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

[Audio Gap]

[Audio Gap] about the earnings. Like beat earnings and obviously, orders are better, but can you just remind everyone what were the 3 kind of big takeaways from the earnings that you want to share with everyone?

Yes. So first of all, I think we overachieved relative to both our guidance and street expectations within the quarter. And so if you go back, rewind 3 months to mid-May when we were putting together the guide, there was -- actually kind of the first 2 bullets, one was around orders. We had measurable outperformance in orders relative to our expectation 3 months prior, and then we did a good job of converting that order outperformance into revenue outperformance and then flowing that through down into EPS. So the first 2 kind of focused on the 2 various components of outperformance within the quarter. And then second -- the third point was the business, we remain focused and committed to value creation. We recognize that we're in a bit of a mixed demand environment on the downside of the cycle in a number of our end markets, but we continue to focus on execution, on driving profitability. Business continues to generate 25% operating margins. We remain committed to our capital allocation strategies, and again, remain focused on creating value for shareholders.

I think it's very important. And you've used M&A pretty successfully over the course of the year to expand the addressable market, and we'll come and talk about that. But one of the things that caught my attention was you talked about the return of returning to long-term growth. And just to remind us, is that the 5% to 7% growth? Is there another target? And what are sort of the puts and takes you see over the next 3 to 5 years to get to that growth rate?

Yes. So our long-term target remains at 5% to 7%. That's the target that we put into place at our Analyst Day in March of 2023. And admittedly, at the time that we were standing on the stage down the road here at the New York Stock Exchange, we didn't foresee what the next 18 months were going to look like. And so -- but if you think about where we are now and how we got here, going back to 2020 and the disruption in 2020 relative to COVID, but then we had 2 periods -- a 2-year period of growth that was well above long-term market norms, right? You saw the post-COVID balance in '21, [ qualified ] by the supply chain disruptions that were itself a significant demand driver in '22. And then we've seen the business pull back from those levels now with about 18 months of below market growth. And so as we look forward, admittedly, at this point, visibility is still somewhat limited. I think the long-term secular drivers in our industries remain largely intact. But we think the most likely center at this point is a bit of a gradual climb out from where we are today, at least that's our base case. And if we get something that's steeper, we'll deal with that when it comes. But I think we're going to be looking towards things like AI that's driving wireline or continued evolution of 5G. There is still a lot to go there. The semiconductor -- the increase in semiconductor demand and the onshoring of semiconductor demand are going to be some of the things that are going to drive our business back to growth.

It's always good to under-promise and over-deliver. Investors like that, because I've seen too many of the other ones. And it sort of works, because I think about coming into calendar '23, I just remember how much backlog you came in coming out of the pandemic, and there was that disconnect between strong backlog, strong visibility versus revenue growth was slowing as you were burning off that backlog and looking at order replacement. So it's nice to see that you've kind of turned the corner on that. Let me ask first about M&A because it's been -- I'd like to say, and correct me if I'm wrong, but I feel like it's a core part of the Keysight story. As opposed to some companies, they do like occasional deals or maybe a transformational one, but I feel like tuck-ins and growth is kind of a core part of the story. So maybe remind us how do you think about -- because you're in so many end markets, how do you think about M&A? And how do you think about what the ultimate TAM for Keysight is?

Yes. So there's a couple of parts to it. I think the TAM as we currently define it is a $21 billion or $22 billion TAM, roughly half of that in commercial communications and the other half spread relatively equally between the industrial end markets and the aerospace and defense end markets. And I think -- I mean, you're right, we've done a fair amount of M&A over the -- we have a 10-year history now as an independent company since spinning off from Agilent in 2014. I think we've done 22 or something deals. Most of those smaller tuck-ins, some things that are a bit larger. And I think we're looking at a couple of different things, right? A lot of the tuck-ins are really designed around plugging small gaps in the portfolio, bringing in talent, in some cases, getting early seeds into new markets like quantum computing. We did some small tuck-ins in the quantum computing space. And I think some of the larger deals we've done have been more around expanding our addressable markets. So I look at the most recent acquisition of size that we closed, which was our acquisition of ESI, which closed in the first quarter of this fiscal year, really aligned with the stated objective to get more presence in these design engineering tools. Keysight's always had a presence in that space -- not always, for the last 30 years, [ this has ] a presence in that space via our RF and microwave EDA franchise, but this acquisition of ESI moves us more into kind of broader multi-physics computer-aided engineering types of simulations, and it allows us to provide a more complete engineering toolkit to our customer set.

It's funny because you mentioned Agilent. And I just -- I mean I go back and think about the genesis of HP, the original HP, and how many companies out there are part of that original core engineering talent and how built into the DNA it is of the company and all the companies, Broadcom and HPE, HP, et cetera. But I want to ask about orders because, as I said earlier, like in -- if I go back to the beginning, first half of calendar '23, orders were softening at a time when you had a lot of backlog, so revenue growth was better, and it wasn't unique to Keysight. A lot of your peer group and hardware were experiencing the same thing coming out of the pandemic with all these backlogs. Orders started to turn the corner last quarter, right? I think if I saw it correctly, it was the first positive year-over-year order quarter in 6 quarters.

Six quarters. That's correct.

And that's a big deal. So I guess the first question is, like how sustainable is that? Or another way to ask it is what's driving that? Is that like a specific area of the business that's driving it? Or is it broad-based? Or is it green shoots? Like how would you characterize what's driving orders and how sustainable you think that is?

Yes. I think it's a great question. And so first of all, there were definitely positives to take from the corridor that we just put up. I mean we -- I have already mentioned the order outperformance relative to what we were expecting 3 months prior, and that order outperformance was really focused in 2 end markets. The first one is our wireline business. You can think about AI as the driver of investments in wireline, and certainly, there's been no shortage of discussion about that market opportunity here recently. And the second was in semi, where we saw, again, really the first signs of positive signs here. I think the question in semi is was this just a blip, or is this something that will sustain over the next couple of quarters, wherein on the wireline side, we have now seen multiple quarters of year-over-year growth, and so I think we have a higher degree of confidence in that as a business that is truly inflecting. And so the question gets to about sustainability, though, and I think we continue to remain cautious, right? Even if you said both semi and wireline were inflecting, it's still less than 50% of Keysight's total revenues, revenue exposure. And so we remain cautious. We still haven't seen that broad catalyst is going to lift the business in its entirety. And so we remain cautious going into 2025. We'll take the positive data points that we have and go. But again, I point back to those secular drivers that are going to ultimately drive growth in our business. And we -- again, we believe those remain intact. We just need a little bit of help, industry-specific help, across the markets we serve.

I can tell you one thing. I've been going to investor conferences for a lot of years, and I could tell you, when you look up at 4:30 in the afternoon and there's a pretty full room, you're doing something right. So I'm going to stop here. I have a bunch of more questions, but I'd love to open it up to the audience. If anybody had a question they want to ask, just raise your hand, or you can just yell it out.

Right here on the front.

So I mean I think there's a lot of talk about how just these transceiver failures can really cripple chip-to-chip communication and really hinder some of these training jobs. And it seems like your customers for this would be the ones who may want to maybe test more of these components. But when we're getting into these Level 2 and Level 3 data center networks, you're going to have maybe up to 10:1 transceiver to GPU ratios. So I guess I'm just trying to think like how many of these transceivers would someone -- would one of your customers even want to test? Because like on one hand, you have just the importance of a failure here being so crucial, but on the other hand, you have such a high number of these things going to the data center. So -- yes.

Yes. So I'm going to back up a little bit and then I'll come and answer your question more directly. So first of all, I think it's worth noting that if you think about our exposure to wireline, first of all, there are a lot of touch points. It's network silicon guys, it's transceiver folks, it's network equipment manufacturers, it's hyperscalers, it's connectivity. There are a lot of touch points in that ecosystem. The other thing that I've noticed, our business at the Keysight level is roughly 55% to 60% R&D, 30% manufacturing and the balance in kind of post-deployment operations. But our wireline business skews even more heavily towards R&D. We're primarily helping our customers in the R&D lab rather than the manufacturing line. Now we do, do some transceiver test in manufacturing, and all I'll say is that across numerous end markets, we see exactly the phenomenon that you're talking about, right? Early on, things get tested, 100% test. And as yields get better and as cost of failure gets lower, they start to simplify or reduce the testing. So this is a phenomenon that we've seen across many end markets over the years. And so we'll have to see how it develops here, but again, the primary exposures that we have in the space are on the R&D side rather than on the manufacturing side.

That was a good question. Going once, going twice. Any other questions in the field? All right. Let's go back then. And my friend gave me a good lead in to AI because it's obviously a hot topic, and over the last year, every company I've met, even if they're not anywhere in technology, they're trying to get in on this AI momentum. And you really have multiple plays here. So can we talk about where does Keysight play in the AI ecosystem? And have you quantified it as a percent of sales that you're either, directly or indirectly, you think you get from AI?

Yes. So again, I just touched on the fact that we have lots of touch points, and I'll expand upon that a little bit more. So I think, again, when we think about our commercial communications exposure, the first kind of subdivision of that we do is wireless versus wireline. And I think if you look at our -- and just rough math, wireless is 55% to 60% of commercial communications, wireline is 40% to 45%. That's the rough breakdown of the 2 businesses. And right now, it's pretty clear that AI is the #1 driver across that wireline ecosystem. In some way, it's impacting just about all of those revenue streams. And as I think about it, think about it that there's kind of another segmentation one can do. You can look at kind of what are the infrastructure-related investments, I would kind of put this in the category of rising tides raise all boats, right? So this is primarily the investment in increasing ethernet speeds, 100-gigabit to 400-gigabit to 800-gigabit to 1.6-terabit, and again, primarily servicing that ecosystem in R&D, physical layer and protocol there, and helping to essentially to drive that infrastructure development that's necessary for these AI work streams. And then there's another set of revenues that are much more AI specific, right, helping [indiscernible], as an example, where we provide tools now that help the hyperscalers essentially emulate their AI workloads but in a lab environment before they deploy a network. Obviously, if they can optimize those data flows, they can deploy a smaller but more efficient network, cost them less money, as an example. We're selling to the folks that are making the connectors that are connecting these GPUs together, so you could build these networks of GPUs. And so again, there continues to be a lot of data points, both kind of network-focused in how do you improve the overall performance in a network, and then things that are much more specific to the AI solutions that are in place today.

Okay. And is that -- all in with those business would be more than 10% of the company revenue?

Again, so rough math, if it's 40% to 45% of commercial communications, you're about $1 billion total for wireline.

Okay. Perfect. And going off script, I want to ask you a question because I hadn't thought about this emulate -- the hyperscalers emulating their environment. Given that up until, I don't know, a few months ago, it was pretty much just GPUs, maybe Google had their TPU, that was kind of it. Now I look at all the companies that are making custom ASICs or various other solutions for AI, that proliferation of other non-NVIDIA GPU, is that a driver for Keysight to test all those different [indiscernible]?

I think it is, and I think it will be. And one of the things that we talked about on our call is how the AI and wireline markets is still a relatively consolidated market, right? The primary revenue drivers for us are all with big household name types of companies. And we saw a similar phenomenon in 5G, right? Through the first multiple years of 5G, it wasn't a handful or 2, but it was multiple handfuls of customers. But now we probably have 400 or 500 customers in 5G. And what we haven't yet seen, on the wireline and AI side, is this democratization of the market opportunity, and I think that's something that's going to come and it's going to create more opportunity for Keysight.

The other thing I hadn't thought about today, I was sitting in the Marvell meeting a couple of hours ago, and they got a question -- because they also have like a communication networking business, and somebody was asking the question of how much an opportunity instead of just directly to hyperscalers is that telecommunications interconnect of all those data centers as a driver for them. So I would imagine that would at least have some effect for Keysight.

Absolutely, we're selling to both. Sure.

All right. Let's talk about wireless, because as I said, for 2 years, every time people talk about Keysight said, it's a 5G play. They could care less whatever else was going on under the hood there. You said earlier 5G is not done. There's still other opportunities, but -- there's 6G, but there's probably other things. So can we talk about just generally what drives the wireless business from here?

Yes. So first of all, not only is 5G not done, 4G is not done, right? We still have significant 4G business.

My grandmother has that phone.

Yes. So I think the point being is that these are very long cycles, and people are continuing to invest in the current -- so if you think about our customer set, right, in this ecosystem, right, they're investing across the suite of technologies, right? They still have some folks that are working on legacy 4G projects, eking out performance and sustaining the life and monetization of those networks. The lion's share of their teams are working on 5G-related programs, and then they have early-stage teams that are working on research for 6G, early-stage research for 6G. And what we're going to see is a continued evolution of those, right? The 4G ones are going to continue to roll off, the 6G ones will eventually become the dominant ones. But right now, the lion's share of that market is still happening in the 5G space. People are continuing to invest in things like non-terrestrial networks and satellite communications; open radio access networks or O-RAN, you may hear that term; stand-alone versions of 5G instead of the versions that most of us are using today, which ride on the back of the 4G network, getting to a truly stand-alone 5G network; other use cases as well as the performance enhancements that are going to come in future revs of the standard. So reminding everybody, if you go back to 3G, 3G, 3.5G, 3.9G, and then we went to 4G, and it was LTE, LTE-Advanced, LTE-Advanced Pro. We're still in the very first instantiations of 5G, and there's a lot of evolution that still has to happen before we're going to be ready for 6G deployment, which we're probably looking at commercialization sometime in 2029. People are pointing to the L.A. Olympics in 2028 is where you might see some early-stage trials. And so for Keysight, you'll probably back up a couple of years from that and say, "Hey, we're going to start to see meaningful ramp of our 6G opportunity in the '26, '27 time frame," is kind of the time frame that we're looking at. And so you're going to see this evolution from 5G to 6G over that window in time.

And you're in the lab, so it's not like we're going to see it in deployment?

We're primarily in the lab, that's right.

Yes, exactly. I still can't get a good signal in parts of Manhattan, but that's a different discussion. Let's talk about semis, and you already mentioned there's some green shoots and there's some opportunity there, and orders are getting a little better, but it's still early days and one day at a time kind of thing. But remind us what parts of semiconductors are you most exposed to? What are your key end markets within semis?

Yes. So I mean you can really think of ours -- our primary customers of the foundry customers. And it's -- I get this question a lot. What are you seeing in memory? What are you seeing in logic? And the answer for us really is we're agnostic. We almost -- our solution is totally independent of what's actually on the chips themselves. We make a solution called a parametric tester that's really a check on the foundry process itself and has the fab essentially maintain process control over the manufacturing environment. And so I mean, I can tell you, we have seen some -- you can tell a little bit, some relative strength in memory and other places, but for Keysight, we're really agnostic to what's on the chip itself. It's much more of a fab process check.

They must think you're KLA. That's what -- software and services, and it's interesting because a lot of companies, particularly equipment companies, have been building their software and services franchises. It's up to about 29% of your mix. It's obviously a good margin, sticky business. Where do you think that can go? Or if even Keysight has a target and say, look, over time, 5 years out, we'd like it to be this percent of the business? Or is it not -- it's more of just depends on placements and...

So first, let me just correct the record. In the most recent quarter, it was 39%, not 29%. So 39% of revenues, roughly mid-20s on software and roughly 15% on services is how we get to that number. And both of those businesses have been outperforming the broader business. And so I think what -- we haven't put out a target, so I'll start by that. We haven't put out a target where we think this heads other than we expect to continue to increase as a percentage of the overall mix. And I think what you'll see is on an organic basis, you'll continue to see kind of incremental improvement as those businesses outperform. But software, in particular, is an area of focus for M&A. I've obviously referenced the recent acquisition of ESI, and so I think there is potential for us to potentially accelerate that mix shift via M&A as we find targets, particularly software targets.

Okay. I'm going to go once more back to the room, and then I have a couple more questions.

One there and one here.

Go ahead. You're next. Go ahead.

Yes, I got to amplify the signal, I guess. But yes, so I guess like on that note of the software attach rate, like I'd just love to get -- when you have like newer maybe hyperscaler customers in the wireline segment, like is there kind of a difference in maybe the appetite for these guys to have that higher attach rate in the software side?

Yes. I think the answer to that question is yes, but maybe for reasons that are different than you're thinking, right? I think as our business has evolved, we're providing more complete solutions to our customer set, and so you think of the hyperscalers, obviously, when we're helping them emulate data traffic, that -- there's hardware as part of that solution, but there is also a significant software component to the offering that we have and a required component to address the specific problems that they're working on. And so I think that's what we're looking for is where are the points across -- we did the same thing in 5G, where we helped folks that are developing modems and devices emulate how those modems and devices will perform on a network, but in a lab environment. So we essentially recreate the network environment in a lab, and you could do the same things for folks who are working on base stations or via O-RAN for part of the base station solution. You can emulate the data traffic so they can see how those -- how that network tool will deploy -- will behave when flooded with network traffic, but again, in a lab environment. So this trend towards simulation and emulation, more broadly, I think is helping these customers, whether wireline or wireless, all of the complex challenges that they're looking to solve.

That's a lot more just specialized to their own needs.

To the needs of the industry. That's right. And there was a question in the back over here.

I just want to circle back on semi. So when you talk about the exposure to foundry, would you be able to elaborate? Is it more on legacy foundry or advanced foundry? Or should we think about analog manufacturing? What type of foundry?

Yes. I would say our differentiation is highest at -- higher at the more modern process architectures, right? There are more -- there are other ways to address this at, say, 45 nanometers and above. So more modern process architectures, our differentiation is higher, and therefore, the value that we bring to the foundry customer is higher.

Again, like KLA, on process control and advanced nodes. I'm going to switch gears to the financial model because we wouldn't have the CFO here if we didn't ask a couple of financial questions. So one I want to ask about gross margin, the other about just sort of revenue growth. But in gross margin, I was looking -- so last quarter -- well, let me say this. Margins have been very resilient. Despite sales declines, margins have held up in like 64% last quarter, not far off the peak from last year in the mid-60s. So how do you think about margin resilience? And how do you think about margin expansion going forward when we do start to get the recovery? What are the kind of puts and takes around margin?

Yes. So I think the 2 biggest impacts on margin for us are, in fact, volume and mix, right? And that can be -- volume is obvious, right? We've seen a low double-digit core pullback in our business. And then the mix side, it can be not just the mix of software versus hardware, but even within the hardware portfolio, there can be a substantial difference in various components of the portfolio. And so we've seen both of those things impact our business unfavorably over the course of the last couple of years, but as you've noted, the margins have been pretty resilient with roughly the peak to trough, 200 basis points -- 200 basis point shift. I think we certainly see opportunities for us to get back to 66%, and then we'll see where we take it from there. I think it's going to be a function of recovery, and we're going to need revenues to get back to growth. And then we have seen some of the largest pullbacks in some of the higher -- highest margin portions of our business. And I think as those things recover, we're going to be well positioned to see those gross margins bounce back.

Yes. But then I would imagine semis is going to be one of the higher.

Semis is certainly a favorable market, yes.

Okay. Revenue growth. This was a challenging year. And by the way, I'm not asking you to endorse consensus. I'm just saying if I look at consensus, they had sales down about 9% this year and up mid-single digits over the next year. The company, I think, has been pretty clear to talk about gradual recovery, not trying to overpromise. So again, not asking whether it's going to be mid-single or not next year, but just how do you think about the recovery over the next 2 to 3 years? What are sort of the drivers of that? And the reason I ask is because if you look a lot of times, what happens is -- there's a saying on Wall Street, the bigger the dip, the bigger the rip. Like the -- if your sales are down 9%, it's not hard to get back to good trajectory of growth. So how do you think about sort of the -- not slope of recovery, but just how do you think about the recovery in next year? What again are the puts and takes that you see for Keysight?

Yes, I kind of see us kind of being a little bit industry specific, right? We've talked about AI is driving wireline, but we haven't yet seen that catalyst in many of our other end markets, right? So if we just step through them, again, I think we've seen a bit of a pause in 5G investment even in the R&D lab. Some of the big customers in that space are going through restructurings. They have inventory challenges. We've seen refresh cycles on cellphones slow. So I think as that customer set starts to get confidence in their own business, we'll see R&D investments once again start to inflect, and that will become a positive trend for our wireless business. In semi, we've seen delays in some of these big fab programs in Arizona and Texas and Ohio and other places around the world. I think as those fab investments get back on track, again, that's going to be a positive catalyst for our business. In our general electronics business, we generally watch PMI pretty carefully. It has a little bit of a higher manufacturing focus than many of our businesses. And I think there's a little bit of a watch for there, because I do think there was some overcapacity that got put into place in response to the supply chain. So we probably need a little bit of sustained -- to sustain PMIs above 50 for a while to allow that capacity to be absorbed and once again, spur new investments. The one area where, at least in the current environment, we've seen actually some incremental weakness is in auto, particularly with a lot of the press that you've seen around EV, pull back in the adoption rates of electric vehicles, lower-cost Chinese electric vehicles leading to some tariffs and other things. And I think we'll have to see how that develops. I think people in the industry are rethinking what that investment model looks like and a little bit of a watch for. So I think the point being that there are strong secular drivers, but these things don't always happen in phase, right? And the post-COVID recovery plus the supply chain, things were in phase. Kind of everything was up and to the right, but that's pretty atypical for our business. It's much more common that you have some things that are firing when other things are in maybe a little bit less favorable state, is a more common scenario.

I agree. I agree. Okay.

You have a question right here.

Oh, sure. Thank you.

So we have Michael Dell saying today that he's expecting a big refresh cycle in PC market. Many are also anticipating the same in smartphones towards the end of this year. Can you maybe comment on how Keysight is sort of geared into that and whether you share the view on a big refresh cycle driven by AI at the edge?

Yes. Well, I -- so I'll leave the commentary on the current refresh cycle to those that are closer to it, but I think we do see investment cycle driven by AI at the edge. Again, for us, it's primarily in the R&D lab. We think that's a necessity. As you think about how AI is happening, a lot of that stuff is going to need to happen at the edge device, and so we do see that as a significant opportunity for us, again, to help our customers with that primarily in the R&D lab. I think as -- if they are correct and that you see a significant refresh cycle both in PCs and in handsets, I think that will be, net-net, good for business. I mean there's evidence to suggest that chip inventories are still increasing. I think that's a headwind for some of these businesses. If we can start to get inventories under control, get confidence in, again, the business performance, these are -- many of these companies are our customers, they will once again start to reinvigorate their investments in R&D, and that's what we really need to see.

Can you talk about the -- just touch a little bit on the Spirent transaction, sort of the strategic rationale, why you're buying it, what you get from it?

Yes. So we're somewhat limited, we could say, given the U.K. takeover panel rules, but this is a business that is exposed to the same broad market themes that we're exposed to around wireless and wireline and automotive, even some aerospace, defense. And so there is a pretty significant overlap with our selling and sales and marketing infrastructure, our ability to leverage those to drive incremental revenue as well as some cost efficiencies through those line items. There's some significant market expansion that's attractive to us, both in terms of their network assurance business and their precision location business. Those things would be net additive to us from a market perspective and extend our reach within the broader commercial communications ecosystem. And I think if you step back at a high level, we saw an opportunity to buy a business that ultimately we think could be accretive to Keysight's gross and operating margins and create value for shareholders.

I'm just going to tag on to that and ask about your capital allocation strategy, just how you think about -- particularly, you've done a lot of deals over the last decade, but how do you think about returning cash to shareholders at a balance sheet?

Yes. I come back to kind of Satish's third comment from our recent earnings call. We continue to remain focused on value creation and operating the business as efficiently as possible, and I think capital allocation goes with that. And our capital allocation priorities have not changed. I think one of the things that we're proudest of as an organization is that we have been able to sustain our investments in R&D over the course of this down cycle, and that's our #1 capital allocation priority, is to sustain the organic growth of our business. And the fact that R&D is basically flat this year, even in the face of a 9% decline on the top line, I think it's positive and positions us really well as we look forward to fully capitalize on some of these technological transitions we're talking about, whether it's AI, AI at the edge, 6G, autonomous driving, quantum computing, these are all areas where we're investing. I think beyond that, we're looking to strike a balance between returning capital to our shareholders and doing M&A. We've talked about some of the successes that we've had in M&A. We closed the ESI transaction at the beginning of the year. We did 2 smaller tuck-ins, a company called Riscure, which is chip level security, and we did a small communications tool provider called AnaPico in Switzerland here in the most recent quarter. But at the same time, we're actively returning value to shareholders as well. Over the last 4 quarters, we've returned about 78% of about -- we returned 78% of free cash flow to shareholders via our buyback program. And so again, looking to strike that balance with, again, an emphasis on continuing to make the investments that are necessary to drive the organic growth of our business. That's the #1 priority.

Okay. I'm out of questions, so you and I are the only thing standing between these people and a bar. So...

All right. Anybody else?

Any last questions? If not, we're going to call it here. Thank you very much. Neil, thank you for your time. It was awesome. Jason as well. And you'll be around for a couple of minutes. And if not, thank you, everyone, for joining day 1 of our Citi Tech Conference. We'll be back tomorrow in force, and I think we even go through Friday. Thanks, everyone, for joining.