CEO Claims Investors Got It Wrong on $30 Billion Shortfall as ON Stock Plummets 9%

ON Semiconductor (NASDAQ:ON | ON Price Prediction) hosted an investor day designed to reset the story around its power portfolio and its AI data center runway, and the shares fell through the presentation. The stock closed the regular session down 9.02% at $66.60, extending a one-month drawdown of 20.12% even as the year-to-date return remains positive at 22.99%.

CEO Hassane El-Khoury spent his CNBC hit explaining a gap in the slide deck. The total addressable market shown on stage was smaller than the one investors had been carrying in their models since June, and the growth math behind the AI data center opportunity was anchored to a consensus assumption rather than a proprietary one. Both explanations may be technically correct, yet the sellers still showed up during the event itself, which is the only question that matters when a management team schedules a day to make its case and loses ground doing it.

Where the $30 Billion Went

Hassane El-Khoury presented a total addressable market of $213 billion at the investor day, well below the $243 billion figure the company cited in June when it announced the Synaptics (NASDAQ:SYNA) acquisition.

His explanation was procedural. “Back in June, when we talked about 243, we were talking about the thesis and the strategy with the acquisition of Synaptics, according to ON Semiconductor. Today, the focus of today is core onsemi business because once you announce the deal, we’re two separate companies,” El-Khoury said.

That is how antitrust review works. Acquirer and target run independently until closing, and presenting a combined market opportunity before regulators sign off invites problems the company does not want.

The trouble is that the Synaptics transaction is expected to close in mid-2027, and a shareholder pricing the combined entity has to underwrite nine more months of separate reporting while rates and cyclicality do their own work.

Conservative or Unfalsifiable?

El-Khoury framed the AI data center growth assumption as prudent, saying the company “talked about a 30% market, which is in the middle of the range that everybody’s talking about” and then “anchored on a 10% on top of whatever market you want to pick.”

Anchoring to the middle of the range is consensus dressed up as prudence.

The 10-point outgrowth claim is the more interesting number because it is testable only through design wins. Management already disclosed real ones, including expanded content in the NVIDIA (NASDAQ:NVDA) MGX ecosystem, two power-supply platform wins with Great Wall in China, and AWS power-supply and battery-backup content in the United States.

Until those wins translate into revenue that visibly separates ON from the market it is anchoring to, the outgrowth is an assertion, and the buy side is entitled to want more than that on a day billed as a strategic reset.

Margin Bar and an Automotive Deflection

El-Khoury called the company’s 53% gross margin target a milestone on the way to a higher bar. Getting there requires the revenue mix to shift meaningfully toward silicon carbide, gallium nitride, and Treo analog content, which carry richer margin profiles than the legacy portfolio the company has been exiting.

The Q2 non-GAAP gross margin of 39.3% gives you the current baseline, and management’s own sensitivity of 25 to 30 basis points of gross-margin improvement for every point of utilization shows how much of the bridge depends on volume that has not yet arrived, disclosed in the Q2 2026 8-K.

Asked about the recent stretch, El-Khoury pivoted to the long run, saying ON “grew 72% our revenue in automotive from 2019 to 2026.” That is a fair rebuttal to an unfair quarterly comparison, and also a convenient way to avoid discussing an Analog and Mixed-Signal segment that was down 2% year over year in the most recent quarter.

Bull and Bear Case for ON Stock

The bull case is that today’s selloff has already repriced a business whose per-vehicle and per-rack content is measurably rising, and whose AI data center revenue is guided to more than double in 2026 with silicon carbide inside AI applications growing nearly 60% year over year. The forward P/E near 17x against an average analyst target of $106.88 is where that argument lives.

The bear case is that automotive semiconductor demand is cyclical, the Synaptics deal adds integration risk before it adds earnings, and a company that has to explain its own slide deck on stage is not controlling its narrative. The trailing P/E of 47x leaves little room for another miss.

The variable that decides between them is design-win disclosure. If ON can name the AI data center customers and rack content behind the 10-point outgrowth claim over the next two quarters, the investor day gets re-read as a buying opportunity. If it cannot, the anchoring language starts to sound like a placeholder.

Contact [email protected] for any questions or corrections.