Kioxia Q2 FY2026 slides: record revenue tops full FY2025, stock falls

Kioxia reported Q2 FY2026 revenue of JPY 2,480.0 billion, up 40.3% sequentially and 453.2% year over year, exceeding its full FY2025 revenue; non-GAAP operating profit was JPY 1,988.3 billion at an 80.2% margin. The company guided for Q3 revenue of JPY 2,500.0 billion, just 1% higher sequentially, while citing a JPY 55 billion FX headwind; CY2027 production capacity is fully allocated and more than 50% of CY2028 gigabyte shipments are covered by long-term agreements. Kioxia completed an JPY 800.0 billion buyback and announced an interim dividend of JPY 90 per share pre-split, but shares fell 7.14% to $13,470 in after-hours trading.

Analysis

The key question is whether the earnings step-up reflects durable scarcity pricing or a short-lived price reset. If verified, flat Q3 bit volumes alongside roughly flat sequential revenue imply little near-term unit growth; the Q4 acceleration therefore depends on negotiated pricing and delayed data-center shipments, not simply AI demand. That makes the margin outlook unusually exposed to price reversals. CY2027 allocation and CY2028 contracts improve visibility, but do not establish contract pricing, take-or-pay protection, or returns on the planned capacity build. A rapid K3 expansion could eventually turn today’s scarcity into oversupply; slower smartphone and PC demand adds a potential outlet for displaced supply. High-capacity QLC adoption could also pressure HDD suppliers over time, but substitution economics and workload mix need confirmation.

More importantly, the reported scale and margins are sufficiently extraordinary that we would not trade the figures at face value. Verify Kioxia’s primary filing/presentation, units, accounting definitions, share-price currency, and the buyback/dividend details before inferring earnings power or valuation. If confirmed, the after-hours decline may reflect a reset from peak-cycle expectations: a 1% Q3 growth guide leaves little cushion if negotiated prices weaken, while yen moves are a meaningful near-term sensitivity. The thesis fails if subsequent disclosed ASPs and operating profit hold or rise while Q4 shipments convert as indicated; it strengthens negatively with price concessions, weaker guidance, or evidence that capacity additions outrun demand.

Key Decisions for Investors

• No directional position on the reported results until they are reconciled to Kioxia’s primary disclosure; the unusually high stated revenue and operating margin make source/units verification a gating item, not a footnote.
• If confirmed, treat the next 1–3 months as a pricing-and-conversion test: monitor Q4 data-center shipment realization, ASPs on the unpriced Q3 volume, and subsequent guidance. Do not annualize the reported peak-quarter margin.
• Watch for a conditional relative-value expression rather than a standalone NAND bet: if Kioxia’s pricing/shipments weaken while peers’ disclosures remain resilient, consider underweighting Kioxia versus better-supported memory exposure; if pricing remains firm and contracts prove protective, reassess rather than fade the sector automatically.
• Revisit the longer-term supply risk as Y7/K2 utilization and K3 investment plans become clearer. Evidence of capacity growth outrunning market bit demand, or weaker PC/smartphone absorption, would invalidate the scarcity-pricing thesis and raise downside risk.