ON Semiconductor: Overbought Ahead Of Slow Fundamental Recovery, Downgrade To Hold

ON Semiconductor’s AI data center revenue may double in 2026, but it remains a small share of total revenue and meaningful design-win-driven upside is expected mostly from 2027 onward. The article describes FQ1'26 performance and FQ2'26 guidance as underwhelming, while warning that the recent rally has run ahead of fundamentals. The stock’s premium P/E of 35.1x is seen as already pricing in much of the AI growth story, leaving limited margin of safety.

Analysis

The key market mistake is treating AI as a near-term re-rating story rather than a deferred earnings stream. If the incremental design-win revenue only becomes material in 2027+, then the current multiple is effectively discounting a cleaner margin mix and much faster AI monetization than the business can actually deliver over the next 2-4 quarters. That leaves ON exposed to a classic “story ahead of numbers” de-rate if guidance continues to underwhelm while the stock has already pulled forward several quarters of optimism.

Second-order, the likely winners are the customers and peers with faster AI-cycle exposure and/or better operating leverage. If ON’s AI contribution is still small in 2026, then the supply-chain beneficiaries are probably elsewhere in the semiconductor stack where AI content is already visible in revenue and gross margin inflection. On the other side, analog and power names with similar valuation anchors may see a sympathy reset if investors start repricing how much AI optionality is actually worth before it shows up in reported sales.

The timing matters: the next catalyst window is not a product-cycle headline, it’s whether FQ2/FQ3 execution validates any acceleration in backlog, mix, or margins. If that evidence does not show up, the stock can underperform for months even without a hard fundamental break, because positioning looks crowded and the premium multiple limits support. The contrarian case is that consensus may be underestimating how quickly a single large design win can change the 2027 setup, but that is a longer-duration call and not a good reason to pay peak-like multiples today.

Key Decisions for Investors

• Short ON into strength or via a tactical call spread hedge over the next 4-8 weeks; thesis is multiple compression if guidance remains soft and AI revenue stays too small to alter FY26/FY27 numbers.
• Pair trade: long a more directly monetizing AI semiconductor name vs short ON for 3-6 months; seek relative performance from names with near-term AI revenue translation and lower dependence on 2027 optionality.
• If holding ON longs, buy downside protection into the next earnings/guidance event; preferred structure is a 1-2 quarter put spread to cap premium outlay while preserving convexity on a de-rate.
• Do not chase the rally until there is evidence of AI mix inflection in reported revenue or margins; use any post-earnings bounce to reduce exposure rather than add.
• For contrarians, only consider a small starter long after a 15-20% pullback and a stabilizing guide, because the current setup offers limited margin of safety at a 35x-type multiple.