ON Semiconductor Q2 Results Beat Expectations, AI Data Center Revenue Expected to Double
Alina Collins
ON Semiconductor posted Q2 revenue of $1.604 billion and adjusted EPS of 74 cents, both above consensus, while guiding AI data-center revenue to more than double in 2026 — shares rose over 6% after hours.
Where exactly did the quarter beat?
Q2 revenue hit $1.604 billion versus the FactSet consensus of $1.589 billion; adjusted EPS came in at 74 cents versus the expected 71 cents.
This means → both headline numbers cleared the bar by a modest margin — not a blowout, but enough to push the stock past $85.92 in after-hours trading, up more than 6%.
Why is the AI data-center business the standout?
ON supplies power semiconductors — chips that manage electricity flow — to AI-chip data centers. That business generated roughly $250 million in 2025 revenue and is now guided to double to about $500 million in 2026.
In plain terms = every new AI server needs efficient "power-management" chips, and ON sells exactly that.
CEO Hassane El-Khoury credited the ramp to the strength of ON's smart-power portfolio and expanding customer adoption across the power tree — the full chain that delivers electricity from the rack down to each individual chip.
How big is $500 million in ON's overall picture?
Even after doubling, AI data-center revenue would account for roughly 8% of analysts' projected 2026 total revenue of about $65 billion.
This means → ON's bread and butter is still automotive and industrial power management plus image sensing — auto customers alone make up more than half of total sales.
In plain terms = AI is the fastest-growing new engine, but the car business remains the ballast.
What is the market watching next?
Whether the AI data-center unit can keep delivering on the doubling trajectory will be the key proof point for ON's power-chip penetration into AI infrastructure.
ON shares had already rallied about 50% year-to-date through Monday's close; the after-hours pop of over 6% shows the market is still willing to pay up for the AI narrative.
This reflects a broadening of the "AI picks-and-shovels" trade — investor enthusiasm has moved beyond GPUs into power management, on the logic that as long as AI compute keeps expanding, the power layer benefits too.
Content is for reference only, not financial advice.