This AI Storage Stock is Up 560% so Far This Year. Is it Too Late to Buy? (Hint: It's Not Sandisk)

Kioxia’s flash memory sales are surging on AI data-center demand, with revenue up 59% in fiscal 2025 and 37% in fiscal 2026, while net income more than doubled in 2026. Analysts expect revenue and net income CAGRs of 118% and 232% from fiscal 2026 to fiscal 2028, and the stock still trades at just 8.5x this year’s earnings. The article argues the shares remain cheap despite a 560% YTD rally, with a potential U.S. ADR IPO also highlighted as a catalyst.

Analysis

The market is starting to re-rate memory as an AI bottleneck business, but the bigger second-order effect is that supply discipline is now the real earnings lever. With a concentrated manufacturing footprint and an extended JV structure, incremental volume is less important than whether the industry keeps rational pricing into the next capacity cycle; if it does, operating leverage will stay unusually high for longer than the market expects. That makes the current setup more like a constrained infrastructure oligopoly than a classic cyclical semiconductor trade.

The near-term upside is not just multiple expansion, but a continued mix shift toward higher-spec data center and enterprise products, which should keep margins ahead of consensus even if unit growth moderates. The hidden beneficiary is Western Digital via the JV economics: as access-fee economics and production utilization improve, WDC’s exposure is a cleaner, lower-volatility way to express the same AI memory thesis with less single-name execution risk. By contrast, any lagging smartphone/PC exposure becomes the weak link if pricing rolls over, because those segments tend to absorb the first margin compression when supply loosens.

The main risk is timing mismatch: the market is pricing a multi-year AI supercycle, while the next 1-2 quarters will likely be dominated by shipment volatility, customer inventory normalization, and IPO-related supply. If the upcoming listing increases float without a corresponding lockup discipline, the OTC rally can retrace hard even if fundamentals remain intact. Consensus may be underestimating how quickly sentiment can reverse in memory when investors shift from scarcity narrative to capacity-announcement narrative.

This is a good place to prefer relative-value exposure over outright momentum chasing. The asymmetry is attractive if one can get paid to wait for post-IPO volatility or a broad tech pullback, but at current levels the better trade is to own the structurally advantaged operator against a more diversified but less leveraged peer. The market is probably not wrong on direction, but it may be too aggressive on near-term magnitude and too complacent about financing, float, and inventory-cycle risk.

Key Decisions for Investors

• Long Kioxia only on a post-IPO pullback or first 2-4 weeks after listing; target entry after float stabilizes, with a 6-12 month hold. Upside comes from continued AI mix shift and scarcity pricing; risk is a sharp de-rating if lockup supply and sentiment hit together.
• Pair trade: long SNDK / short WDC over the next 3-6 months. SNDK has the cleaner direct AI memory leverage and stronger momentum, while WDC offers a more diluted exposure set; this isolates the memory upcycle with less macro beta.
• Sell downside in the form of cash-secured puts on SNDK or Kioxia into any post-earnings strength. Premium should remain rich as the market prices elevated volatility; this is attractive if you’re willing to own on a 10-15% drawdown.
• If already long Kioxia OTC, use the expected IPO window to trim 25-50% and re-enter on the first post-listing liquidity flush. The risk/reward improves materially once the new issue clears price discovery.
• Avoid chasing at current levels if memory equities add another 15-20% before a formal listing. At that point, the setup shifts from fundamental re-rating to flow-driven momentum, where drawdowns can be abrupt.