ON Semiconductor’s results, guidance beat estimates but stock slips after hours

ON Semiconductor beat Q1 2026 expectations with EPS of 64 cents on revenue of $1.51 billion versus consensus of 61 cents and $1.49 billion, respectively. Q2 guidance also topped estimates, with adjusted EPS of 65-77 cents and revenue of $1.535 billion-$1.635 billion above the midpoint consensus. Despite the strong report and improving AI data center demand, the stock fell about 4% after hours, likely reflecting its roughly 89% YTD run-up.

Analysis

ON’s print is less about the modest beat and more about evidence that the cycle bottom in industrial/auto semis is becoming a real inventory restock, not just a one-quarter pause. The mix matters: power is accelerating while mixed-signal lags, which suggests customers are resuming orders first where end-demand is most visible and deferrable inventories are thinnest. That usually precedes a broader recovery in 1-2 quarters, but it also means multiple expansion can outrun fundamental revision if investors front-run a cleaner 2026 setup.

The second-order winner is the SiC ecosystem, not just ON. If AI data center power demand is compounding from a small base, the market may be underestimating how much of that growth comes from higher-voltage power conversion and thermal management rather than pure GPU unit growth; that benefits adjacent power module, discrete, and materials suppliers with similar exposure. The main loser is any supplier whose thesis depended on a slow AI capex digestion story — the hyperscaler re-commitment makes that thesis harder to sustain, and ON’s guidance reinforces that AI spend is still flowing into power infrastructure even if some software/LLM names wobble.

The stock reaction looks more like positioning than fundamentals. After an ~89% YTD run, a beat had to be exceptional to justify another leg higher, so the post-print fade is a signal that good news is increasingly fully owned. That said, the risk/reward improves on pullbacks because the guidance midpoint still exceeds sell-side expectations while the business appears to be exiting the trough; the key risk is that auto/industrial normalization stalls again before volume leverage shows up in margins.

Contrarian view: the market may be over-indexing on the AI narrative and underpricing the recovery in the core industrial/auto base. If AI decelerates from 30%+ sequential growth to merely strong growth, the multiple could compress; but if industrial turns while AI holds, earnings revision breadth could expand materially over the next 2-3 quarters. In that setup, the right trade is not chasing the gap, but buying dips on proof of sustained order momentum.

Key Decisions for Investors

• Buy ON on a 3-5% pullback over the next 1-2 weeks; use a 2-3 quarter horizon. Risk/reward is favorable if the market is simply digesting a crowded long, with upside from earnings revisions and downside limited unless auto/industrial re-rolls lower.
• Pair trade: long ON / short a slower-moving industrial semiconductor peer with weaker AI power exposure over the next 1-2 quarters. The thesis is that ON gets both trough recovery and AI power re-rating, while the short leg only gets the cyclical beta.
• Initiate a small tactical long in the SiC/power ecosystem via ON or a basket proxy on any weakness; target a 6-12 month window. The catalyst is continued AI data-center power intensity and EV/industrial electrification, with asymmetric upside if power demand broadens beyond hyperscalers.
• Use options to express upside without paying full multiple: sell cash-secured puts or put spreads 1-2 months out on ON after the post-earnings volatility settles. This monetizes elevated implied vol while positioning for a drift higher if guidance is confirmed next quarter.
• Avoid chasing the AI capex basket purely on headline spend announcements; prefer names where power infrastructure is already showing through in bookings/revenue. The risk is that software/compute beneficiaries lag while power and thermal suppliers capture the nearer-term spend.