Memory-chip stocks emerge as a $3 trillion AI market force
Portfolio managers are tracking “memi” stocks as AI data-center spending boosts memory-chip makers and reshapes fund performance.
By Maya Lindqvist · Senior Technology Correspondent
3 min read
A cluster of memory-chip stocks has become one of the main equity-market winners from the AI data-center boom. Portfolio managers have started calling the group “memi,” a shorthand for memory semiconductors, as demand from large technology companies strains supply, Fortune reported.
The informal group has no official market classification or ticker. Fortune reported that Amazon, Google, Meta and Microsoft are among the companies competing for chips needed in AI data centers, with much of the supply coming from Micron Technology, SK Hynix and Samsung, each now valued at $1 trillion or more.
The concentration matters for investors because the same memory-chip trade is showing up across funds that may appear diversified. Fortune reported that exposure to the companies is helping drive returns in U.S. small-cap, international and emerging-market strategies, even as a pullback in big tech spending could pressure the stocks.
AI demand lifts memory suppliers
Micron shares are up 240% this year, and the company’s market value is $1.1 trillion, according to Fortune. SK Hynix listed on Nasdaq this month after raising $26.5 billion in what Fortune described as the largest U.S. listing by a foreign company, while Samsung has gained 116% this year on the Korea Exchange.
The “memi” term came from Harbor Capital’s 2026 midyear outlook call with investors, Fortune reported. Spenser Lerner, Harbor’s head of multi-asset solutions, said the wordplay was intentional because memory has become a central part of the semiconductor story.
Lerner said semiconductor stocks tied to AI have surged this year, including companies focused on memory, power, analog and compute chips. He said those gains have come as the Magnificent Seven technology stocks — Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla — have lagged that part of the market.
Nvidia’s processors remain closely associated with AI, but memory chips have taken on a larger role because AI models need DRAM to store data during training, according to Fortune. Harbor said in its outlook that DRAM pricing has stayed firm and that added supply is unlikely to matter before 2028, which could support pricing power for Micron, SK Hynix and Samsung for two more years.
Micron CEO Sanjay Mehrotra said on the company’s June earnings call that AI’s demand for memory bandwidth was “insatiable,” according to Fortune. Micron reported quarterly revenue of $41.5 billion, up 74% from the prior quarter and 346% from a year earlier; DRAM revenue reached $31.3 billion, or 76% of total revenue.
Returns spread across markets
Roundhill Investments launched a memory-chip ETF under the ticker DRAM in April, with Micron, SK Hynix and Samsung among its largest holdings, Fortune reported. The fund has returned 162% since launch, after peaking at 180% in June.
Lerner said memory and related chip companies are also influencing broader indexes. MSCI data cited by Fortune show emerging markets returned 43.51% over the trailing year, with much of the gain tied to Samsung, SK Hynix and Taiwan Semiconductor Manufacturing Company; Lerner said earnings growth drove most of the performance, and Korea and Taiwan now account for 51% of the index.
In developed markets outside the U.S., the MSCI EAFE index returned 20.8% over the trailing year, while MSCI’s Japan index gained 29.5%, according to Fortune. Lerner said much of that strength came from Japanese chip-equipment and memory names, including Tokyo Electron and Kioxia Holdings.
MSCI’s small-cap index returned 30.2% over the trailing year, and Fortune reported that its largest holding is Sandisk, a maker of flash-memory cards. Lerner said equity performance this year can be explained in large part by AI bottlenecks and where cloud-computing spending is flowing.
Fortune noted that memory chips have a history of boom-and-bust cycles, as manufacturers add capacity and later create oversupply that pushes prices lower. Some investors argue AI infrastructure spending could change that pattern, but a reversal in demand would test whether the memi trade can last.
This story draws on original reporting from Fortune.