The goal is to make AI models more efficient, one of the core objectives for the leading research labs.
For investors betting on years of high HBM demand, the possibility cuts straight to the business model. If TurboQuant meaningfully reduces memory needs. it could slash demand for AI memory chips—the critical component used to train large language models from companies including Google. OpenAI and Anthropic.
Deutsche Bank, in a Tuesday note, told investors to “continue to brace themselves for continuous AI-related disruption.” The bank pointed to TurboQuant as evidence, saying the release triggered a sharp decline in the share price of the biggest memory providers.
Still, the bank added that “it remains to be seen” whether the TurboQuant technique will create a structural shift in demand.
On the other side of the trade, there’s a counterargument built around timing and supply. Jon Cunliffe. head of investment office at wealth manager JM Finn. said there is scope for production to increase meaningfully over the next three years. easing supply constraints—particularly if AI demand grows at a more normal pace. He also warned about what today’s stock prices assume: that prices stay high for a long time. that companies remain disciplined about not over-investing. and that profit margins remain much better than in the past.
He added another pressure point that can feed volatility: “the sector has experienced a high degree of momentum crowding in recent weeks,” which has made it “vulnerable to a shakeout.”
Even the cautious can’t predict the turning point
Forecasting when memory supply could exceed demand is “an impossible task,” investors are told—but caution still matters, because the current pricing assumes the good times last.
Andrew Lapping. chief investment officer at Ranmore Fund Management. warned that investors should be wary of an industry with “historically average returns on capital that is priced to make very high returns in future.” He put the risk in simple terms: “A leopard does not often change its spots. ” describing how a structural shift in the memory sector may still be unlikely.
Korea’s market ride adds another layer of concentration risk
The warnings don’t stay inside chip factories—they show up in entire market indexes. Samsung and SK Hynix have been driving South Korea’s Kospi higher across 2025 and 2026. Together, the two stocks make up over 50% of the entire index.
Steve Brice. global chief investment officer at Standard Chartered. told CNBC’s Squawk Box Asia on May 13 that he believes peak optimism around Korean equities is “not too far around the corner.” He said that when he was in Korea the previous week. he and his team were advising clients to take profits on parts of their portfolios and rotate into a globally diversified portfolio.
Yet not everyone is pulling back. Some banks remain bullish on both firms. Nomura estimated that SK Hynix stock could reach 4 million won and Samsung Electronics could reach 590,000 won over the next 12 months. On those estimates, Samsung would be up about 20% from current prices, while SK Hynix would double.
The market’s central tension is still the same: investors are chasing momentum in a sector that has historically rewarded growth with sharp reversals—and now new AI-related efficiency ideas could reshape demand in ways that are impossible to price perfectly, until the cycle turns.
memory stocks HBM AI boom TurboQuant Google Samsung SK Hynix Micron SanDisk Kospi market cycle