Goldman Sachs raised its 12-month price target for Kioxia, Japan's NAND flash memory giant, to 116,000 yen on June 30, maintaining a buy rating. The core thesis: NAND supply-demand is tighter than previously expected, and the price upcycle could extend into mid-2027, with some segments lasting until 2028.
Second Upgrade in a Month
This marks Goldman's second target hike within a month. On May 31, the bank upgraded Kioxia from Neutral to Buy, lifting the target from 48,000 yen to 93,000 yen. Less than 30 days later, it raised the bar again to 116,000 yen. Kioxia shares have surged over 660% year-to-date, making it the best-performing semiconductor stock globally.
Goldman raised its operating profit forecasts for Kioxia for FY3/27 to FY3/29 by 9%, 19%, and 29% respectively, with EPS estimates up by 10%, 19%, and 29%. On a calendar-year basis, the bank expects NAND average selling prices to surge in 2026 and grow another 38% in 2027, well above its previous estimate of 27%.
Supply Brake: DRAM Priority Keeps NAND Tight Until 2028
NAND has historically been the most oversupplied and cyclical memory product due to more players than DRAM or HDD. But Goldman argues this cycle's profit peak could be higher and last longer than the market assumes. The key is on the supply side: major memory makers continue to prioritize capex for DRAM over NAND capacity. With AI demand expanding, meaningful new NAND supply is unlikely to hit the market before 2028.
Demand is supported by three pillars: rising enterprise SSD demand, substitution from tightening HDD supply, and potential impact of U.S. export controls on some Korean manufacturers' Chinese factories.
Goldman's Two Key Bets and Five Risks
Kioxia's management has signaled a stronger focus on price and margins, avoiding long-term contracts to lock in shipments. Since some first-quarter bit shipment price negotiations were still ongoing at the time of its guidance, Goldman expects Kioxia's Q1 FY2027 operating profit to reach 1.417 trillion yen when announced on July 31, above the company's own guidance of 1.298 trillion yen and the Bloomberg consensus of 1.36 trillion yen.
The investment logic rests on two pillars: Kioxia's cost competitiveness as the world's third-largest NAND maker, and its growing exposure to the data center segment — the fastest-growing part of the NAND market. As AI servers and enterprise SSDs demand high-performance storage, Kioxia could capture higher margins in the upcycle.
Goldman also flags five risks: AI investment slowdown, rise of Chinese NAND players, cost pressure or capacity utilization volatility, sharp yen appreciation, and weakening NAND demand from non-AI applications.