Keysight Technologies, Inc. (KEYS) Earnings Call Transcript & Summary

Okay, everybody. I think we'll go ahead and get started this morning. Thanks for joining us for the Keysight presentation. I'm Rob Mason, the senior analyst with Baird that covers advanced industrial equipment, which, of course, includes Keysight as kind of a cornerstone of our test and measurement coverage. I want to introduce Neil Dougherty, who's with us Senior Vice President and Chief Financial Officer; and Kailash Narayanan, who's Senior Vice President and Head of Keysight's Communications Solutions Group. This is going to be an open forum. Please send your questions up if you have any. We got plenty of questions to get us started.

We're just going to jump right into Q&A. So I think Neil, first off, just update us kind of on state of business, what you're seeing in the marketplace. You reported your quarter just a couple of weeks ago. You talked about some baseline expectations for this year. What's baked into that and just kind of run through different areas of the business to give us a -- level set us on where you've seen.

Yes, absolutely. Right now, it feels like things have broadly, I guess, you'd say, stabilized in the end markets if you take them in aggregate. We really didn't see much change in our outlook over the course of the last 3 months. And so it does feel like, obviously, a bit of a pause across most of our end markets. There are some positive signs with our wireline business inflecting positively driven by AI. That has been enough to turn growth in our broader commercial communications business, positive on the order side for the first time in 6 quarters. So that's positive. But we do believe that those longer-term secular themes that have been driving growth in our business remain intact, right, that there's a lot of -- on the wireless side, 5G evolution that still has to happen as we look forward to 6G sometime in the late stages of this decade, when we transition to 6G, wireline already inflecting. Aerospace defense has been the one business that never really saw a downward inflection on the order line and this continued to be quite stable and positive with growth in line with expectations. And then on the EISG side, which went into the downcycle about 6 months after the communications business, probably have a few more quarters of tough comps there. But the AV and EV investments to drive next-gen auto are still intact. In the semi business, you still have the significant efforts to ensure supply and re-onshoring and pretty dramatic growth expectations for chip production over the intermediate term. So the long-term secular themes are intact, and we're just kind of taking things 1 quarter at a time, managing the business with discipline, focus on winning every competitive deal that's out there, so we can maximize the business in these conditions as we wait for the coming market inflection.

Okay. Okay. We'll dive into some of those different verticals. But just again, if you think at a high level, the way you're managing the business today, managing the cost side of the business, you're speaking to OpEx being down this year, kind of low single digits. Help us on -- and you've built the business, constructed the business to have some variability in cost as revenues flex. How should we think about that decline in OpEx versus kind of a structural efforts cost takeout that you've also been executing on as well?

Yes. So I mean, there's kind of 2 elements. So as Rob has suggested, we do expect our OpEx to be down low single digits on a year-over-year basis this fiscal year, but virtually all of that savings coming from the SG&A line items. We're keeping our foot on the gas from an R&D perspective, want to make sure that we have the right portfolio of solutions to capture the upturn when it comes. But we're taking actions to reduce spending in SG&A. So there are -- as Rob has suggested, we have some pretty unique structural flexibility elements in our cost structure. Most notably, that 100% of our employees have a portion of their pay that's variable and fluctuates with business performance. So that is right now cycling down as it was designed to do, and it's -- obviously, our salaries are the single biggest component of our cost structure. So that's providing significant savings in this period of time. In addition to that, we've taken action, our head count is down 4% to 5% over the course of the last year. That's a combination of absorbing attrition. We provided an early retirement incentive. We did -- we've done some pretty surgical workforce restructuring where it was necessary in order to further reduce savings -- reduce spending and drive savings. So again, continuing to invest in the portfolio and the future growth of our business, but then again, staying disciplined in this portion of the cycle.

How do you think about those, I guess, structural decisions that you made impacting incremental margins if you -- as we come at some point to a recovery?

Yes. I mean I think some of those things are permanent in nature. You're finding efficiencies, ways to do things in a lower-cost fashion or just stop doing them entirely. Other portions of them may come back over time, but it would be feathered in. And so I think it does drive, give us an opportunity when the upswing comes to do a good job converting top line growth into profitability. We've talked about when our business grows mid-single digits or better. We can drive 40% operating leverage, and we would expect to do that again as the business recovers.

And just maybe last question on the cost structure, and I'll toss one over to Kailash. Gross margin has been -- was a focal point coming out of the second quarter as well because there was some flex there within the various businesses. I think that was mainly sales mix driven. But as you think about gross margin, gross margin stability and any volume leverage as we go forward on a recovery, will we see that? Or is your gross margin such that it's more mix dependent?

Yes. Gross margin is mix dependent. I mean, I think Keysight as a total company has done a pretty good job. We're right there around 65% gross margin. Gross margins have been more stable in Kailash's business in the communications side in the upper 60s, 67%, 68% range. We've seen a little bit more variability in the industrial business because there's frankly a broader dispersion of gross margins in the portfolio in that business. So there's greater opportunity for mix to have an impact. They have -- on the hardware side, they have amongst the highest margins, some of the highest margin software -- hardware that we have in the business, but then they also have more of our low-end instrumentation, a little bit more of a manufacturing focus, those instruments tend to be lower margin. So there has been more of an impact on the industrial side. But I think as markets recover, we remain committed to the financial targets that we put out at our Analyst Day, which is now 15 or 16 months ago, to continue to migrate gross margins north into the upper 60s and operating margins into the low 30s, 31%, 32% range. We may need to push out the time line to deliver those given this downmarket environment that we're in currently, but we remain committed to those financial targets.

I wanted to flip to the communications business, Kailash. Again, I think one of the standouts in the quarter was the inflection we've seen around those wireline portions of your business. Can you, I guess, step back, just give us a quick overview in terms of what wireline comprises as a part of the overall commercial communications part of the business? And then where your involvement is in the AI data center build-out, which seems to be a significant catalyst for portions of your business?

Yes. Thank you, Rob. So when you look at commercial comms, it's roughly a $2.2 billion, $2.3 billion business. The wireline portion of it is somewhere between 40% to 50%. So that's where it is. And right now, the AI/ML application and investment is a tailwind for that business. It's a driver, and we grew the wireline business double digits this past quarter. And AI/ML is both a direct and an indirect driver of this business. We're seeing multiple use cases, right? If I step back, there's just a sheer capacity expansion of data centers and data center infrastructure. So that's one. You have increase in data rates and speed grades. So that's creating a new wave of design and deployment of different types of complements, switches, transceivers, amplifiers, everything that make up these data center networks. And then three, protocols, the way in which GPUs and CPUs and memories communicate with each other. So there is an expansion in that. So all of that, the capacity expansion, the speed grade expansion and the protocol expansion is creating a need for design, emulation and test capabilities, which is what we're benefiting from.

And you've talked about actually some of those emulation tools being in higher demand and particularly, I guess, at hyperscaler type customers. Are the -- should we think those are new customers -- new customer categories for this business? And just where are -- where do you think those customers are in terms of their adoption of these type -- those types of tools?

Today, if I look at this business, we're primarily driving 3 use cases. So you see GPUs, server design and manufacturing. That's a use case. You have these GPUs in racks, GPUs stand-alone. There are lots of interconnects and things like that, that go on. All of that requires testing -- design validation and testing. So that's one use case. And it tracks a bunch of customers, component customers that we serve. The other use case is this build-out of these data centers, these AI clusters. It's not just the GPUs, but you need optical components, you need transceivers, you need switching fabric. So you have players participating in that. So that's sort of the second use case that we're covering. And the third use case is the hyperscalers who want to emulate AI/ML workloads. This is a new type of workload. They're trying to understand network latency. They're trying to understand how well their GPUs are being utilized. And these require solutions to help model all of that in a lab setting prior to deployment. So we're serving that customer base. So when I look at it, if I sort of build it out for the benefit of the audience here, you have chipset players think about the Marvells, the Broadcoms of the world, the GPU makers. They're looking at chip interface testing, chip R&D -- chipset R&D, we're serving that base. So one level up, you have these component makers. Think about the Coherents, the Lumentums, the InnoLights. They're building a lot of transceiver components and things like that, that require power integrity wavelength test, compliance test. These are solutions that we provide to that ecosystem. And then you shift one level up, think about the network gear makers like the Ciscos and the Junipers and the Aristas of the world. They need capability to emulate traffic and AI/ML workloads, different speed grades. We need to address cybersecurity concerns for them. So we provide a lot of emulation capabilities to that group. And then you go one level up, you have all of these ODM companies, the Quantas, the Compals, the Foxconns of the world, making server blades using some of these components and chips and we -- they need to figure out how their servers interop with other servers in a network and we provide capabilities to test that interop. And finally, at the top of the food chain, you have the big hyperscalers that need capability to emulate traffic, that need capability to emulate utilization rates, latency times and so forth. So it's a -- this is sort of the tier we're selling our current solutions, but we also have the opportunity to drive newer use cases and newer solutions that we have introduced. And it's attracting newer players. We're seeing adoption, not just from these existing traditional network players like the Ciscos of the world, but you also have many new players that are getting attracted to each of these tiers driven by AI/ML because somebody wants to make a specific digital signal processor, somebody wants to make a specific amplifier, an optical amplifier or somebody that wants to make a retimer chip. All of those customers are new customers that need our capabilities.

When you think about that new customer set, what would their history be with Keysight in the past? Because Keysight obviously has been around for many, many years and well established in some of these more traditional customers. How are you winning that new business if they're less familiar with Keysight historically?

Yes. So maybe we compartmentalize that into customers that we work with, the big names that -- some of those names that I mentioned. Clearly, they're known to us, and we're a part of many different standards consortia. We partner R&D to R&D with these customers. So clearly, they are aware and know about our capabilities. And we're doing proof-of-concepts with them, which ultimately leads to a sale. And in general use cases like traffic emulation, we already have a presence. We have capabilities that we're expanding. So that's sort of on the existing installed base. But this installed base is adopting faster driven by AI/ML. Typically, it would have been more of an IT sort of upgrade that drives some of these -- some of this adoption, but AI is now serving as an accelerator to adopt these new speed grades. We've always talked about enabling technology megatrends like 400G and 800G and terabit. Now there is a reason, AI/ML is sort of the application that's driving that adoption. So that's kind of on the existing customer base. These newer customers, we have aggressive as well as creative go-to-market options. Some of this, we know because we're engaged with some of these top players. We know what that ecosystem is, how it maps out, what their supply chain is. Again, we attend many standards bodies. We are lead members, comembers. SONiC is a good example, which is -- it's a cloud-based operating system, we work with Microsoft, and we know newer players. DASH is another alliance. These are all network and alliances that have many players and through that, we have a go-to-market reach on many of the newer players.

I see. And then if -- you talked about research functions, production functions. If we apply the traditional lens, we often look at Keysight under with production test, R&D-related tests, maybe there's an operational element in this as well. How does your business, the wireline business, in particular, shake out under that lens in terms of mix of business where you're serving?

Yes. It's largely levered as is all of commercial comms. It's largely levered to the R&D workflow. So when you think -- when we talk about a speed grade increase, every component in the network has to be redesigned. You have to come up with new designs that require R&D capabilities. It requires simulation, emulation capabilities. And we have a solution that covers the electrical, optical spectrum, the physical protocol application spectrum. And we -- and that's the rich portfolio that we offer in the R&D space. So that's one element. And that's the larger element, whether you're designing a component or you're designing a subsystem, the rest of the setting has to be emulated. So you could be a component maker, but you need to know how the network is going to hit you. You might be a subsystem, you might deploy a subsystem like an AI cluster, you need to know how it operates in the context of the overall network. We provide emulation capabilities to those customers. So that's sort of the R&D workflow. The manufacturing workflow is just in the sheer volume of deployment as some of these components, these optical transceivers that I talked about, that's a smaller portion of the business, but that I don't know if we have sized it, but I'll just say that the mix is -- it's very high margin manufacturing business. So we're comfortable about preserving our margins as we move forward.

Relative to company averages, the wireline business has a higher exposure to R&D and a lower exposure to manufacturing.

Very good. Flipping over to the wireless business, that side of the commercial communications group. It's been in kind of, I guess, stable area the last few quarters. The questions often come up, though, what is the catalyst to start to see that accelerate. It sounds like you think there's still a lot of opportunity and build-out still to go with 5G and maximize the opportunity set there. But what's the catalyst that starts? Is there something on the horizon from a standards point? Or is it just more of a macro driver recovery that would catalyze that business?

Yes. As you pointed out, the business has been on a tear and has suffered a bit of a pullback. It does feel like we're bouncing along the bottom here. We look at -- at a macro scale, you look at 1 billion 5G subscribers, still substantial room to grow there compared to where LTE is. You look at the number of networks that have been deployed, it's still less than 1/3 of the number of LTE networks that have been deployed. There is room for growth there. And then newer use cases and newer device types. About a couple of thousand devices have been announced or launched and we see that expanding as the decade moves forward. So as we come out of this bottom, we do see these trends as providing secular drivers for steady investment. So that's one element of it. The second element is just standards progression. Release 15 was commercialized back in '19. You had the C-band expansion in '22, '23. And then now we're looking at Release 17. So R&D activities are proceeding in Release 17, and we have 18, 19, 20 coming up leading to 6G, which is Release 21 at the end of the decade. These standards progress kind of in an overlapping manner. So as we speak, there is deployments that continue to happen using Release 15 and 16 in some applications. R&D activity is going on with Release 17 and 18. And then when we look ahead the next couple of years, it'll -- there will be heavier R&D in 18 and 19 -- Release 18 and 19 when Release 17 gets deployment. So there's this overlapping wave of the previous technology or previous version of the standard gets deployed when the next version of the standard gets invested in R&D, and that's sort of how it progresses setting up a secular driver. When I look at catalysts, I would look at FR2 in China is a catalyst ahead of us. There's been talk about that. It hasn't happened yet. There is full-scale 5G stand-alone deployment, which delivers the full promise of 5G, driving newer use cases like industrial and private networks, which will pull newer types of devices for controlling robots and whatnot. And now with -- so that's another catalyst 5G SA. And AI PCs and AI smartphones, there's talk about that. In the next 18, 24 months, you're going to see a new wave of SKUs in that space that will drive more data generation, data needs, newer applications and increased subscriptions and adoptions. That's another catalyst that we're looking at.

Is there anything on the technology side -- from a technology challenge standpoint that could change the economics for the providers and improve their ROIs and perhaps accelerate deployment or is it just more about the progression that we have to go through standard-wise?

We're -- there's a lot of quality of experience and quality of service improvements that need to be made. There's a heavy focus right now on energy efficiency. Networks consume a lot of power. There's innovation that is required. We have capabilities to drive that. So that's one area. Satellite connectivity is something that is still being figured out, especially in terms of scale. So that's another area that's going to require ongoing R&D investment. The deployments and the subscription growth that I talked about will require investment to better utilize the networks to provide better quality of service, that will all require network upgrades and investments. So these are areas where we're helping. We have capabilities to also drive 5G stand-alone. That will require heavy investment. I mean there is -- certainly, there's been a pullback on the operator side, not able to fully monetize what they've deployed. It's a little bit of a chicken and egg. As the subscriptions increase, the networks are not going to be able to support this quality of experience requirements, and that will require newer upgrades.

Yes. Maybe just before -- while we're still on the communications segment, I'm not sure what you can really say about it, but Spirent is an opportunity that you see to expand the addressable opportunity market within communications. It brings some new capabilities as well. What can you say about the value proposition to make an offer to acquire Spirent?

I think there's -- we've identified $1.5 billion of SAM expansion opportunities in network assurance and in precision location. Both of those things, we believe, are highly complementary and additive to Keysight's current portfolio. I think we see opportunities for us to leverage our go-to-market strength across the broader Spirent portfolio and to leverage from a cost perspective our general and administrative infrastructure in a way that will allow us to make this acquisition accretive to both gross and operating margins post integration. And so I think we're excited about the opportunity to create value for shareholders, expand our market presence. And yes...

And the 2 capabilities in particular that were called out where network analytics and positioning is incremental -- viewed as incremental, how do those help Keysight and how would you be able to leverage those?

I mean you look at the positioning and location technologies, it's got applicability in comms, in aerospace, defense and in automotive. That's something that we don't have that is clearly complementary and you look at what's -- the trends that are going to play out over the next several years, that's very useful. Network live -- network monitoring, assurance, be it cybersecurity or mission-critical network infrastructure, all of that requires operational network, live network capabilities. So we think that is strategic and that's complementary to our business.

Yes. Just to touch up real quickly on maybe some of the industrial markets. The semiconductor, in particular, you talked about in the last coming -- on the last conference call, perhaps some optimism as we get to the back part of the year and moving into 2025. Is that optimism more around revenue coming through or potentially order inflection in that side of the base?

Yes. I mean it will start with orders. I think, obviously, the semi industry longer lead times, where Keysight plays, our lead times are significantly shorter than industry averages. But relative to Keysight's other products, they tend to skew a little bit longer. So we certainly see it. We'll see it on the order side first. We've already seen a little bit of an uptick as it relates to the memory side of semi. So that's at least one green shoot, if you will, that's actually being realized. I think the other things that we're seeing and hearing are really more about monitoring the industry, right? The comments that the big fab operators are making about, getting their -- some of the new fab construction finally to come online. These types of things are going to be net additive. And again, as we look forward to '25, we'll have to see what the actual timing of the inflection is, but it certainly seems like there's some momentum that's building.

So just flip back to M&A. Spirent, as mentioned, is pending. That's an acquisition that you're probably more waiting on regulatory approvals than just the voting process, approval process itself.

Shareholder vote was held a couple of weeks back and passed. We've kicked off the regulatory process. But as you know, that is a process which we will go through.

But you did complete ESI Group earlier this year. Just can you give us a status update in terms of where the integration is and how you're seeing the incremental opportunities start to bubble up there. That was viewed as more of a revenue synergy opportunity play for Keysight to broaden yourself out in some different verticals?

Yes. So we closed the ESI acquisition at the very beginning of this fiscal year, early November. So they've been part of the Keysight portfolio now for 7 or 8 months. Things are progressing very well. It's, again, a significant market expansion for us, but they focus on the broader auto industry and we believe their products are directly extendable into our aerospace, defense end markets as well. And so we're just in the process of getting it kicked off. The sales cycle is a little bit longer. But right now, the business is tracking at or ahead of our expectations, and we're very pleased with where we are early in the process.

If you think back or think about overall capital deployment and your priorities there, give us some perspective on your comfort level with leverage. I mean, this is 2 sizable transactions in fairly close proximity, certainly close proximity for Keysight. Just -- would your comfort level as we go forward, assuming Spirent closes later this year, what the expectation would be?

Yes. So the first comment I'd make about capital deployment is the thing that I'm really pleased about is that we've been able to keep our foot on the gas from an R&D perspective, even as our business has gone through a little bit of a market -- downward market inflection because I think it's going to position us really well as we come out of this. I think from an M&A perspective, as you said, we've done 2 transactions, ESI, a little over $1 billion, already closed. And then Spirent a little over $1 billion. ESI is a little under $1 billion. Spirent's a little over $1 billion, would close sometime early in the first half of our next fiscal year. I think from a leverage perspective, we're targeting somewhere in the low 2 turns gross leverage is where we think the right level is for the business over time. There have been periods of time, most notably when we did the Ixia acquisition back in 2017, where we've stretched the leverage up into the low to mid 3 turns of leverage kind of a range, but with a quick delevering back to target. So I think those historical data points give you a sense of what we're comfortable with, but I don't think we're going to need to stress the organization from a leverage standpoint to complete these 2 acquisitions.

Yes. Very good. Well, we're at time. We'll break there. There is a breakout session afterward if you have any additional questions for Keysight. Join us there. Thank you, gentlemen.

Thank you.

Thank you, Rob.