A particular kind of unease settles over a stock when the price and the fundamentals are running in opposite directions — not violently, not in a way that forces a conversation, but quietly enough that most participants seem content to look away. ON Semiconductor is sitting in that space right now. Shares have climbed from $58.55 in mid-March 2026 to $68.65 by mid-April, a move that feels less like a re-rating and more like a tide lifting everything it touches.
This entire article is about one variable: operating margin. Not revenue recovery, not AI adoption curves, not analyst ratings distribution. Just whether ON Semiconductor’s operating income — which collapsed from $1.77 billion in 2024 to $84.2 million in 2025 — can credibly begin reconstructing itself over the next twelve months, and whether the current price gives that reconstruction enough room to matter.
The Margin That Needs Explaining
Start with the number itself. An operating margin of approximately 1.4% on $5.99 billion in revenue is a structural signal that something in the cost base didn’t flex the way management or analysts expected when volumes softened. The 2024 figure was 25.0% — a gap so wide it describes a fundamentally different operating reality. And yet the stock hasn’t been priced like a company in existential trouble, which suggests the market’s working assumption is that the 1.4% belongs to a trough that has already passed, that the business underneath is intact, waiting for demand to return.
That assumption might be right. Capex-to-revenue dropped from 9.8% in 2024 to 5.7% in 2025, which on its face looks like retreat — management pulling back investment during a downturn. But read alongside R&D intensity actually rising, from 8.7% of revenue to 9.7%, a different picture starts to form: the company reduced physical capital outlay while protecting the engineering spend that generates its product differentiation in power management ICs. That’s a defensible prioritization, and it suggests the compression in operating income reflects demand-side pressure more than a structural erosion of the cost model — though that framing grows harder to defend if the inventory correction runs longer than the market currently expects.
The semiconductor inventory cycle for power management ICs — components deeply embedded in industrial and automotive applications, both of which have been soft — doesn’t move on the same schedule as the AI narrative currently driving sector sentiment. Inventory normalization in those end markets has historically lagged by one to three quarters relative to the broader cycle, and if the glut in those channels persists into late 2026, pricing concessions could suppress margin recovery even as nominal revenue begins to stabilize. The stock price, sitting at $68.65, doesn’t appear to be pricing that risk with much seriousness.
It may be right not to.
What the R&D Spending Protects on the Other Side
ON’s differentiation in power management for AI infrastructure — specifically the silicon carbide and intelligent power modules that sit at the interface of high-current density and thermal efficiency — is the reason the bull case holds together at all. These are components where product-level pricing power persists across cycles, and the decision to maintain R&D at 9.7% of revenue through the downturn is essentially a bet that such pricing power will be available on the other side of the inventory correction. If that bet pays off, the operating leverage embedded in a business that has already compressed costs significantly could translate into margin expansion that surprises on speed, not just magnitude.
Over the next 12 months, operating margin recovery toward the 10–15% range is more likely than a revenue-led re-rating, provided industrial and automotive channel inventory clears by mid-year 2026 — if that clearance stalls past Q3 2026, the current valuation at $68.65 becomes difficult to justify on any forward earnings basis.
The analyst community’s posture — 20 Buy or Strong Buy ratings against 24 Holds, with no Sells as of April 2026 — is consistent with this reading. It’s a community that believes the floor has been found but isn’t willing to call the ceiling. The absence of Sell ratings in particular suggests a degree of conviction that the 1.4% margin is a cyclical artifact rather than a structural destination.
The counter-scenario deserves serious attention. If automotive electrification demand — one of ON’s core growth narratives — continues to disappoint relative to build plans, and if industrial capex in Europe and North America remains suppressed by rate-driven caution, then the recovery in power management IC volumes may be shallower and longer than the current stock move implies. A margin recovery that reaches only 6–8% by late 2026, rather than the 12–15% range the bull case requires, would leave the stock roughly fairly valued at current prices — without the upside that makes the position interesting. This breaks if auto and industrial destocking extends two or more quarters beyond current consensus expectations, because the cost compression that makes the operating leverage story compelling has a shelf life — at some point, sustained low utilization starts eroding the fixed-cost base itself rather than coiling it for recovery.
The $84.2 million operating income figure deserves one more moment of attention, because the math of recovery from a low base is what makes this setup asymmetric. Historically, ON has demonstrated the capacity to generate operating margins well above 20% when utilization rates normalize and pricing stabilizes — 2024’s 25.0% reflected a genuine period of supply-demand balance, not an aberration. If operating income recovers to even half the 2024 level, approximately $885 million at a similar revenue base, the margin picture shifts from distressed to merely recovering, and the stock’s forward multiple compresses rapidly. A 10% move in operating income is noise at the current $84 million level, but from a $400–500 million base, that same percentage starts to reshape how the market reads the earnings trajectory. The math becomes interesting before the full recovery completes.
$84.2 million on $5.99 billion. Cyclical recovery in power management is already reflected in the share price, but structural margin expansion back toward historical highs is not yet — and whether the inventory cycle cooperates with the timeline the market seems to be assuming is a question the next two earnings reports will begin to answer, one that matters more than anything the broader sector rally is currently suggesting.