Earnings call transcript: Kokusai Electric Q4 2026 results exceed forecasts, stock drops

Kokusai Electric beat revised FY2026 forecasts with revenue ahead by JPY 5.1 billion, adjusted operating income ahead by JPY 3.2 billion, and adjusted net income ahead by JPY 2.3 billion, but shares fell 12.83% after investors focused on a 2% revenue decline and an 18% drop in adjusted operating income. Service revenue rose 27% year over year to 40% of sales, bookings exceeded expectations by about JPY 30 billion, and the company reached net cash a year early. FY2027 guidance calls for revenue growth of 19%, but management flagged ongoing China weakness, higher CapEx, and about JPY 8 billion of additional SG&A.

Analysis

The key signal is not the headline beat; it is the mix shift toward service/retrofit demand that is masking a cyclical equipment pause. That usually matters more for the ecosystem than the company itself: when customers stretch existing tools via upgrades instead of greenfield orders, it supports installed-base profitability but delays broader WFE recovery and keeps multiplier effects on upstream consumables and component suppliers muted for 1-2 quarters.

For NVDA, the China-visit optics are a near-term sentiment tailwind, but the more durable read-through is that advanced-node spending is still being pulled forward by AI rather than memory normalization. If this company’s non-China demand is accelerating while China remains soft, it implies AI capex is becoming increasingly bifurcated: compute leaders keep spending, while legacy and geopolitically constrained markets stay constrained. That dynamic should support AI-exposed semiconductor equipment and foundry names relative to China-dependent memory proxies.

The market appears to be punishing reported conservatism more than fundamentals. With bookings running ahead of plan and management explicitly describing customer schedules moving earlier, the setup suggests second-half estimates may be the right place for upward revision, not the first half. The overreaction is likely in the stock, but the rerating will be capped unless investors see evidence that the service mix can convert into higher operating leverage rather than just revenue smoothing.

Contrarian angle: consensus is likely underestimating how much of the 2027 uplift is already visible in order behavior, while overestimating how much China weakness is secular. If lead times are already stretching and capacity is the constraint, even modest pickup in China NAND/DRAM could force a fast revision in both top line and margin assumptions. The risk is that the current valuation still prices in a cleaner margin rebound than the company itself is guiding, so timing matters more than direction.

Key Decisions for Investors

• Add LRCX on 3-6 month weakness versus the group: this print supports a relative long in service/installed-base exposure and AI-driven WFE upside; target a 10-15% rerating if next quarter order commentary confirms early pull-ins.
• Stay long NVDA on any dip tied to China policy noise: the read-through is still positive for AI capex breadth, and the risk/reward is better via the dominant compute beneficiary than through equipment laggards; use 1-2 month horizon for event-driven strength.
• Short a China-memory-sensitive equipment basket against LRCX/NVDA-related AI exposure: the article reinforces that China remains the slowest-growing end market, making the pair attractive if AI capex keeps accelerating; look for 5-8% relative underperformance over the next quarter.
• Consider buying out-of-the-money calls on LRCX or a semicap ETF into the next earnings cycle: bookings momentum plus earlier customer timing creates upside asymmetry if guidance gets revised upward; cap risk by limiting premium to a small notional.
• Avoid chasing the standalone move in the Japanese name after the selloff; the better trade is to fade overdone pessimism only if follow-up channel checks confirm H2 pull-ins and no further China deterioration.