Keysight (KEYS) Earnings Aug 18: Orders +56%, Guide +1%

TL;DR

• Keysight reports fiscal Q3 after the close on Tuesday, August 18, call at 4:30pm ET. The quarter ended July 31. Company guide: revenue $1.730-1.750 billion and non-GAAP EPS $2.43-2.49.
• Orders hit $2.051 billion last quarter, up 56%. Revenue that quarter was $1.72 billion. Book-to-bill of 1.19, and roughly $330 million of it went into backlog rather than the income statement.
• Then Keysight guided this quarter's revenue to $1.74 billion at the midpoint, about 1.2% above Q2. A company sitting on that order book set a low bar for itself.
• The stock closed Friday at $357.82, against a 52-week range of $152.85-$374.96 and 32.9x forward earnings. It has more than doubled off the low.
• Consensus sits one cent under the top of the guide: $2.48 against a guided $2.43-2.49, on revenue of $1,748.1 million against a guided top end of $1,750 million. Our own earnings calendar puts the whisper at $2.63.
• The $360 straddle expiring August 21 prices about ±9.4%, roughly $34 on a stock whose revenue guide implies a quarter it should clear comfortably.

The Board

Orders ran 19% ahead of revenue last quarter. The guide for this one barely moves.

When Does Keysight Report Earnings?

After the close on Tuesday, August 18, with the call at 4:30pm ET, per Keysight's own announcement. The fiscal quarter ended July 31.

It lands the same evening as Toll Brothers, Mercury Systems and La-Z-Boy, on a day that starts with Home Depot and housing starts. Keysight is the only one of them that reads on AI capex.

Twelve analysts average $2.48 of EPS on $1,748.1 million of revenue, with the EPS spread running $2.45 to $2.55. Both sit at the top of Keysight's own guided ranges, which is the setup this site keeps flagging: when consensus lands on the ceiling of the guide rather than the middle, a merely in-line quarter reads as a miss.

Orders Grew 56%. The Guide Grew 1%.

Take the two numbers from May's release side by side. Orders of $2.051 billion, up 56% year over year and the first time Keysight has crossed $2bn in a quarter. Revenue of $1.72 billion, up 31%. Non-GAAP EPS of $2.87 against a $2.32 consensus, and $472 million of free cash flow. Details are in Keysight's Q2 release.

Divide orders by revenue and you get a book-to-bill of 1.19. About $330 million of demand booked in the quarter did not ship in the quarter.

Now the guide. $1.730-1.750 billion of revenue, so $1.74 billion at the midpoint. Against $1.72 billion in Q2 that is 1.2% sequential growth, from a company that just added a third of a billion dollars to its backlog. Management called for full-year growth in the high 20s and raised the outlook, and still guided the next three months to roughly flat.

I read that as capacity and revenue recognition rather than softening demand. Test instruments ship against acceptance schedules, and an order booked in April for a data-centre validation programme does not become revenue in July because someone wants it to. But it does mean Tuesday's revenue line is close to a formality, and everything that matters sits elsewhere in the release.

The Number That Decides Tuesday

Orders. Specifically, whether they hold above $2 billion a second consecutive quarter.

A single $2bn print is a lumpy quarter in a business that sells to hyperscalers and defence primes. Two in a row is a run rate, and at 32.9x forward earnings the stock is priced for the run rate. Morgan Stanley's upgrade to Overweight with a $400 target explicitly framed Keysight as an AI capex proxy, which is a different valuation regime from the 20-somethings multiple test-and-measurement usually carries.

The arithmetic in the guide is worth checking against itself. EPS of $2.46 at the midpoint is about 43% growth, on revenue up about 29%. Both decelerate from Q2's +31% and +69%. Operating leverage is still there; it is just narrowing, and it narrows further if orders come in at, say, $1.7 billion. That would still be roughly 30% growth year over year and it would still read as a deceleration against a $2.05bn comparison. Backlog conversion is the bull case and the trap in the same paragraph.

Also worth a line: Keysight closed the VPIphotonics acquisition and has a public collaboration with Lockheed Martin on airspace protection. Neither moves a quarter. Both explain why the aerospace and defence half of the business is growing alongside the AI half, which matters if you think the AI order book is a single-cycle event.

The Options Angle

An implied ±9.4% is roughly $34 on Friday's $357.82. For context, that is more than the entire market cap of a mid-size test company, priced as a two-day range on a $61 billion one.

I do not want to buy that. The revenue guide is the easiest part of the print, the EPS guide has a six-cent width, and Keysight has beaten its own non-GAAP number in each of the last several quarters. A straddle at 9.4% needs a genuine surprise, and the most likely surprise here is an orders number that disappoints relative to $2.05bn, which is a directional event rather than a volatility one.

I also do not want to sell it naked into a name sitting 4.5% off its 52-week high with AI capex as its swing factor. July taught this site that lesson expensively: realised moves ran well past implied in name after name, and the losses came from being short the wrong tail. Defined risk or nothing.

What I would actually take is conditional and I am not pre-positioning for it: if orders print above $2bn again and the stock gaps, a call spread financed into the pop is a cleaner expression of the run-rate thesis than owning the print. If orders come in below $1.9bn, the multiple is the problem and no amount of backlog commentary fixes it that night.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Conviction	Breakeven
1	Pass	Long straddle	$360 strike, Aug 21 expiry	~9.4% of spot, about $33.60	$357.82, Aug 14 close	±9.4%	6/10	needs >9.4% either way
2	Pass	Short premium (any naked structure)	n/a	n/a	$357.82, Aug 14 close	±9.4%	7/10	n/a

The implied move and the $360 strike come from our own earnings calendar dataset, dated August 15. I did not source a live quote for the straddle itself, so row 1 is priced off the implied move rather than a filled premium. The full ledger is at /data/track-record.

The One-Line Read

Keysight booked $2.05bn of orders and then guided revenue up 1%. Tuesday is not about the quarter it already told you about. It is about whether $2bn was a spike or a floor.