Kioxia Shares Hit New High After AI Fuels Huge Profit Increase

Kioxia said it expects operating profit of ¥1.3 trillion in the three months to June, topping its full-year profit from the prior year and sending shares up 16% to a record high. Management also said NAND supply should stay tight through 2027, with demand strengthening across customers and long-term AI data center contracts being discussed for 2027-2028. Analysts lifted price targets after the report, and the company is weighing dividend payouts and enhanced shareholder returns.

Analysis

The key market implication is not just that NAND pricing is tight; it is that the industry may be entering a rare period where contract visibility extends far enough to compress the normal reset risk in memory equities. If large AI data-center buyers are locking supply for 2027-2028, that effectively pulls forward valuation support for the whole NAND complex, because investors will start capitalizing mid-cycle earnings as if they are closer to structural earnings power. That usually benefits the highest-beta memory names first, but it also raises the bar for any “temporary oversupply” thesis in the next two quarters.

Second-order winners are likely the equipment and materials vendors tied to storage buildouts and advanced packaging, even though NAND itself is the headline. If storage demand is being pulled by AI infrastructure, the incremental capex may spill into controller, substrate, test, and data-center adjacent components with less direct market attention than GPU supply chains. The more important read-through is that memory pricing strength can improve capex discipline across the ecosystem: once producers believe pricing is sustainable, they tend to prioritize supply growth over share gains, which keeps margins elevated longer than consensus expects.

The main risk is that the market extrapolates this into a straight-line upcycle, when NAND remains more elastic than HBM and can still self-correct if customers pause orders after securing long-term inventory. The critical time horizon is 6-12 months: near-term momentum is likely intact, but if handset/PC demand stays weak while AI-related buying slows, the tightness narrative can fade quickly and multiples will compress before fundamentals do. Also, any announcement of aggressive capacity adds by the largest incumbents would matter more than macro data in reversing sentiment.

The contrarian view is that the move may be underpriced, not overdone: the consensus may still be treating this as a cyclical earnings spike, when it may actually be the start of a longer-duration re-rating for memory assets. The better signal is shareholder return policy and contract tenor, not spot prices; if capital returns are meaningful, this can support the equity even after earnings growth normalizes. For now, the setup favors owning the supply chain with the cleanest operating leverage and avoiding names where balance-sheet stress could force future supply growth just as the cycle peaks.

Key Decisions for Investors

• Add to high-beta memory exposure on pullbacks over the next 1-3 sessions; use a basket of NAND-linked equities or proxies with the goal of capturing continued multiple expansion if contract visibility into 2027-2028 is confirmed.
• Initiate a 3-6 month long/short: long memory suppliers with stronger pricing power and capital-return upside, short any weaker competitor or upstream beneficiary where earnings quality is more cyclical and less contract-backed. Target 15-20% relative outperformance if memory stays tight.
• Buy call spreads on NVDA over 3-6 months as an indirect beneficiary of storage tightness and AI data-center capex persistence; the trade is lower delta but favorable if AI infrastructure spending broadens beyond accelerators.
• Set a tactical alert for any announcement of incremental NAND capacity or customer inventory normalization; if visible, take profits on 30-50% of the memory long within 2-4 weeks, as that is the clearest catalyst for a sentiment reversal.
• If listed exposure to Kioxia is accessible, consider a momentum-following long with tight downside stops for 1-2 weeks, but size modestly given gap-risk after a 300% year-to-date move.