When Does onsemi Report Earnings? August 3, and the Stock Fell 14% Into a Quarter It Already Guided

TL;DR

• onsemi reports Q2 2026 after the close on Monday, August 3.
• Consensus: adjusted EPS of about $0.72, up roughly 36%, on revenue of about $1.59 billion, up about 8%.
• Both sit inside management's own guide of $1.535bn to $1.635bn and $0.65 to $0.77, near the midpoint of each. The headline beat is close to pre-agreed, again.
• The stock closed $81.61 on July 31 and is down 14.2% over the past month, while the average sell-side target sits at $113.12, about 39% above the price. Somebody is badly wrong.
• The line that decides the reaction is AI data centre revenue, which management now expects to double year over year in 2026 after originally guiding high teens.

When Does onsemi Report Earnings?

The short answer: after the close on Monday, August 3, 2026.

It is the quiet chip report of the week and the interesting one. Everybody will be watching AMD on Tuesday evening, which is the loud read on AI compute. onsemi reports the night before and covers a completely different part of the same build: the power that goes into the rack rather than the silicon that does the maths.

What the Street Expects

Line	Q2 2026	Comparison
Adjusted EPS	~$0.72	up about 36% year over year
Revenue	~$1.59B	up about 8%
Company revenue guidance	$1.535B to $1.635B	consensus sits at the midpoint
Company EPS guidance	$0.65 to $0.77	consensus sits at the midpoint
Q1 2026 actual	$1.513B revenue, $0.64 adjusted EPS, 38.5% gross margin
Options implied move	~10%	May's print implied 9.4% and realised 4.3%

Consensus is parked exactly in the middle of the guided band on both lines. That is the most common shape in this season's previews and it means the same thing every time: clearing the number is the default outcome and is worth close to nothing. The reaction gets set by the September guide and by one segment.

The Board

Consensus at the midpoint of the guide. The argument is entirely about what happens next.

The Gap Nobody Wants to Talk About

The stock is $81.61. The average analyst price target is $113.12. That is a 39% gap, and it has widened rather than closed because the stock fell 14.2% over the past month while targets did not move with it.

A gap that wide is not a buy signal on its own. It is a disagreement, and disagreements resolve in one of two directions. Either the sell side is slow to mark down a semiconductor name into a cooling tape, which happens constantly, or the market is front-running a guide that the analysts have not modelled yet.

What makes this one worth sitting up for is that the fundamental news has been getting better, not worse. Q1 was the quarter onsemi's automotive revenue grew year over year for the first time after seven consecutive quarters of decline, and the quarter AI data centre revenue more than doubled year over year and grew more than 30% sequentially. The CEO's phrase was that the business has "moved beyond the cyclical trough". The stock has fallen 14% since roughly that point.

The Three Things to Watch

1. AI data centre revenue and its guide. This is the whole print. Management raised its 2026 expectation for this line from high-teens growth to doubling year over year, which is an unusually large mid-year revision and the reason the stock re-rated in the first place. onsemi sells the power delivery silicon that sits between the wall and the GPU, and every gigawatt of announced capacity needs it. If that line decelerates, the entire 2026 story in this stock is wrong.

2. Gross margin, which is the honest scoreboard. Q1 came in at 38.5%, and management guided to continued sequential expansion. Analog and power businesses live and die on utilisation, so the margin line tells you whether the fabs are full independently of what the revenue line says about pricing.

3. The gap between GAAP and adjusted. Q1 GAAP diluted EPS was $0.08. Non-GAAP was $0.64. That is an eight-fold spread, driven by restructuring, and it is the kind of thing readers should look at directly rather than take on trust. A company mid-restructuring can produce a perfectly clean adjusted number while the audited one says something much less flattering. Check whether the charges are actually shrinking, because a "one-off" that repeats for four quarters is a cost.

The Silicon Carbide Position Is Real

The part of onsemi that is genuinely differentiated is EliteSiC. Its silicon carbide held roughly 55% share of new EV models displayed at the 2026 Beijing Auto Show, and the company has taken 900V platform wins with NIO and Geely plus expanded IGBT traction inverter business with North American manufacturers.

That is a strong position in a market with an awkward property: Chinese EV volume is the growth, and Chinese EV pricing is the problem. Winning 55% of designs in the world's most price-competitive vehicle market is a share statistic that does not automatically become a margin statistic. Watch what automotive does to blended gross margin, not just to revenue.

The other cross-current is cost. We have written about which companies get hit next by memory prices, and analog chipmakers are less exposed than the handset and PC names: Qualcomm reported EPS down 20% and blamed exactly this. onsemi's input problem is substrate and wafer cost, not DRAM. That is a genuine relative advantage this season and it is worth saying out loud, because the market has been selling semis as a block.

The Options Angle

• Options price a move of about 10%, and the honest counterweight is that onsemi has only matched or beaten its implied move in four of the past eight prints. In May the options implied 9.4% and the stock moved 4.3%. Paying 10% for a guided quarter is a poor bet on that record.
• The setup argues for defined-risk directional exposure rather than outright premium. You are not betting on the quarter, which is guided, you are betting on the AI data centre guide, which is binary in tone even when it is incremental in dollars.
• Do not hold this through Tuesday as well. onsemi prints Monday, AMD prints Tuesday, and semiconductor correlation this month has been close to total. A right call on Monday can be erased by Tuesday evening before you can act on it.
• Live prices for individual strikes could not be sourced, so the plays below are quoted against the July 31 close and the implied move.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Breakeven
1	Bullish, defined risk	Call spread	~$85 / ~$95, first monthly after Aug 3	debit not sourced	$81.61 (Jul 31, 2026 close)	±10%	needs a close above ~$85, +4.2%
2	Pass	Long straddle	at-the-money weekly	~10% of spot	$81.61	±10%	needs a move beyond ±10%, about $8.16
3	Pass	Short premium into the print	any	credit not sourced	$81.61	±10%	loses beyond ±10%

Rows 2 and 3 are both logged as passes so they get scored in both directions. We are saying onsemi moves more than a normal earnings reaction but less than the 10% a straddle needs, which is a narrow claim, is exactly what happened in May, and deserves to be marked right or wrong.

The One-Line Read

onsemi reports after the close on Monday August 3 with consensus of $0.72 on $1.59 billion sitting at the midpoint of a guide the company set itself, so the quarter is close to pre-agreed and the argument is entirely about the AI data centre line management has already told you will double this year: the stock has fallen 14.2% in a month while the sell side left its targets 39% above the price, which is not a valuation opportunity so much as an unresolved argument, and Monday evening is when one side of it gets marked.