Maybe investors don't appreciate about the story. Where should we spend a little bit more time kind of digging in?
Sure. Yeah, yeah. It's a broad question, but I guess I would start with the infrastructure business. That's probably the area of the most interest right now, and it's something that we've been investing in for the last probably six to seven years now with our wireless infrastructure products as well as our optical and data center products. It's kind of fallen in that category. That business will do, I'll call it, $160-$170 million this year. It's been all grown organically in that time period, and it was something that really broadened our capability, leveraged on existing technologies that we had into some higher growth, exciting markets, and we've made really good traction there. The wireless infrastructure hasn't been like this year has been not as exciting as telco spending has been down, so you're seeing that impact, but that's expected to recover next year.
And then on the data center side, lots of excitement around DSPs and data center build-outs on these transceivers.
Yeah. Okay. Well, maybe we'll kind of double-click in there. We'll start with infrastructure. I think you talked about some exciting opportunities in last call, earnings call. You were talking about some of those. Maybe walk us through some of the growth opportunities in that segment and kind of rank order some of them of importance and size of how we should think about looking into next year.
Sure. Yeah. So, I mean, the one that maybe I'll start with optical. So, look, optical this year we said we'd do greater than $30 million. That was kind of earlier in the year. We thought it would be a little bit lower. We upped that number a bit. This is driven by our 5-nanometer Keystone product. This is going these 800-gig data centers that are being deployed right now. We're still selling some 400-gig and 800-gig, but the 800-gig data center opportunity is really happening kind of late this year and really even in 2025. Some of the bigger hyperscalers are just going to start to deploy in the first half, and then there's a couple even in the second half of next year. So that's what's driving this.
The DSP market as a whole, I'm sure you know this, Joe, but for everyone else, it's a market that did probably $800-$900 million last year. It's expected to double over the next couple of years. So really exciting opportunity as AI rolls out. You kind of get the front end as well as some of these back-end networks around the clusters themselves where all of these transceivers play.
So I guess as we look into next year, things starting to really ramp on the 800-gig side, like $30 million could go to what?
Yeah. I mean, I think this year we expect probably exit the year somewhere between $12 and $15. I think a lot of folks think next year that number is probably $60-$70 million. Pretty conservatively, kind of depends on how fast some of these 800-gig data centers roll out. Maybe there's a little bit of upside if you see more rollout next year. Naturally, we're working to gain more market share. I mean, our goal here is to win 20%-25% of this business. And so that's a big upside from us. And we're winning those sockets right now, and as they move into production, they're in various forms of qual right now. Some are out of qual and in production this quarter.
Some are just entering qual, or some of them are months into a qual that expect to come out in, call it January, February, go to production in Q2 would be another example.
Okay. And that qualification time frame, kind of, is there an average we should think about, or it kind of depends on the customer, I guess?
It really does. I mean, some of the more conservative guys, I mean, it could be six-to-nine months in some of these soak tests, but then you've also got some of the other, I'll call it mid-tier guys that might do it in two-to-three months.
Okay. And I guess, is there any, in terms of as you kind of get more proliferate and into this business or just kind of get bigger in size, do you expect that to change more or that some of those longer cycles become shorter?
The qualifications themselves, I mean, are still going to take that long, right? I think as we get bigger, and we've seen this even over the last 12 months, we're working with a lot more customers today than we ever have. Now you've got more qual slots underway. The other, I think, interesting thing is as you work with a lot of the module vendors, once they get you qualified at XYZ's data center, once that one gets done, then they want to go off and sell that to three or four other customers that they have. They can pitch it as, "Hey, they've already been qualified at XYZ data center," right? Then the things go a little faster. Then it's kind of like a domino effect.
We can start to really see more business fall quicker because some of those quals have been done. The qual itself may still take the same amount of time, but hopefully sometimes they'll rely on somebody else's data.
Okay. Okay. That's helpful. And maybe it's a newer opportunity for you guys. You're going against some pretty sizable competitors, right? Can you walk us through just kind of what's the competitive differentiation and why are you winning?
Sure. Yeah. Look, I mean, I think the simple and short answer is power. I mean, that is the differentiation. That is where the pain point is for most all of these data center guys as you're hearing them buy nuclear reactors and the like. I mean, there's big dollars being spent on a lot of this infrastructure, but a lot of it comes all the way back to power, and so the more we can reduce the power, the better. That's really our differentiation. Our architecture itself is differentiated. The nodes matter as well. We have jumped ahead on having a 5-nanometer solution, so we've definitely got an advantage there. But that's going to keep progressing. I mean, five will go to four, will go to three, and so that's going to continue, so our real differentiation is power. We've got to continue to support the customers.
We've got to differentiate that solution. Cost-wise, we're competitive. I mean, this is not a business that you're competing on price, and so look, right now there's really two big guys in the space, one of which owns probably more than 50% of the market, and they're focused on really big opportunities, so a lot of times they're walking away and not giving the proper amount of attention to some of these other, call it tier one or tier two data center providers, and so that really opens up the door for us if we can jump in and have the right solution.
So I mean, if we fast forward a couple of years and you kind of think about where your market share is today versus longer term, I mean, where would you fall out as being, "Okay, that was a success"? And what's your kind of expectation for a longer term kind of market share dynamic?
Sure. So we've always had a goal of being at 20%-25% of this market. We think this is very achievable. And a lot of these applications where you're going into a data center, often they'll split the business between two people. If it's a bigger data center, maybe you see them split between three people. So I think it's very reasonable to see 30%-50% in some of these bigger data center opportunities. So really we feel like that 20%-25% share is very reasonable. I think the other thing to note is that often in these data centers, they require a massive amount of work just to get through interop, to get through quals. And so they're not going to spend that time if they're not serious about using you in a serious way, right? They're not going to give you 5% share.
They're going to give you 30%-50% share in that opportunity.
So as you think about the visibility into that business, and I mean, what's the type of kind of you talked about the qual cycle, but the lead time or visibility that your customer is saying, "Hey, look, we need to have this data center online Q1, Q2"? How do you think about that and forecasting the business?
Right. So it's tough. I mean, it's tough right now because as you can imagine on the data center side, they themselves are kind of waiting to the last minute. They want to order within six-week lead times. They want to make sure that they've got their power supplies in place or whatever situations that they have on their side so they'd like to order later. We've got, call it in this area, it's probably 20-plus week lead times. And so you can imagine, so right now we're really booking to finish up Q1, Q2 of next year is really where our heads are at, and we're pushing customers to place those orders. If they don't place those orders, often they're at risk as well. And we typically nor do our competitors, but you don't hold a lot of inventory here.
Sure. Is there any from a component standpoint or manufacturing capacity for you guys that we should be thinking about as being potential gating factors to upside, or how should we think about that?
Yeah. I mean, look, our primary foundry that we use for these products is TSMC, and I mean, the volumes aren't huge, so I think we feel comfortable. I mean, that's the major gating item, and I shared with you the lead time. So the amount of volumes are manageable, and so it hasn't been that much of an issue.
Okay. Okay. Maybe shifting gears a little bit to the wireless infrastructure side. I think that's been an area of challenge, right, just given telco CapEx. I think last earnings call, you guys talked about maybe that seeing some signs of improvement looking into next year. So can you talk about kind of what gives you that confidence, and how should we think about just the visibility there?
Sure. Yeah. So you're right. I mean, it has been tough over the last, call it, 12 to 15 months. We're starting to see some signs of improvement. You're absolutely right. We have just kind of two pieces of that business: our access business, which is mostly transceivers, and then the other side is backhaul. And backhaul can be a little bit more spotty. I mean, it kind of ebbs and flows a little bit. Access is a lot more like 5G rollout, big deployment, big CapEx dollars, right? And so we went through that, what, a year and a half ago, and we've seen that moderate a little bit. But I think overall we're starting to, I think 2025 will be a little bit better. 2026 will be a little bit better than that, right? So I think we're seeing the signs of improvement.
I think on the MaxLinear side, what can we do? I mean, this is what can we do to grow our content, to grow our revenue dollars kind of despite the market, right, so from a market share standpoint, I think we're doing a good job, particularly on the access side. That's a smaller piece of the business, but it's a big market opportunity where we're up against a couple of bigger players, and I think we are doing a really good job of, one, taking share and, two, adding to our content. So we've had a transceiver historically. We have a new product called Sierra. It has a digital front end, so now we're more than doubling the content just in that one application.
So that really opens up a lot of doors and enables us to grow revenue faster kind of regardless of whatever telco spending does as long as we've got more content. Same thing applies to backhaul. Backhaul, we had a, I mean, we typically had a modem over the years, and then we've added a transceiver. And so some of the growth that we saw in 2023 was due to that added content. And so that's something else that we have in the toolbox that we can go off and use.
Okay. Okay. Maybe sticking in the infrastructure side still, storage accelerators. I think you've seen some momentum there from the Panther III. Can you talk about just that business long term? How meaningful could it be to MaxLinear? What's driving that business from just a secular standpoint?
Right. Yeah. No, so storage accelerators is kind of a new area for us, and it is really kind of exciting right now with what you see going on in the data center, enterprise applications as well, where they're using just a lot of processing power is being used out there, so this accelerator, you could think of it as almost like a compression engine as well where you're offloading from some of the existing processor cores. So this enables ultimately that customer to run more efficiently, more cost-effectively, and offload some of that data, and so we've got a unique offering. It's done in hardware versus some of the competitors have done this in software. Like Intel has a solution, for example, in software called QAT. This kind of takes that to a whole new level. It's done in hardware. The performance that it has is much, much better.
And so we're getting lots of interest. And I think this is a lot driven by some of the demands that they have on just the amount of data that's being processed, right? So it's been a pretty small business. We've been in it for the last probably five-plus years, but we have a new solution that came out this year in the spring. We're getting more traction in it. You ask how big can it be. We've said it can be $50-$75 million over the next, call it, two years. And historically, this business has kind of run $10-$15 million. So there's some decent upside here over the next couple of three years as more and more customers want to adopt this technology and moving away from the software solution to the hardware solution.
We've done a fair amount of work with AMD as well. They have a lot of stuff that they work from a reference design standpoint, a lot of things that they offer to their customers through applications, through their website. And so they love it because it's been a nice extension of those processor cores. And so if they can't get the customer to buy more processors, then this is a nice solution for them.
I guess the differentiation for hardware versus software, why?
It's just more robust. I mean, you can offload more data, is kind of the simple answer to that rather than in software. You just can't move as much data.
Gotcha. Okay. Okay. I think maybe just to round out the infrastructure industrial or I'm sorry. Well, maybe just we'll, sorry, we'll bounce back to that in a bit, but broadband connectivity, talk about that. I mean, I know that's been an area that's been a challenge, right, coming out of the pandemic. I think revenue's down like 80% peak to troughs. How do we think about normalized revenue there, and when should we kind of think about that being part of the story?
Yeah. So look, we definitely saw kind of COVID period, big uptick in 2021, and then kind of supply chain crisis hit in 2022. So we've seen six, seven quarters of softness here as they work through inventory. There's a lot of build ahead here that happened, and so we've been burning off that inventory. Things have been getting better. I think we've made a tremendous amount of progress. I think the inventory itself is probably no longer a problem. I think we're bottoming out on that front. We're starting to see bookings and backlog improve. The outlook is good for next year. I think a lot of the folks have this growing north of 30% next year. Now, you ask specifically when does it get back to the previous level? Look, I don't.
Normalize it.
Yeah. Normalize. Okay. That's fair. I prefer to talk about normal because I think that peak that we saw in 2022 was a bit artificial. And so the answer is somewhere in between. It's not as low as it is right now, but it's nowhere near what we saw in 2022. I think so there's a couple of pieces here in broadband. One, we continue to see spending. I mean, CapEx dollars are going out. You've seen the telcos continue to be relatively aggressive as far as crossing more homes, things like that. Keep in mind, we're not a, I mean, I think of us, we don't really care who succeeds, whether it's fiber, cable, fixed wireless access, any of those guys can succeed. We typically do this gateway that sits in the home. It has a processor in it. It has maybe an A- to- D converter in it.
It has some power products in it, Wi-Fi, Ethernet. So there's $40-plus of content in that box. There's a few things going on. I mean, one on the PON side or the fiber side, this is a market that we've been really underexposed to over the years. It wasn't a big market for us. It was less than $10 million a few years ago. I think this year it's probably a little north of $50. I think this can be a $100-plus million-dollar product line over the next few years as we win more share as the market deploys further. We've seen a lot of success. On the cable side, I mean, cables also having to compete with fiber. They're going through an upgrade cycle right now, which is kind of interesting.
Most of those CapEx dollars have actually been going towards infrastructure, meaning like the head end. They've got to upgrade the nodes. They've got to upgrade the network itself before the home gets kind of facilitated, right? So the good thing is a lot of that money's been spent. So now we're going to start to see that money switch over to more CPE and upgrade that piece of the puzzle, which is really where we play. So I think the inventory situation's behind us. We're also seeing some upgrade cycle with the technology itself. The other piece of this, aside from the processing or the modem capability of what's in that box, you've also got Wi-Fi, and Wi-Fi falls in the connectivity bucket. And Wi-Fi is happening. You're seeing a recovery happen there, but you're also seeing Wi-Fi 7 start to come into play.
I mean, we'll see more of that ship in 2025. Most of the work that we're doing today is getting more penetration on Wi-Fi 7. The good thing there is that it's our content increase. And so aside from unit increases, we also see a content increase from Wi-Fi 6 to Wi-Fi 7. So maybe it goes from maybe $12-ish dollars downwards to $15 per box. And so that's something that really we control. And then maybe the other product to point out is Ethernet. Ethernet falls into that connectivity end market. Ethernet, the reason it's in that end market is because in each of these gateways, there's an Ethernet port that's being upgraded from one gig to two and a half gigs. So you're seeing a content increase there. We're also gaining more share on those gateways.
What we've also kind of done with our sales force over the last two years is really to branch out beyond these broadband gateways, and we've gone into the industrial markets. We've gone into the enterprise market as well to get more traction because we've got a two and a half gig solution that really works perfectly for them. There's billions of these ports out there that need to be upgraded, and so we've seen a lot of success with some of the bigger players in the enterprise market as well as in the industrial market, and that'll help kind of complement what we're doing on the gateway side.
Maybe to back up a little bit, kind of walk through some of the broadband business. I mean, DOCSIS, you kind of talked about we're starting to see maybe some upgrades. The move to DOCSIS 4, I think it's a slow process, right? But I guess how do you think about that over the long term? And then can you remind us of the content uplift there as well?
Sure. Yeah. No, those are really good questions. See if I can hit them all. So there is a DOCSIS 4.0 transition happening. Most of the market has been DOCSIS 3.1. Now, there is a. I'm going to call it kind of an in-between approach that's DOCSIS 3.1 Ultra, which is also an upgrade, and it is also competitive with PON. And so they can kind of get away with somewhat of a lesser approach. For us, the content increase is still very good, and it is an upgrade. DOCSIS 4.0 as well. It's running upwards of 50% increase in content. The rollout timeframe that you talked about, you say it's running slower. I would say, so DOCSIS rollouts never run fast. But I would say that it's happening more or less in line with what our original expectations were. So we have that coming in into 2026.
I mean, you'll see a little bit of revenue next year, but it won't be a big number, so we really counted on this in 2026. There's been some rumors around a slowdown due to some of the technologies, FDX versus extended spectrum. I wouldn't say that that has been a big hindrance. Most of the problem has been on the network side or on the node side, getting that infrastructure upgraded before the other end can be rolled out, so the good thing is for us, I mean, we continue to sell more DOCSIS 3.1, DOCSIS 3.1 Ultra, ultimately 4.0 in the timeframe that we're expecting, and again, this is not a market where unit volumes are huge, but if you're seeing a 50% increase in content, it's a big deal.
Right. Right. Okay. Okay. Maybe on the PON side, I mean, I think that that's an area that you've got another large competitor, right?
Sure.
But you've talked about strong design-win traction. So again, can you kind of help us understand the value that you're providing to the customer that's driving those wins?
Sure. Yeah. So look, so maybe first of all, I'm going to frame. PON is about two times bigger than the cable market, and it's a market that we've been underexposed to. So it's an exciting area for us. How do we differentiate? I mean, a lot of cases, I mean, it's processing power. It has been a big differentiator for our SoC. That's why we've been winning. But you also need to have a firmware software capability that sits on top of all these chips. You really do need to have a comprehensive solution, and we do. And so that's been compelling. The largest player in this market has been Broadcom. I think we've been making good progress against them. I mean, they're never an easy competitor, but I think customers really have seen us.
We've competed head to head in the cable market for a number of years and more or less split that market with Broadcom. On the PON side, they've really been a much more dominant player, not necessarily because of technology, just because there hasn't been anyone else there. So over the last three years, I think we've made really good progress in penetrating that market, proving to those service providers that we can be a player here.
Okay. Okay. Maybe switching over to connectivity a little bit. We talked a little bit about the Wi-Fi 7, but can you remind us how do we think about the ASP uplift that's associated with that relative to Wi-Fi 6 and Wi-Fi 6E?
Right. So I would say the majority of revenues this year will be around Wi-Fi 6. So it varies a little bit, but I mentioned the $12. Somewhere around $12. There are lower-end solutions that run lower than that, and there's some more higher feature sets that'll run higher than that. But that's Wi-Fi 6. Wi-Fi 7, on average, I would say it's $15. Again, there too, you've got some low-end solutions in Wi-Fi 7 that might be $12 or $13, but you've got some higher-end ones that are well north of that. I'd say one thing that we've seen in the market, we've seen kind of a typical cyclical dynamic where in the boom times you're buying the really high feature set. And now I would say we're kind of mid-tier now is really what's playing well, and that's where we're seeing more of it.
Okay. Okay. Maybe we'll shift gears a little bit. I think just quickly on industrial and multi-market, I mean, that's been an area that's also had some corrections, but everyone's been seeing industrial corrections. I mean, where do we kind of fall out for that business in 2025? Can we see flat revenue growth in 2025? How should we think about that?
It has been tough. I think it continues to be tough. I think if you look at a lot of the kind of analog peers and industrial, I mean, it's been pretty rough. And I would expect that there's probably another six or eight months of that. I think from our perspective, I think we do see growth next year. I think it's low single-digit type growth, which is somewhat normal for this type of business, is kind of somewhere in the GDP range, 3% or 4% a year. I think that's what we would expect in this business next year.
Okay. Any sort of end market that's maybe driving that a little bit more?
I mean, that's the biggest one. I forgot to mention, and it might be worth maybe clarifying. We did have a bit of a step down. We've had some products that were shipped into China that were subject to some export issues in Q2. And there's probably $15-$20 million worth of business that'll come out of that number.
Yeah, the comp.
That won't come back. So that has been a challenge. I mean, that's more of a geography, not a specific application. But from a geography standpoint, this has had more exposure into China. And China's been soft, right? So hopefully we'll see a little bit of recovery there next year, kind of despite some of the geopolitical challenges that exist.
Okay. Okay. Maybe shift gears a little bit. Let's talk about the target model that you guys have put out there. Cycle obviously has been hurting the ability to kind of hit those targets. But I guess talk about your confidence level as kind of demand or the business normalizes from inventory reductions and things, like getting back to kind of some of those targets of, I think, 65% gross margin, 35% op margin. Maybe just walk us through some of the drivers there.
Sure. Yeah. So gross margins have been something that we've always had a lot of focus on. Our goal is to get to the mid-60s, as you pointed out. We kind of held in. I mean, we were doing not quite mid-60s at a little north of $1 billion. Now revenues are down quite a bit. But that said, we held in at right around 60% gross margins that kind of dipped into the high 50s in the most recent two quarters with the lower revenue run rate. So revenue has definitely had an impact there. But I feel confident that we can get back. Hopefully next year, we're exiting the year at 60% and then start to kind of grow from there. So I'm confident that we can get up to that 65%.
One other maybe comment on the mix, because I was kind of talking more about the scale itself and the impact that it has. I mean, we're a fabless company, so it's not all about scale. A lot of it is about mix. Our infrastructure business, which we've talked about, has the higher growth opportunity in the company, and those are markets that have higher growth engines, so that actually has a higher gross margin to it, so our optical business, our wireless infrastructure business all have higher than corporate gross margins, and so as they grow and become a bigger percentage of revenues, I would expect that'll start to influence gross margins also.
What about content growth in terms of the mix of higher content solutions that you're offering in terms of gross margin additive? Is that gross margin additive?
It is. So the couple of examples I gave with the modem and the transceiver, the same thing with the DFE. I mean, those will be higher gross margin products. Now they happen to fall in the infrastructure end market. But if you can add more silicon onto that same product, I mean, more than likely it's going to be a little bit better gross margin. That's correct.
Okay. Okay.
I was just going to go back to your question on the model. From an OpEx standpoint, yeah, long-term target of being 35%. I think that's very reasonable in the markets that we play in. Infrastructure, it's been a little heavier spend as of late as we get more traction. I mentioned that we've grown this business organically over the last six years, so there's been a heavy spend. They're big markets. Data center, for example, also needs a pretty hefty spend on masks, for example. We have to definitely invest there, but I do feel confident that we can get our operating margins back up north of 30% over the next, call it, two to three years as revenue improves, as spending comes down. We've also had, so we kind of have two sets of markets.
I've got high growth engines like data center that do have a pretty heavy spend, but it's got a big growth driver, then you participate in other markets that have seven-year product life cycles. You're doing a chip every seven years, so we don't have to spend as much there, and so we've got some spending that's actually going to cycle down over the next two years, and so I think you'll start to see us drive operating margins up from here and get back north of that 30%. So as the revenue improves and then as the spending comes down, we did some cuts. I mean, it might be worth mentioning. We did some cuts in Q3. It was about 20-25% of OpEx. So it was a sizable cut, but it really follows the investment cycle that I mentioned.
With the lower revenues, naturally, we want to get our cost structure aligned. But we're also at the tail end of a big spending cycle. Wi-Fi 7 chips were completed. Our Rushmore chip is kind of coming to the tail end. That's our 1.6T chip. We've got our PON chip just completed. Our DOCSIS 4.0 chip completed. So that's where you had, whatever, 200 people of ASIC designers, well, now I need 20 application engineers to go win this business at the customer. So we're in a much stronger position, and I don't think we give up anything on the growth side.
Is there anything to think about as we kind of go into 2025 and variable comp things reset in terms of increases there?
Yeah. So, well, okay, if I think of the kind of shape of spending, so we certainly see payroll taxes and comp increases in the first part of the year. So that is correct from a shape standpoint. But I don't think there's anything hugely meaningful. I mean, from a stock comp standpoint, I mean, we're doing that in equity, and that would be what you would typically expect.
Yeah. Some companies have talked about just as demand has corrected, so those kind of targets to hit, incentive targets to hit variable comp have reset lower. So if demand does improve, hopefully it does improve, right?
You're seeing more of variable comp. That's correct.
Yeah. Okay. Maybe in the 30 seconds we have left here, can you talk a little bit about just capital allocation, how to think about just the balance sheet and shareholder return?
Sure. Yeah. So right now, I mean, kind of post this correction, balance sheet, we got $125 million of debt, $150 million of cash. We did burn cash in Q3 with the lower revenue number and the spending. So that's one of the things that we're just laser-focused on right now is getting back to cash flow positive. And now with the reductions in headcount, we have had some severance costs, which hit cash. That'll improve. We'll start to see cash increase throughout 2025. And so from an allocation, we're not doing deals. We're really just focused on generating cash right now and improving shareholder return.
Okay. Perfect. We'll end it there. Thanks for joining us.
All right. Thank you. Appreciate the time.