Mizuho raises ON Semiconductor stock price target on AI growth By Investing.com

Mizuho lifted its ON Semiconductor price target to $120 from $70 and kept an Outperform rating, citing stronger fundamentals and growth prospects. The company said AI data center revenue should rise 100% year over year in calendar 2026 to more than $500 million, while March-quarter EPS of $0.64 beat the $0.61 estimate and revenue of $1.51 billion topped the $1.49 billion consensus. Gross margin expansion, higher utilization, and 2027 earnings growth expectations support the bullish outlook.

Analysis

The market is signaling a rotation from a single-vendor foundry narrative toward a more fragmented AI supply chain, and that matters more for valuation dispersion than for near-term unit volumes. The immediate beneficiary is less the obvious underdog and more any supplier with credible advanced-node capacity, because even a small diversification push from a hyperscaler can re-rate “optional capacity” ahead of realized revenue. That creates a second-order read-through for Intel: the stock can trade on strategic relevance before it proves durable share capture, but execution evidence will still be required over the next 6-18 months.

For ON, the key point is not simply that growth is improving, but that the earnings mix is shifting toward higher-quality, higher-margin content just as utilization normalizes. That combination tends to compress the market’s skepticism quickly: when a cyclical semiconductor name moves from operating deleveraging to margin expansion, the multiple can expand faster than consensus models assume. The risk, however, is that the market is already capitalizing too much of the 2027 margin story today; if auto remains flat and AI ramps are back-end loaded, the stock can stall even with good quarters.

TSM’s near-term risk is not lost share so much as narrative pressure: any customer diversification headlines create an overhang on multiple even if actual wafer allocations barely change. Apple-related diversification talk is especially important because it forces investors to handicap dual-sourcing economics, which usually favor bargaining power over outright displacement. The contrarian view is that the move may be overdone in both directions: Intel gets credit for strategic progress too early, while TSM may be penalized for a supplier-happiness issue rather than a demand issue.

Key Decisions for Investors

• Long ON / short SOXX into the next 4-8 weeks: ON has the clearest path to upward estimate revisions from margin leverage and AI mix, while the basket is more exposed to names where the diversification narrative is a valuation headwind. Use a 2:1 upside/downside structure and trim if ON approaches analyst-day expectations too quickly.
• Add a tactical long INTC position only on dips over the next 1-3 months: the stock can continue to re-rate on strategic relevance, but treat it as a headline-driven trade, not a fundamentals conviction buy. Use tight risk controls; if no concrete design-win evidence emerges, the trade loses relevance quickly.
• Short-dated put spread on TSM for 1-3 months: the point is not a fundamental short, but to monetize multiple compression from customer diversification headlines. Best risk/reward if the implied move is still cheap relative to headline sensitivity.
• Pair long AAPL / short TSM only if new sourcing commentary becomes more explicit: this is a hedge against Apple signaling supplier optionality, which improves procurement leverage and can modestly pressure TSM sentiment. Keep sizing small because the second-order impact is more on terms than on shipment volumes.
• Hold ON through the next guidance cycle, but consider taking partial profits into strength: the stock has already priced a good portion of the 2027 margin expansion story, so the cleaner opportunity is on pullbacks or around confirmation events rather than chasing at fresh highs.