Kioxia Holdings

Phase A — Understand the business

Company Overview

Kioxia is the world's #3 NAND flash memory maker and the only pure-play among the majors — it makes flash, and essentially nothing else. There is no DRAM division, no logic foundry, no HBMHigh-bandwidth memoryMemory stacked in layers beside a processor so data reaches it faster. The usual bottleneck in AI chips., no diversification. When NAND is good, Kioxia is the best-levered name in the industry; when NAND is bad, it has nothing to hide behind. That single fact is the whole thesis.

Lineage. Formerly Toshiba Memory — the flash business Toshiba was forced to sell in 2018 to plug a hole blown open by the Westinghouse nuclear disaster. A Bain Capital-led consortium bought it for ~$18B, renamed it Kioxia in 2019, and after two aborted attempts (killed by the 2023-24 NAND collapse) finally floated it on the Tokyo Stock Exchange on 18 Dec 2024 at ¥1,455/share for a ¥886B ($5.6B) initial market capMarket capitalisationThe share price multiplied by the number of shares. What the market says the equity is worth..

What it sells (FY2025, ended Mar 2026):

• SSD & Storage — enterprise/data-center SSDs, client PC SSDs. ~55% of sales and the AI-demand engine; Q4 FY25 segment revenue ¥600.3B (+99.8% QoQ).
• Smart Devices — embedded flash (UFS/eMMC) for smartphones. Q4 FY25 ¥337.3B (+81% QoQ). This is the lower-margin, seasonal-smartphone-driven leg that dragged the Sep-2025 quarter.
• Other — ~¥65B, ~6%.

Customers. Named/known: Apple (Bloomberg repeatedly frames Kioxia as "the Japanese supplier to Apple" ); NVIDIA (deep SSD-architecture collaboration — a joint program targeting 100 million IOPS SSDs vs today's ~3M ceiling ); hyperscalers and enterprise-server OEMs buying the LC9-series 122.88 TB NVMe drives. Contract structure is memory-commodity: short-cycle, ASP-driven, no take-or-pay — except that in the current shortage Kioxia confirmed its entire 2026 NAND output is pre-booked/sold out, a rare demand-visibility window for a commodity maker.

Structure. The manufacturing base is a 50/50 joint venture with Sandisk (spun out of Western Digital in Feb 2025) — see Lens 2. This JV is the defining operational feature of the company.

Supply Chain

Upstream → Kioxia → end customer, named at each node:

• Equipment (the real chokepoint): litho from ASML (and Nikon/Canon for non-EUV layers), deposition/etch from Lam Research, Applied Materials, Tokyo Electron (TEL), metrology from KLA. 3D-NAND layer-stacking is deposition/etch-intensive, so Lam and TEL matter more here than EUV — a structural difference from the leading-edge logic supply chain.
• Wafers & materials: silicon from Shin-Etsu / SUMCO (both Japanese — a home-field advantage), plus specialty gases and photoresist (JSR, TOK).
• Fabrication — the Sandisk JV: the Yokkaichi plant (Mie) and Kitakami plant (Iwate) are the world's largest flash-production sites, run 50/50 with Sandisk. Output is shared; each partner markets its own bit-share independently. The JV agreement was extended through 31 Dec 2034 in Jan 2026 — de-risking the single biggest structural question hanging over the company (a Sandisk exit).
• Fab economics: Fab2 (K2) at Kitakami began operation Sep 2025, capable of 8th-gen 218-layer 3D flash on CBA (CMOS directly Bonded to Array) wafer-bonding. The JV received an up-to-¥150B Japanese government subsidy for Yokkaichi + Kitakami advanced-node capacity — a sovereign-industrial-policy tailwind.
• Packaging/test: in-house + OSAT partners; downstream the bits go into Kioxia-branded SSDs, or as raw NAND/UFS to module makers and directly to Apple/hyperscalers.

Chokepoints & single-source risks: (1) Total dependence on the Sandisk JV — Kioxia does not own 100% of its own production; capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs., output cadence, and technology roadmap are co-decided. (2) Geographic concentration in Japan — Yokkaichi and Kitakami are both in Japan; a single-country earthquake/typhoon exposure (Yokkaichi has taken contamination hits before, historically). (3) Equipment lead-times — adding NAND capacity means 12-18 month Lam/TEL/ASML tool orders, which is why 2026 supply is fixed and cannot flex to demand. This lens is names-complete: ASML, Lam, AMAT, TEL, KLA, Shin-Etsu, SUMCO, Sandisk, Apple, NVIDIA.

Competitive Advantages (moats)

Be honest: NAND is a commodity, and commodities have thin moats. What Kioxia has:

• Oligopoly membership (the real moat). NAND is a five-player rational oligopoly — Samsung (~32%), SK Hynix (~19%), and then a cluster of Kioxia (~15.3%), Micron, Sandisk, YMTC in the 13-15% band. The whole industry has learned discipline: after the 2023 bloodbath, everyone cut output ~15-20% and prices recovered. The moat is not Kioxia-specific; it is that the club has (for now) stopped fighting a share war.
• Process/IP — BiCS + CBA wafer-bonding. Kioxia co-invented 3D NAND (BiCS FLASH) with Toshiba/SanDisk. Rather than win the raw layer-count race (it lags SK Hynix/Samsung on layers), it bets on CBA bonding — building the CMOS logic and memory array on separate wafers and bonding them — for density/speed efficiency. BiCS 10 = 332 layers, +59% areal density, +33% transfer speed vs 218-layer, sampling summer 2026, mass production 2027. Credible but not clearly ahead.
• The NVIDIA / HBF optionality. Kioxia is researching High-Bandwidth Flash (HBF) — stacking up to 16 NAND dies + a logic die to give HBM-like bandwidth at far higher capacity. If AI inference forces a memory tier between HBM and SSD, Kioxia's flash-native HBF could be a genuine new moat. This is the single most interesting non-consensus asset in the story — call it a free option.
• Scale + Japanese cost base + government backing. World's-largest-fab scale, home-country silicon/equipment ecosystem, ¥150B subsidy.

Bargaining power: weak-to-neutral vs customers in normal times (buyers multi-source), temporarily strong now (sold out through 2026, ASPs doubling). Over suppliers (ASML/Lam/TEL): weak — those are the true chokepoints. Verdict on moat: shallow-but-real via oligopoly discipline; the only durable edge would be HBF winning a new memory tier.

Segments

No our figures exists (web-only), so all figures are. Kioxia reports two core segments:

The trend that matters: the mix swing is the P&L. In the Sep-2025 quarter, seasonal smartphone demand pushed lower-margin Smart Devices to ~35% of sales, and that mix shift (not weak demand) is what caused the earnings miss and the -23% crash. As the mix rotated back toward server/PC SSD through Q4 FY25 and into FY26, margins exploded — Q1 FY26 guided OP margin ~74%. Geographic breakout is not disclosed granularly; the demand pull is US hyperscaler AI capex + global smartphone replacement. Whoever models Kioxia is really modeling the enterprise-SSD share of the mix.

---

Phase B — Measure performance

Earnings Result

The FY2025 arc (fiscal year ended 31 Mar 2026) is a textbook memory V-shape — one of the sharpest on record:

Full-year FY2025: revenue ¥2.3376T (~$14.7B), +37% YoY; net profit ¥554.4-554.5B (~$3.48B), roughly doubled YoY from ¥272.3B.

The forward guide is the blow-off top. For Q1 FY26 (Apr-Jun 2026) Kioxia guided net profit ¥869B (a 48x YoY jump), revenue ¥1.75T (~5.1x YoY), operating profit ¥1.298T (~29x YoY, OP margin ~74%) — blowing past the ¥405.6B QUICK consensus by >¥300B. The stated driver: US-dollar NAND ASPs doubled within a single quarter. ROIC guided >60%.

Balance-sheet flags: the up-cycle is repairing the balance sheet fast — Kioxia targets a net-cash position by end of Q1 FY26 and leverage <1x, equity ratio rising toward ~60%. The classic memory-cycle flag — revenue driven by ASP not volume ("significant ASP increase, partially offset by reduced bit shipment") — is explicitly the case here, which is the tell that this is a price cycle, and price cycles reverse. Market reaction to the arc: shares are up ~54x from the ¥1,455 IPO and >660% YTD 2026; the market has fully embraced the up-cycle.

Earnings Calls (sentiment trend)

No transcripts on the shelf; sentiment reconstructed from coverage of the last ~4 prints:

• Aug 2025 (Q1 FY25): defensive/bottoming. Management framed the -74% profit drop as trough, pointed to mix and "slow NAND recovery". Tone: cautious.
• Nov 2025 (Q2 FY25): cautious even amid recovery — flagged the September mix drag, guided nine-month OP ¥229.8-269.8B below the ¥420B street number → -23% crash. The market read management as under-promising into strength; the recurring phrase became "supply tightness" + "sold out."
• Feb 2026 (Q3 FY25): decisively bullish — record revenue, "entire 2026 capacity locked in advance," ASP-led upgrade. Stock +19.6%.
• May 2026 (Q4 FY25 / FY26 guide): euphoric — the 48x profit guide, ADS-listing announcement, "era of cheap 1TB SSDs is over," capacity "sold out through 2027".

The shift over time: from "we are bottoming, please be patient" (Aug) → "tight and sold out, but we won't over-guide" (Nov, which the market hated) → "records as far as we can see" (Feb-May). The thing management stopped saying is anything about production cuts (a 2024 refrain); the thing they started saying is "sold out through 2026/2027." Tone is now maximally optimistic — which, for a commodity name, is itself a late-cycle sentiment signal.

Comps

Kioxia's cleanest peer is Micron (MU) — the only other name with a real, freshly-sourced multiple. Samsung and SK Hynix are conglomerates where NAND is a slice; Sandisk is the JV partner. Multiples are or n/a.

Reading it: both pure-cyclicals (Kioxia ~7.9x fwd, Micron ~6.8x fwd) trade at low headline multiples — the classic "cheap at peak earnings" trap. A single-digit P/E on a cyclical near a cycle top is a warning, not a bargain: the E is inflated. Kioxia sits at a ~20% discount to SK Hynix on 2027E, which is defensible given SK Hynix's DRAM/HBM diversification and superior NVIDIA HBM position. Dividend yield and 5-yr avg ROE: n/a (Kioxia has <2 years of public history). The comp table's honest message: Kioxia is not mispriced relative to peers; it is priced as a peak-cycle pure-play, and the debate is entirely about where we are in the cycle.

Stock-Price Catalysts

With <2 years of listed history the "5-year" window compresses, but the moves are violent and revealing:

• 18 Dec 2024 — IPO at ¥1,455; muted debut (~¥886B cap), "barely-noticed".
• Through 2025 into 2026 — the melt-up: +660% YTD 2026, ~54x from IPO, on the AI-NAND-shortage narrative. 52-week range ¥2,270 → ¥112,700 — a ~50x intraday-year range, which by itself tells you the volatility regime.
• 14 Nov 2025 — -23% in a day on a Q2 guide below the ¥420B street number.
• Feb 2026 — +19.6% on the record Q3 + "sold out 2026".
• Nov 2025-Feb 2026 — Bain selldown shocks: each Bain block sale (~$2.1B Nov, ~$3.5B Feb) pressured the price on supply/overhang.
• 15 May 2026 — US ADS-listing announcement (investor-base expansion).

What the tape reacts to, ranked: (1) NAND ASP direction / supply-demand guidance — the dominant driver; (2) the Bain overhang — mechanical selling pressure independent of fundamentals; (3) AI-demand headlines (NVIDIA collab, hyperscaler SSD). It does not trade on smartphone units or company-specific execution — it trades as a high-beta option on the NAND cycle plus a PE-unwind clock. That is the single most important behavioral fact for sizing a position.

---

Phase C — Judge people & books

Management

• Track record / transition. Nobuo Hayasaka (70) led as President & CEO from 2020, steering Kioxia through the 2023-24 trough and the Dec-2024 IPO. He steps down 31 Mar 2026; Hiroo Oota (63) becomes President & CEO 1 Apr 2026, with Hayasaka moving to senior executive adviser. Oota is a 40-year Toshiba/Toshiba-Memory lifer (joined 1985), an engineering/semiconductor insider — continuity, not disruption. Fresh CEO into a euphoric tape is a timing watch-item.
• Skin in the game. The controlling economics sit with Bain Capital, not management — and Bain is selling (below). Employee alignment is real, though: ~600 Kioxia employees became (yen-)billionaires on stock options in the run-up — a genuine culture-and-retention positive, and a mild sell-pressure risk as options vest.
• Capital allocation. The defining decisions are capex and the JV. FY26-28 capex guided to ~¥470B/yr avg (+66% vs FY25), R&D to ~¥230B/yr, with a third Kitakami fab and M&A under consideration. This is pro-cyclical capex into a peak — the eternal memory-maker sin. Discipline signal: they extended the Sandisk JV to 2034 (good) and target net-cash + leverage <1x (good). ROIC guided >60% at peak is not a through-cycle number.
• Red flags (governance). The Bain-controlled, PE-exit structure is the central governance overhang — decisions are influenced by a seller optimizing its own exit, not by permanent owners. The board blends Bain + Toshiba + independents. No related-party-abuse or promotional-accounting flags surfaced.
• Archetype. Professional-manager / engineering-continuity run, sitting under a PE owner in wind-down. Appropriate for an operationally mature commodity maker; the risk is that ownership is transitional and capital allocation is being made near a top.

Forensic Red Flags

Web-only; every figure. No filings to audit line-by-line, so this is a risk map, not a reconciliation.

• The commodity-accounting core risk: inventory. In memory, the swing account is inventory valuation. In the 2023 downturn, gross margins went single-digit/negative industry-wide; Kioxia (and peers) took write-downs. Right now inventory across NAND suppliers has normalized to 8-10 weeks vs >20 at the 2023 peak — healthy today, but the account that would flip fastest in a reversal. Watch it.
• ASP-not-volume revenue quality. Management states revenue growth is ASP-driven with falling bit shipments. Price-led revenue is lower-quality/less-durable than volume-led — a structural yellow flag inherent to the cycle, not misconduct.
• JV accounting complexity. The 50/50 Sandisk JV means capex, output, and some costs are shared/equity-method-adjacent — genuine analytical opacity in cost attribution and true owned-capacity. Not a fraud flag; a transparency discount.
• Leverage repaired but recent. BB+ (speculative) with a $2.2B bond stack (5yr @ 6.25%, 8yr @ 6.625%). The high coupons price the cyclicality; the fast de-leveraging is real but only because of peak cash flow.
• SBC / dilution: the ~600-employee options windfall implies meaningful stock-based comp and future dilutionDilutionIssuing new shares, so each existing share owns a smaller slice of the same company. as options exercise — quantify from filings when the ADS F-1 lands.

Regulatory findings (required sub-section). Read regulatory/regulatory-findings.md (Step 0):

• SEC EDGAR (LR + AAER): 0 findings — Kioxia has no CIK and files nothing with the SEC, so no EDGAR enforcement search is possible.
• Non-SEC enforcement (web search "Kioxia Holdings" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR fine OR penalty) enforcement): no material enforcement action, consent decree, or antitrust penalty against Kioxia surfaced in coverage. (Historical industry note: DRAM/NAND price-fixing probes have periodically touched the memory sector broadly; nothing specific and material to Kioxia found.)
• Item 3 / Legal Proceedings: n/a — no 10-K exists (no EDGAR filer). The forthcoming US ADS F-1 will be the first SEC-grade legal-proceedings disclosure; treat its Risk Factors as a priority read when it lands.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER, all zero by construction), web search, and the absence of a 10-K, as of 2026-07-06.

---

Phase D — Project & stress-test

Forward Projection

No our figures, no share count on the shelf, and Kioxia reports in yen with a fiscal year ending March — so EPS-per-share modeling from primaries is not possible web-only. I therefore project operating profit / net profit trajectory (the metrics the company and Goldman actually guide) rather than fabricate a per-share EPS. Every input labeled; output.

Anchors (all ): FY2025 (ended Mar-2026) net profit ¥554.4B, revenue ¥2.3376T. Q1 FY26 guide: revenue ¥1.75T, OP ¥1.298T, net profit ¥869B — i.e. a single quarter guided above the entire prior fiscal year's profit. Goldman models FY2028 consolidated OP toward ¥10T with ~80% gross margin.

FY2026 (ending Mar-2027) — profit path:

• Base: annualizing the ¥869B Q1 guide is naïve (cycles peak mid-year), but even a decelerating H2 puts FY26 net profit in the ¥2.5-3.5T range. Revenue ~¥5-6T. This is a ~5-6x profit jump on the ASP doubling.
• Bull: shortage persists all year, ASPs hold, HBF/enterprise-SSD mix richens → FY26 net profit ¥3.5T+, revenue ¥7T, tracking toward the Goldman FY28 ¥10T-OP path.
• Bear: ASPs are a price spike, not a plateau; capacity (industry-wide, incl. Kioxia's +66% capex and 3rd Kitakami fab) floods 2027; mix reverts on a smartphone-heavy quarter → H2 FY26 decelerates hard, FY27 profit halves or worse as the cycle rolls. In the last down-leg NAND revenue fell -39.9% in a single year and Kioxia posted outright losses.

The honest projection: FY26 is nearly locked to be a record (2026 output sold out). FY27-28 is a coin-flip on the cycle — the same +66% capex that funds the bull case builds the bear case's oversupply. Kioxia has no DRAM/HBM cushion, so its through-cycle earnings power is structurally more volatile than Micron's or SK Hynix's — it will out-earn them at the top and under-earn (or lose) at the bottom.

Forecast log: skipped per --watchlist rules (no our model create in the unattended loop; this is not a committed base case worth Brier-scoring). If promoted to a thesis, the scoreable binary would be "Kioxia FY2027 (ending Mar-2027) reports a YoY operating-profit decline" — the cycle-turn bet.

Bull vs Bear

Bull case. The AI data-center build-out has structurally re-rated NAND from a PC/phone commodity to an AI-infrastructure input. Enterprise SSD is exploding (Q4 FY25 SSD segment +99.8% QoQ), 2026 output is sold out, ASPs doubled in a quarter, and the industry is capacity-constrained because tool lead-times mean supply can't flex for 12-18 months. Kioxia is the highest-beta pure-play on exactly this — no diversification means maximum torque. Add the HBF option (a flash tier between HBM and SSD for AI inference), the NVIDIA 100M-IOPS collaboration, a fast-repairing balance sheet (net-cash target, leverage <1x), Japanese sovereign backing (¥150B subsidy), and a ~7.9x forward multiple at a ~20% discount to SK Hynix. If "the era of cheap SSDs is over" is a regime not a spike, the stock is early.

Bear case (2-3 permanent-impairment risks). (1) It's a commodity cycle, and cycles turn. Revenue is ASP-driven on falling bits — the definition of a price spike. Kioxia's own +66% capex + a possible 3rd Kitakami fab + industry-wide expansion is the oversupply seed for 2027-28; when it hits, a pure-play with no DRAM has nothing to cushion the collapse, exactly as in 2023 (losses, -70% contract prices). (2) The Bain overhang is a mechanical share-price weight — Bain has cut from ~51% to <30% and is still selling into strength; every block sale caps the stock regardless of fundamentals. (3) Single-digit P/E on peak earnings is a trap, not a discount — the E is inflated; on normalized through-cycle earnings the multiple is far higher.

Pre-mortem (18 months out, thesis broke). It's early 2028. NAND ASPs peaked in mid-2026 and rolled over as Kioxia's, Micron's, and Samsung's new capacity (all greenlit at the top) came online into a decelerating AI-capex digestion phase. Kioxia's mix swung smartphone-heavy in a soft consumer year; the pure-play that had the most torque up now has the most torque down and prints a loss quarter. Bain finishes exiting into the weakness. The stock, which round-tripped a 54x move, gives most of it back. Everyone knew it was a cycle; the only question was the month it turned, and the +66% capex told you they'd overbuild.

Are multiples too high? Headline no (~8x); cycle-adjusted yes — this is peak-earnings pricing. Contrarian view of what the market is refusing to see: the bulls are treating a price cycle (ASPs doubling on flat/down volume) as a secular demand regime, and are ignoring that the same capex funding today's bull thesis is manufacturing 2027's oversupply. The market is also under-weighting that Kioxia, uniquely among the majors, has zero diversification to soften the turn.

Devil's Advocate (short-seller)

Dismantling the bull case:

• What structurally breaks the money machine: NAND ASP reversal. Full stop. ~85%+ of the profit swing is ASP, and ASPs are set by industry supply/demand, not by Kioxia. In 2023 flash revenue fell -39.9% in one year; a repeat turns ¥869B/qtr profit into losses, because Kioxia has no DRAM or HBM to offset — the exact weakness that forced two IPO delays.
• Revenue concentration: exposed to (a) hyperscaler AI-capex durability — if AI-server SSD orders digest/pause, the high-margin leg evaporates; (b) Apple/smartphone seasonality — the low-margin leg that already caused the -23% crash; (c) the Sandisk JV — Kioxia doesn't fully control its own capacity or roadmap.
• Why the moat is weaker than bulls think: it's oligopoly discipline, not a company moat — and oligopoly discipline is exactly what breaks in a downturn when someone (historically Samsung) fights for share. Kioxia lags on raw layer count; its CBA/HBF bets are promising but unproven at scale, and Samsung/SK Hynix have deeper R&D pockets.
• Most dangerous competitor bulls underestimate: SK Hynix (already a 14% Kioxia shareholder via converts) — it has HBM leadership, DRAM diversification, the best NVIDIA relationship, and could out-invest Kioxia through a downturn. Also YMTC (China), subsidized and share-hungry, a structural price-floor threat.
• Worst capital-allocation setup: pro-cyclical +66% capex at the top under a PE owner exiting — capital decisions made by a seller, near a peak, in a commodity. That is the textbook value-destruction configuration.
• What must hold for today's price: ASPs stay elevated through FY27, AI-SSD demand doesn't digest, the industry stays disciplined despite everyone expanding, and Bain's remaining ~30% clears without cratering the stock. Four things, all outside Kioxia's control.
• Valuation if growth disappoints 20-30%: a 20-30% revenue miss in a fixed-cost fab business is a >50% profit miss (operating leverage cuts both ways), which re-rates the "cheap 8x" to an expensive 20x+ on the lower E — and the stock, being a high-beta option, would fall far more than the earnings.
• Single permanent-impairment scenario: a 2027-28 oversupply glut (AI digestion + industry capacity flood) that pushes NAND back to cash-cost while Kioxia carries elevated capex commitments — plausibility moderate-to-high on a 2-3yr view; this is not a tail, it's the base rate of the NAND cycle.

Management Questions (ordered by information value)

1. What is your through-cycle operating-margin and ROIC assumption — not peak — and at what NAND ASP level does Kioxia go operating-loss-making?
2. Your FY26-28 capex is +66%; a third Kitakami fab is under study. What is the industry-wide bit-supply growth those investments imply for 2027-28, and how do you avoid recreating the 2023 glut you're barely out of?
3. How much of current revenue growth is ASP vs bit-volume, and what is your bit-shipment growth guidance for FY26 specifically?
4. Bain is below 30% and selling. What is the expected timeline and mechanism for the remaining PE exit, and what governance changes follow full exit?
5. Quantify the HBF opportunity: TAM, timeline to revenue, capex required, and where it sits between HBM and SSD — is this a real new profit pool or a research line item?
6. What share of SSD & Storage revenue is AI/hyperscaler data-center vs traditional enterprise/PC, and how concentrated is it among your top 3 customers?
7. On the Sandisk JV: how are capex, output allocation, and technology decisions governed 50/50, and what is your true owned wafer capacity vs marketed bit-share?
8. Where are you on the layer-count roadmap vs Samsung/SK Hynix, and does CBA bonding close, hold, or widen that gap through BiCS 10/11?
9. What is your plan for the balance sheet at the next trough — how much net cash do you carry into a downturn given the 6.25-6.625% bond coupons?
10. How exposed is the mix to smartphone seasonality, and what structurally reduces the margin volatility that caused the Nov-2025 miss?
11. The US ADS listing: timing, size, primary vs secondary, and does it create additional Bain/Toshiba exit supply?
12. Post-Hayasaka, what changes under CEO Oota — strategy continuity or a shift in capital discipline / M&A appetite?
13. What is your M&A thesis — you're "considering" it at a cycle peak with a rich currency; what would you buy and why now?
14. What is Kioxia's answer to YMTC and Chinese subsidized NAND as a structural price-floor threat?
15. If AI-server SSD demand digests for two quarters in 2027, what is your production-cut and cash-preservation playbook — and how fast can you cut given fixed fab costs?

---