All right, perfect. Good morning. Thank you for attending J.P. Morgan's 21st Annual Semiconductor Technology and Automotive Investor Forum here at CES. My name is Harlan Sur, Semiconductor and Semiconductor Capital Markets Analyst with the firm. Very pleased to have Chris Koopmans, Chief Operating Officer, and Ashish Saran, Vice President of Investor Relations at Marvell here with us today. Marvell is one of our top picks in the semiconductor sector for 2023, leadership in cloud data center, 5G networking with their networking compute storage and custom ASIC solutions and an emerging pipeline in automotive. Gentlemen, thank you for joining us today.
Yeah. Thanks.
I will go ahead and kick it off with the first few questions and turn it over to the audience, if you have any questions. Let's talk about the sort of full year outlook. As we enter calendar 2023, you know, there are macro headwinds pressuring certain parts of your business like storage, wired infrastructure, China enterprise networking. That said, though, the team has strategic idiosyncratic programs that are firing this year as well to offset some of the softness. Cloud spending is still growing mid to high single digits percentage.20 On the last earnings call, I think Matt had articulated a growth profile for calendar 23, albeit at a lower rate versus the team's prior expectations. Consensus has the team growing revenues kinda low single digits this year.
Help us understand within your growth outlook for this year, what are the programs and markets geographies that are going to be driving growth for the team?
Sure. Yeah, I think, you know, firstly, I think relative to the industry, I think we feel in a very fortunate position, right? Given we have good secular growth in most of our end markets. As you pointed out, we have a number of kind of, call it, idiosyncratic opportunities. Maybe I'll start with, you know, data center cloud in particular.
Yeah.
Where we've articulated a couple of years back, you know, we made a pretty big strategic shift identifying, working with cloud customers directly for optimizing solutions. I think of these as custom opportunities. We won a significant number of them last year, and we'll start ramping, you know, late this year, but certainly late last year.
Yeah.
You know, growing to about $400 million aggregate this calendar year. That continues into $800 million the following year. That's probably the single largest, you know, incremental revenue opportunity. Within cloud, we also launched our new 400ZR data center.
Yeah.
interconnect program that's ramping, 800 Gig PAM4, again, with an Electro-Optics portfolio, which is really being used by AI interconnect. We have a number of, kind of independent growth drivers. Our automotive business, as you mentioned, this has been an absolute home run. This is Ethernet connectivity. We just crossed $200 million annualized last quarter.
Yeah.
Going from essentially almost nothing, right, a couple of years back, and that, you know, continues to grow this year. Then our 5G business, we expect very strong growth, right? Driven really by broader 5G adoption. You know, India's new geography turning on, the U.S. certainly keeps going. Quite frankly, we have more content gains, which will kick in. You know, we'll see some this year, but certainly a lot more next year.
You know, one of the biggest headwinds, is your data center storage business due to inventory adjustments from your HDD and SSD storage customers. I mean, the team has noted that inventory correction should last a few quarters for the storage business. It looks like after two quarters of inventory drawdowns themselves for your HDD customers, they are anticipating nearline shipments to inflect positively in the March quarter. Is the right way to think about it that Marvell potentially sees the inflection one quarter later, meaning revenue inflection in your storage business in the July quarter? That's the first question. You know, what's the linearity of recovery back to that sort of $800 million annualized sort of storage run rate in your data center business?
Yeah, I think, you know, these things typically do take, call it, a couple of quarters to clean up. Call it our Q4 and our Q1 would be kind of in the same zip code, I would say, right?
Yeah.
Then you start to see recovery from that time period and then, you know, full recovery sometime in the back half of the year. I think it's kind of aligned with your timeline. I think the one big difference as opposed to past storage cycles where, you know, we had a lot more PC exposure...
Yes.
-and a conversion from HDD to SSD, where the revenue would go down, it wouldn't necessarily come back up. I think today it's a very, very different situation. 100% agree with what you're hearing from our customers. Fundamentally, the need for storage, for bit storage.
Yeah.
Within data centers is gonna continue to grow. You're facing a couple of quarters of clearly there was an inventory situation which has to resolve itself over a few quarters, but this is fundamentally a growth business, right? As we look beyond call it towards the end of this year.
Yeah. Certainly, as we, you know, we cover Western Digital, we cover Micron. Certainly, you know, they're still seeing from their cloud customers, you know, petabyte storage consumption in that sort of high 20s , low 30s, sort of year-over-year growth range going forward. I would agree with you.
Yeah. I mean, we would expect unit growth rate within HDDs.
Yeah.
drive side. On top of that, we have a very strong position on the SSD side, right?
Yes. That's right.
We've actually won new designs which are yet to launch. As PCIe 5.0 and PCIe 6.0 start to roll out over the next few years, that again increases our total content and actually grows our share. I think overall, our outlook for our data center storage business is very positive.
You know, the team, as we've mentioned a couple of times here, you know, set to drive $400 million of incremental cloud optimized, what we call cloud ASIC programs in calendar 2023, $800 million calendar 2024. Help us understand how these strategic programs are trending for your customers. You know, how strategic these programs are for your customers' future growth initiatives, whether these could get cut in a downturn. My understanding is that the ASIC pipeline includes custom GPUs, video transcode, AI acceleration, and some custom networking programs. Is there anything that could be missing from that list?
No, I think you nailed the big parts of it, right? It's a lot of sort of compute offload type of things.
In terms of how critical they are, I mean, most of these programs are replacing or are sort of next generation of existing programs. I don't expect that any of them would be the type of thing that were sort of, you know, science projects that they would just kind of move on from.
Right.
They're all very, very critical to their overall infrastructure and to them enabling them to deliver their capabilities to their customers. They're tracking very much on track, very strong, and we're really excited about them ramping this year.
Maybe just to add to that, I think, you know, you're not gonna do these large custom silicon programs really for small projects, right?
Right.
This is where you've essentially expended significant amount of NRE, working with us, plus a lot of internal resources. These really drive down their total TCO, right? That's what they really care about. In fact, if anything, I think the pressure on us to execute even faster has increased, you know, given some of the financial constraints that will be find themselves under. These programs are actually even more critical, you know, as they look forward.
Yeah. As we've gone through the most recent earnings season, even recently in discussions with cloud customers, I mean, you consistently hear them talking about accelerated compute initiatives, right? Which are highly strategic and things that they're not gonna get cut. Obviously, that's a reflection of maybe a lot of these cloud optimized ASIC programs that you're helping these customers to enable. 5 nm pipeline is beginning to unfold, right? The revenue ramp began sort of last year as a part of your cloud optimized ASIC ramps and some of your lead 5G customers. 5 nm contribution should be accelerating this year. In 5G in particular, you know, you have 5 nm wins with Samsung, with Nokia, with Ericsson. Your 5G business is now running at more than $600 million annualized run rates.
You know, India is expected to ramp aggressively starting this year, which should provide some benefit to you guys. The team guided service provider segment to grow this year led by 5G wireless. What are the other product or customer or geographic drivers of the 5G growth this year?
Yeah, I think, you know, first and foremost, it is the conversion across the world, right? Of, you know, what matters to us right now, given we weren't a very significant player in the 4G, right, cycle.
Yeah.
What matters to us is not the total dollar CapEx in aggregate, right? What matters is how much of that converts to 5G, and that's what's really gonna be benefiting us this year, right? Even in the U.S., where clearly last calendar year was one of the bigger years in terms of total spend, but you can't upgrade large geographies in one year, right? Typically, it's a four-year process. I think the U.S. is gonna be one of the bigger contributors as well.
Yeah.
To your point, I think India is gonna be a very large, you know, they've really held back really their conversion to 5G for a number of reasons, now they're ready to go. I think we should expect, you know, pretty significant growth in that particular space. you know, Korea, Japan have really been kind of already been deploying, so those are more think of those as kind of existing run rate businesses.
Yeah.
I would say U.S. and India was the one I would look at for this calendar year. I think we will also expect to see benefits from Europe, although I suspect that's probably more next year.
Yeah.
It's very important if you look at the breadth of our customers, and you mentioned two in particular, both Nokia and Ericsson. And I think, given what will most likely be a shift away from Huawei-
Mm-hmm.
You know, given they just simply can't supply product, right?
Right.
I think. Remember, Huawei had almost 30% to 40% share in Europe.
Yeah.
That's a very big amount of dollars available.
Yeah.
which will shift to customers where we are broadly exposed to. I think that's all in front of us.
You know, 5 nm in whole is gonna be a big, big revenue cycle for the team, right? Thinking, looking forward, you know, you've been putting in place your next generation 3 nm platforms to potentially commence designs for your merchant and cloud optimized solutions. Can you just give us an update on 3 nm? I mean, there's a lot of things to be put in place, right? Working with your foundry partners to make sure the manufacturing timelines are in place. You're developing libraries, IP, discussing with your customers. You know, give us an update there on 3 nm, and are you already engaged on internal designs for your merchant solutions and/or cloud optimized ASIC designs with some of your customers at 3 nm?
Sure, yeah. As you mentioned, you know, this is really a parallel development, right? I mean, we started our 5 nm development, getting all the IP ready and our test chips ready, you know, years ago. Turned those into products last year, which are then turning into revenue this year.
Yeah.
You know, last year we had our 3 nm IP test chip, and we were doing all the development to get all the different IP blocks ready. That's aligned both to our internal merchant solutions as well as what we know is needed for the custom cloud optimized designs that we're already working with our customers on. In many cases, it tends to be the cloud optimized designs that lead the cycle.
Yeah.
that end up being on the very, very front end of the cycle that have the most cutting edge of these, you know, giant compute types of solutions or networking types of solutions. The merchant solutions are also working right behind that, so it's really all of the above.
I'm gonna move on to the data center business, but does anybody have any questions so far? No? Okay. On the 200 and 400 gig optical upgrade cycle with your major cloud and hyperscale customers, you know, the Marvell Inphi team still owns, according to our estimates, 80% plus market share, right? Right now, it's driven primarily by the tier one hyperscalers. The team has confidence that your optical business is gonna grow this year, yet you're working through some near term inventory digestion with your cloud customers here in the January quarter. What gives the team confidence on the growth for the optical connectivity business this year?
Yeah, sure. I think first is on 200 and 400 gig, the adoption is being broad, but it's really three large customers...
Yeah.
Right?
Right.
There is still a fourth large U.S. customer, which is completely in front of us in the market. On top of that, there are obviously the next tier of customers, right? They will inevitably adopt Electro-Optics technologies in PAM4, it's just a question of time. That's certainly in front of us. I think, you know, there was clearly a much bigger inventory situation on the storage.
Mm-hmm.
Given the way that particular market works.
Yeah.
Which is really a commodity product. When the demand is there, right, there's obviously, you know, it's considered fungible, meaning whoever supplies it gets it, right?
Yeah.
Everybody rushes and makes products, and that's essentially what created that situation. You don't have that happening in the rest of the market, right? These are all one-to-one sole source relationships. I think, you know, out of, call it caution, you know, we did guide it down a little bit before.
Mm-hmm.
It really wasn't a big magnitude, quite frankly, right? This is still fundamentally the fabric which drives the total bandwidth capacity. We feel very good about our existing 200, 400 gig business. On top of that, as I mentioned earlier, we have started ramping 800 gig PAM4. In fact, we've introduced our second chip, right, which does that on the DSP side in a new process geometry, along with the accompanying TIAs and drivers. Then 400ZR, which is data center interconnects. When you add it all up together, even if you have a very conservative view on, call it, not getting much help from cloud CapEx this year-
Right.
We still feel very good about the overall business will continue to grow. Then of course, as you look even further out in time, you know, CapEx will certainly start to grow at a faster rate. I think we'll certainly get that acceleration again.
The other new product cycle and optical, and you mentioned this in your prepared remarks as well, is data center interconnect upgrade to the new 400ZR standard. I believe you're ramping into Microsoft now. All of the other tier one cloud hyperscalers have a plan to move to the 400ZR standard as well. Maybe you can provide some color on the revenue ramp profile, your share position, and how should we think about the adoption from the other tier one cloud hyperscalers as we look out over the next sort of 18 to 24 months?
Sure. Yeah, I think as a lot of people remember, you know, 100G DCI data center was kind of the first, you know, project, which was kind of a proprietary relationship with Microsoft and Inphi back that time, which we've now acquired. Then we pioneered the 400ZR, which is an open standard, clearly, you know, the lead customer clearly is the one adopting it. It's spreading, to your point, beyond that. I think one of the big trends driving it, if you think about it, is, if you think back in time, a lot of the cloud customers had these very large mega data centers, you know, very significant within a single location.
As workloads are moving more towards the edge, you're seeing a lot more proliferation of smaller data centers, a lot closer to where the end customer is.
Yeah.
That is now causing, hey, now I need to connect these data centers together. I think the overall usage of DCI, to your point, is spreading to multiple customers. For us, you know, this is gonna be in a ramp phase, I would say for multiple years. Certainly, I think the revenue is obviously, you know, at 400ZR, I would say is, you know, at the level or surpassing the year at 100 fairly quickly because it's a much higher ASP. Then we still have, you know, additional customer adoption, you know, completely in front of us. I think it's not just cloud customers which will drive for the next few years, but over time, you will see the same technology also being used in optical transport, right, in the broader telecom market.
This is a, really a long multi-year cycle is my point.
You know, the team has mentioned 800G several times, and it clearly looks like it's a driver to the business. What's interesting about 800G is that it's not being tied to... When we typically think about optical, we think about optical being tied to a particular switching platform, right? Your 800G platform is actually not tied to a switching platform, it's actually tied to accelerated compute clusters, right? Like Google's TPU cluster and a whole bunch of other different accelerated compute clusters. They tend to seem to be like the first adopters of these new generations of technology. First, is that true that the bigger consumer of 800G is, you know, strategic AI compute initiatives and not so much, you know, the tie between switching and optical?
Yeah, I think that's absolutely correct, right. You're right. Historically, the first use of Electro-Optics was the switch to switch layer-
Yeah.
Right?
Right.
Essentially, typically driven by the 3.2, 6.4, 12.8 cycles, right? 800G PAM4, to your point, is being driven completely by AI.
Yeah.
We've seen, you know, two very large customers driving a significant amount of growth. As you think about the overall CapEx environment, while I know there's concerns about deceleration, the reality is within that's just the big number you see outside...
Yeah.
Includes buildings and a whole bunch of other stuff. When you look at the spending profile, the priority which AI and ML projects get, it remains extremely high. I'm not surprised at all to see the very strong demand we've seen, right?
Yeah.
The strong push from customers to enable this product very quickly. Your assessment is spot on.
As we look out over the next several years, Co-Packaged Optics CPO solutions will initially roll out to support 51.2T switching chips and more significant adoption at 102 terabit per second family of switching products given power and cost savings. What are you hearing from your hyperscale customers on their Co-Packaged Optics development strategies, and can you give us an update on Marvell's Co-Packaged Optics portfolio?
Sure, I'll start. Certainly as you know, our switching and our Co-Packaged Optics initiatives, just like the rest of our initiatives for cloud, are really being done in very tight partnership with our cloud customers. Really we're trying to align to what they need. Ultimately what they would generally prefer to do is to do pluggable as long as they can, right?
Right.
Ultimately, the pluggable is just a, it's a very good technology in terms of how they actually run it and deploy it and maintain it in the field. There is absolutely, they're all looking at the move towards Co-Packaged Optics, and Marvell is working very closely with them in terms of which generation they're gonna do it, and actually whether it is at the switch layer first or actually is it at the AI, as you mentioned, with 800G first? Whether that's actually the first place that they're gonna really wanna start to look at Co-Packaged Optics. It's not yet clear to me that, for example, 51.2T switching is gonna be the first big move for Co-Packaged Optics.
Marvell's working very closely with each of our hyperscale customers, in terms of developing our technology to support those ramps.
Yeah. I think if you think about the whole discussion we had on 400ZR, I mean, that was fundamentally the first high volume production of a Silicon Photonics.
Yeah
... optical, you know, solution. We've learned probably more than a lot of other people have in terms of delivering this in high volume, so I think our ability to co-execute. I think the key thing here is you have to work very tightly with your customers and basically give them the choice, right? Which is, A, it's not an all functions or an end function. There'll be certain applications where they will wanna move very quickly to CPO. There'll be a bunch of mass market solutions where they wanna keep pluggable going as long as it can for obvious reasons, right? Upgradeability, you know, serviceability. There's a number of very good reasons why, I think that is what our customers like, the fact we'll enable both ecosystems simultaneously.
Your Innovium cloud switching product was targeted to drive $150 million in revenues in calendar 2022. Did the team hit this target? Are you anticipating growth this year in cloud switching? Secondly, you know, all of the revenues are coming from your Teralynx 7 platform, which is 12.8 terabits per second, which is equivalent to Broadcom's Tomahawk 3. As we look ahead, right, Broadcom is ramping Tomahawk 4, getting ready to roll out Tomahawk 5 next year. How should we think about Marvell's cadence and roadmaps on the Teralynx platform over the next several years?
Sure. Yes, we said at the last earnings, I think, that we were on track to hit that number.
Yeah.
... for last year. Ultimately it was supply constrained for last year.
Mm-hmm.
for the big customer that we were delivering to.
Yes.
Therefore it was sort of a back half ramp, and therefore, yes, you should absolutely expect growth in calendar 2023 over calendar 2022. That platform is doing very, very well. You know, when we made that acquisition, our initial focus was really to make sure that we were able to ramp and support with high quality and significant supply to that initial customer at the 12.8. Now again, similar to our other cloud optimized platforms, working very closely with initially that first customer and then other customers.
Yes
in terms of what their needs are. Their needs really right now are at the 52T, and that's what Marvell is working on. Aligned, not just the switching layer, but the 1.6T optics that go along with it, right? 'Cause what's a good having a 52T switch if you don't have the 1.6T optics that can go along with it? Our roadmap end-to-end is completely aligned with our customer needs, and that's really the cadence that we're working on is to initially our focus will be to make sure we hit exactly what that very large customer needs at that 52T.
Mm-hmm.
Of course we will be expanding beyond that customer after that, right? Ultimately, our cadence will be aligned to what our customers need and starting with that initial customer.
Before I move on to enterprise, any questions from the audience? On enterprise networking, you know, business is going through a strong inflection by product cycle refresh, new ASIC programs, and the strength of your U.S., you know, tier 1 customers, but offset by muted demand from China networking OEMs. Still, you know, the business nearly doubled over the last two years. Can you just help us understand how much of the growth is strong customer spending versus share and content gains and, you know, this is with your U.S. customers. How do you see your overall enterprise networking business relative to the team's outlook for total company growth this year?
Yeah. I mean, this has been a business which, you know, This is one of the original Marvell organic businesses, right?
Yeah.
In fact, Chris was kind of heading that business back then, right? This has been a, an amazing transformation. I would say the majority of the growth we've seen in the last couple of years has really been, you know, a combination of share gains, both in PHYs as well as in switches, as well as a pretty shift up in content, especially on things like multi-gig PHYs, with ASP significantly higher. I'm not talking percentage points.
Yeah.
This is multiples higher, right? That penetration continues. You know, the market itself has also grown perhaps higher than typical, but it, you know, it's still somewhere in the single digits, right? The end market.
Yeah.
from a customer perspective.
Yeah.
Partly because of supply constraints, right? I think they would wanna grow faster. They have the orders, but, you know, there's been multiple supply constraints, including Marvell. That's obviously starting to clear up. In the meantime, you know, we did see some of the China demand start to come down, which quite frankly, I think is a temporary phenomena.
Mm-hmm.
Again, we're not saying when it comes back, but it will. I mean, China is going to be a big. You know, this is an endpoint-driven market, right?
Yes.
We're talking about enterprise networking, and that is a very large population. It's a large economy. At some point, that comes back. In the meantime, you know, I think for now, we are now finding a position from a supply perspective to meet our U.S. customers' demands, and that's what the focus is. In the long run, what we would go back and say is, look, fundamentally, we think this is a low single digit to mid-single digit end market.
Yeah.
We will still do outperform that, right? Perhaps not to the same magnitude of the, call it 40%-50% year-over-year growth, 'cause we've obviously gained a significant amount of share. I would still say we'll outperform that call it low to mid-single-digit market.
Automotive is the next significant revenue opportunity for the team. As you mentioned, you're already driving, you know, $200 million plus in annualized revenue run rate. Here the team is leveraging its strong position in Ethernet networking and gaining share. Much of the auto products are manufactured on older generation technology nodes, where the market, I think, continues to be in tight supply. Are the supply dynamics starting to ease here in lagging edge technology, mature technologies? Longer term, you know, the team has talked about the auto compute opportunity as a multi-billion dollar opportunity. Maybe if you can give us an update on the progress there too.
Sure. Yes, you're right. Ultimately, we're gaining significant traction with our Ethernet platform. Our Ethernet platform has some in older technology, but it also has, you know, we go all the way into the sort of 16 nm generation with our Ethernet technology. What I would say is generally yes, the older nodes are still constrained, for sure. Having said that, because of a number of strategic moves that Marvell has made within our portfolio in terms of moving products between nodes, we actually have what we need to deliver at these older nodes in automotive right now. We've been very good actually throughout the entire time of not ever becoming a bottleneck for any of our automotive customers, and they've really appreciated that. That traction just continues to grow.
I think we've said we now have eight out of 10 of the top OEMs.
Yeah.
You know, really 100% of the OEMs out there have a plan to adopt Ethernet across the portfolio. They're now starting to move from sort of model-by-model or model-year type decisions to broad portfolio decisions, broad platform decisions across their entire set of models. We just see traction, if anything, accelerating in Ethernet. Now for compute, absolutely. I mean, that's a big long-term opportunity. We've talked about compute everywhere, and we approach the auto compute market the way we approach the cloud compute market, which is really a custom, you know, cloud optimized type of a model as opposed to here's a generic compute product, let's go out there and try to win, share, and compete. That's really not our option. We think some OEMs will adopt an off-the-shelf solution.
Yeah.
We think other OEMs will wanna build their own solutions. We think we're a great partner with all of our, not only our Ethernet and our storage and our automotive technology, but with all of our compute technology that we're delivering to these other markets such as cloud, we're an ideal partner to help them with that. It's a long, I mean, that's really a long-term play.
I think in the call it near to midterm, right, we're going to be adding $hundreds of billions of revenue very quickly over the next few years, right? I think for the next, call it three to five year window, I think our Ethernet market, you know, I think we've sized that in that timeframe, the entire market at about $1 billion, and we said be at least 50% of that market.
Yeah.
We have clear line of sight to a number which is probably north of 500, quite frankly. I think given our share gains, I think 50%'s probably a pretty relatively easy target. Outside that timeframe is where I think the compute opportunity starts to make sense, where there's gonna be large customers in the auto world.
Yeah.
You've heard them talk about this themselves, right? At some point, you wanna control your own destiny, right? If you think about a car of the future, your differentiation is gonna be based on your AI ADAS capabilities, and at some point, you want to have that software hardware stack controlled in a single place, and I think we can certainly help them with that.
You know, on the operations side, on the leading edge, you know, can you talk about wafer and advanced packaging constraints, right? Even with the industry slowdown, you know, demand for leading edge wafers and advanced packaging still seems to be tight and potentially more so for the Marvell team, just given you've got this, you know, several different 5 nm programs that are gonna be ramping this year.
You know, how is the team managing through this dynamic in terms of securing supply and also trying to manage the cost profile both near term and longer term?
Sure. Yes, I mean, look, the most advanced technology is always constrained. You don't have an excess of the most cutting edge advanced stuff. They're very careful about where they, who they partner with, these suppliers, in terms of making sure that they prove out the technology. In prior generations, Marvell wasn't there, right?
Yeah, right.
We weren't one of the leading edge, and so that was a little bit newer for Marvell. However, we have very deep, very strategic relationships up and down within these suppliers. Ultimately they've seen our vision. They believe in our vision. They've in some cases even talked to our customers and understand... I mean, again, since these are cloud optimized programs for specific hyperscalers, they know exactly where they're going, and so they're betting on us. Ultimately, we have the supply that we need. I mean, obviously we have to execute.
Yeah.
We have the supply that we need. We have the commitments that we need from our suppliers to deliver against the numbers that we've talked about and the programs that we've talked about and ultimately, enable this success for our customers. Sure, there are certainly cost constraints, but, you know, we've built this sort of point in time things. I mean, especially for advanced packaging, we've built long-term relationships with several of these sort of interposer substrate type suppliers that give us visibility going out years for the amount of supply that we need and with the cost structures that we need.
On the gross margin front, you know, team is driving 64% gross margin second half of last year. It's at the low end of your target gross margin range. What is the team's confidence level on sustaining gross margins in the 60% range should we enter into a sharper down cycle? On the flip side, coming out of this weak period, like what are the key levers that will drive the margins to the high end of your gross margin target range, which is 66%? Yeah. I think the reality is, mix is probably a bigger determinant, right?
Mm-hmm.
Of our gross margin. As our revenue has scaled up, you haven't seen our gross margin significantly change either because I think this is the whole advantage of being essentially a fabless company, right? Where, yeah, there's some fixed cost leverage you get, but that's a fairly small amount, right? I think we're very happy that we've been able to sustain our kind of 64%-66%. I think where we are at towards a little bit on the lower end is more a reflection of the mix, right? As storage is, which generally tends to be a pretty good gross margin product, is dealing with a couple of quarters and then once that resolves itself right.
I think at a high level, I would go back to, you know, the end market gross margins we discussed on our Investor Day, where I'd say carrier and consumer are the ones which typically tend to be lower, but everything else either in line or slightly above. I think, you know, we remain pretty confident. I think we've done a fairly good job of managing our cost input increases and offsetting them with our customers, but just offsetting them so we're not taking advantage of the situation, but also obviously maintaining our overall margin profile. I think our goal remains, you know, stay in that 64%-66%, and obviously the focus is, you know, drive revenue growth and drive significant operating leverage. That's what we've demonstrated over the last couple of years, and that remains the focus for us going forward.
Great. Well, we look forward to monitoring the progress and execution of the Marvell team this year. I wanna thank you, Chris and Ashish, for participating today.
Yeah. Thank you very much.
Absolutely.
Thanks for your time.
Yeah. Thank you.