Morgan Stanley reiterates KIOXIA Holdings as a Top Pick and raises target price to ¥110,000
AI summary card
Morgan Stanley reiterates KIOXIA Holdings as a Top Pick and raises target price to ¥110,000
The report believes that AI inference-related storage demand and supply discipline will drive strong FCF generation for KIOXIA in 2026-27, and an FCF/EV yield of about 10% could support a higher valuation.
- KIOXIA raised its CY25-28 flash memory EB demand CAGR forecast from 20% to 22%, with inference-related demand CAGR raised to 86%.
- The company expects average annual capital expenditure of about ¥470bn in FY3/27-3/29, only slightly above the ¥450bn planned for FY3/27, indicating continued investment discipline.
- Morgan Stanley expects that even under conservative ASP assumptions, the company can still generate about ¥400-500bn of annualized FCF in FY3/27-3/28.
- Management stated that, depending on the operating environment and growth investment opportunities, a substantial portion of accumulated excess FCF could be returned to shareholders.
- The report raises its US$-denominated GB ASP assumptions and believes the risk of NAND prices falling faster than cost declines in FY3/28 is limited.
Report interpretation
Overview
This report is Morgan Stanley's company research update on KIOXIA Holdings following Investor Day 2026. The core conclusion is positive: the research team reiterates its Overweight rating and Top Pick designation, and significantly raises the target price to ¥110,000. The report believes that AI inference-related storage demand is improving supply-demand conditions in the NAND flash memory industry, and that KIOXIA's FCF generation capability, capital expenditure discipline, product roadmap, and potential shareholder returns together form the basis for a valuation re-rating.
Core views
The report's core view is that although NAND remains cyclical, supply and demand in 2026-27 are likely to remain solid, and expanding AI inference-related storage demand together with supply-side discipline can support ASP and earnings. KIOXIA has raised its flash memory demand outlook without adopting overly optimistic assumptions, its average annual capital expenditure plan remains restrained, and the BiCS8, BiCS9, BiCS10, and AI inference-oriented SSD product lines show progress in technology and product positioning. Morgan Stanley believes that an FY3/28e FCF/EV yield of about 10% can support the share price, corresponding to an implied FY3/27e P/E of about 11x EPS.
Analysis framework
The report mainly adopts a framework of event commentary, scenario valuation, and earnings forecast revisions: it first distills information from Investor Day on demand, supply, capital expenditure, product roadmap, and capital allocation, then adjusts assumptions for ASP, capex, R&D, and EPS, using FY3/28e FCF yield as the core valuation anchor, supplemented by bull, base, and bear scenarios to present the risk-reward range.
Methodology notes
Measure valuation support using free cash flow yield relative to enterprise value
Morgan Stanley believes that in the context of stronger future FCF generation at KIOXIA and potentially increased shareholder returns, FCF/EV will become a more important valuation metric; the base-case target price uses an FY3/28e FCF yield of about 10%.
Internal forecasting and valuation model framework
The report states that unless otherwise noted, key metrics are based on the Morgan Stanley ModelWare framework, with some data estimated by Morgan Stanley Research.
Bull, base, and bear case valuation scenarios
The report uses three scenarios of ¥150,000, ¥110,000, and ¥55,000 to describe the impact of changes in supply-demand, eSSD share, NAND demand, and earnings on the share price.
Overweight means a higher expected risk-adjusted total return relative to the industry coverage universe over the next 12-18 months
Morgan Stanley discloses that its equity ratings are relative ratings; Overweight is not equivalent to a traditional Buy, but it is its most positive equity rating category.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- KIOXIA Holdings (285A.T)Primary coverage target; Japanese NAND flash memory and SSD company
- Strengths
- Rising AI inference demand, tight NAND supply-demand conditions, capex discipline, advancing BiCS product roadmap, strong FY3/27-3/28 FCF expectations, and potential shareholder returns.
- Weaknesses
- The business remains exposed to the NAND cycle, earnings are sensitive to ASP, supply-demand, and exchange rates, and there is execution uncertainty around growth investments and product rollout timing.
- Comparison
- Morgan Stanley's FY3/27 and FY3/28 EPS forecasts are 13% and 9% above consensus, respectively, indicating that its earnings assumptions are more optimistic than the market's.
- Risks
- Slowing NAND demand, capacity expansion by Chinese companies, price declines faster than cost reductions, yen appreciation, and weaker-than-expected eSSD share gains.
- NAND flash memory and eSSD marketCore industry driver
- Strengths
- AI inference, data center, and enterprise storage demand are driving EB demand growth, and the company expects CY25-28 flash memory demand CAGR to reach 22%.
- Weaknesses
- Industry supply-demand and pricing are highly cyclical, and changes in end demand can quickly transmit to ASP and earnings.
- Comparison
- The company raised its previous 20% demand CAGR outlook to 22%, but the report believes this revision is relatively restrained rather than overly optimistic.
- Risks
- Declining end demand, competitor capacity expansion, a shift from tight to loose supply-demand conditions, and price declines exceeding cost reductions.
Key data
- Report date2026-06-02Published after Investor Day 2026.
- Rating actionReiterate Overweight; maintain Top PickThe report says the overall interpretation is positive.
- Target price¥110,000Based on an FY3/28e FCF yield of about 10%, implying about 11x FY3/27e EPS.
- Bull-case target price¥150,000Assumes supply shortages continue through 2026-27 and KIOXIA significantly increases its eSSD market share.
- Bear-case target price¥55,000Assumes NAND demand slows rapidly and earnings deteriorate.
- FY3/27-3/28 annualized FCF forecast¥400-500bnThe report says the company is still expected to generate this level of FCF even under conservative ASP assumptions.
- CY25-28 flash memory EB demand CAGR22%Above the previous 20% forecast, reflecting stronger AI inference demand.
- Inference-related demand CAGR86%Updated outlook disclosed by the company at Investor Day 2026.
- FY3/27-3/29 average annual capex plan¥470bnOnly slightly above the FY3/27 plan of ¥450bn, showing capital expenditure discipline.
- BiCS8 migration targetAround 80% of annual GB shipments to transition to BiCS8 by the end of FY3/27BiCS8 combines planar scaling and CBA technology, emphasizing performance, low power consumption, and capital efficiency.
- BiCS10 timelineSampling begins in summer 2026, with mass production about one year laterTargeting high-capacity demand in enterprise and data centers.
- FY3/27 EPS vs. consensus+13%Morgan Stanley's forecast is above market consensus.
- FY3/28 EPS vs. consensus+9%Morgan Stanley's forecast is above market consensus.
Impact & implications
For investors, the report shifts KIOXIA's investment thesis beyond a simple NAND cyclical recovery toward AI inference-driven structural storage demand, resilient FCF, and potential capital returns. If supply and demand remain tight in 2026-27 and the company gains share in eSSD and data center products, valuation could have further upside; however, if demand weakens, supply expands, or yen appreciation pressures profits, earnings and target price assumptions may face downward revision risks.
Risks
- Deterioration in NAND supply-demand conditions lasting longer than expected, especially if weaker end demand leads to industry softening.
- Capacity expansion by Chinese companies may increase supply pressure and weigh on NAND prices.
- Yen appreciation against the US$ would hurt operating profit; the report estimates that every 1-yen appreciation in the yen reduces annualized OP by about ¥6bn.
- If NAND prices fall faster than cost declines in FY3/28, earnings and FCF assumptions will weaken.
- If AI inference-related storage demand or eSSD market share gains fall short of expectations, the bull-case scenario will be difficult to realize.
- There is uncertainty in capital expenditure, R&D, BiCS8/BiCS10 mass production, and execution of the product roadmap.
- Potential shareholder returns depend on the operating environment and growth investment opportunities and are not a firm commitment.
- Morgan Stanley discloses investment banking business and shareholding-related conflicts of interest with KIOXIA Holdings, and investors should treat the report as one factor in decision-making rather than the sole basis.
What to watch
- Whether NAND supply and demand remain tight in 2026-27, and whether ASP stays elevated.
- Whether AI inference, data center, and enterprise SSD demand support EB demand growth.
- Progress in LTAs with hyperscale and enterprise customers extending into CY29 and beyond.
- Whether the average annual ¥470bn capex plan for FY3/27-3/29 maintains discipline and achieves a low-teens percentage front-end GB cost decline.
- Whether the target of migrating about 80% of annual GB shipments to BiCS8 by the end of FY3/27 is achieved.
- Progress on BiCS10 sampling in summer 2026 and mass production about one year later.
- A progressive dividend policy from FY3/28 onward, and the possibility of an earlier dividend in 2H FY3/27.
- Progress in eSSD market share gains and commercialization of the CM/GP/LC series products in AI inference and server applications.
- The impact of USD/JPY exchange rate changes on OP.