Good day, and thank you for standing by. Welcome to Semtech Corporation's Third Quarter 2025 Earnings Conference Call. At this time, all participants are in a listen-only mode. After management's remarks, there will be a question-and-answer session. Please be advised that today's conference call is being recorded. I would now like to hand the conference over to Mark Lin, Executive Vice President and Chief Financial Officer. Please go ahead.
Thank you, Operator. Good day, everyone, and welcome. I'm pleased to be joined today by Hong Hou, President and Chief Executive Officer. Today, after market close, we released our unaudited results for the third quarter of fiscal year 2025, which are posted along with an earnings call presentation to our investor website at investors.semtech.com. Today's call will include various remarks about future expectations, plans, and prospects, which comprise forward-looking statements. Please refer to today's press release and see slide two of the earnings presentation, as well as the risk factor section of our most recent annual report on Form 10-K for information on risk factors that could cause our actual results to differ materially from those made on this call. Unless otherwise noted, all income statement-related financial measures will be non-GAAP other than net sales.
Please refer to today's press release and see Slide three of the earnings presentation for important information regarding notes to our non-GAAP financial presentation. The press release and earnings presentation also include reconciliations of our GAAP and non-GAAP financial measures. With that, I will turn the call over to Hong.
Thank you, Mark. Good afternoon, everyone. I'm almost six months into my tenure as Semtech CEO, and it has been a very productive period. My numerous engagements with Semtech employees, customers, suppliers, and partners, and the operational and financial progress we have made so far give me high confidence in Semtech's near- and long-term growth prospects. We achieved a very strong Q3 result with net sales, gross margin, and EPS at the high end of our guidance range, while operating income and operating margin exceeded the high end of our guidance range. Further, our Q4 outlook projects continued growth in each of these metrics. In the last earnings call, I shared three near-term priorities, and I'm happy to report that we are making progress on all fronts. First, on strategy, portfolio rationalization, and balance sheet improvement, we have completed an evaluation of our portfolio through our annual strategic planning process.
The purpose of this evaluation extended beyond a delineation of core and non-core assets but provided a more granular assessment correlating investment levels and priorities to multi-year growth curves. In what might be a generational opportunity stemmed from AI-driven product demand, we'll be increasing investment in data center products, which we project to be a long-term and transformational growth engine for Semtech. In some other areas, we expect sustaining investment at the current levels will suffice, and our focus is to improve the contribution margins of these businesses. The net effect is prudent investment levels paired with regular reviews on forecasted return on investment. We are committed to making timely adjustments to the market direction change. Our portfolio has broadly inflected to growth, but I want to ensure my message is clear. I expect the inflection to growth will benefit valuation but will not delay the portfolio rationalization process.
I'm fully aware of the financial and the non-financial benefits of portfolio rationalization, and we are particularly focused on opportunities that accelerate our debt repayment and decrease our leverage ratio. Second, on accelerating growth and driving margin expansion, we have made swift changes to intensify the engagement with customers to provide technical and operational solutions, and our Q3 results and Q4 outlook demonstrate the effectiveness of these initiatives. We continue to see strong tailwinds. Our customers and targeted market are moving toward us. We have instituted a disciplined investment plan, leveraging our design competency and incorporating performance objectives from key end customers and meeting their critical business needs. I believe we have achieved multi-generational roadmap alignment with key customers, and we aspire to become their partner of choice for key technical and product solutions we provide. I expect these initiatives will accelerate sustainable market share gain and SAM expansion.
Another growth driver is Semtech's operational excellence. Semtech has executed to meet customer delivery timelines in the current dynamic environment, with noteworthy instances in our data center and high-end consumer end markets. Semtech's operational excellence in on-time delivery and superb quality contributes to our customer supply chain resiliency, which I know to be highly valued. In this area, I would be remiss if I did not acknowledge the contribution from Semtech's foundry, assembly, and test partners amidst the current ramp. Thank you to our foundry partners that have prioritized our increased demand many times well within natural lead times. Thank you to our assembly and test partners who have quickly installed additional capacity to support our growth. Third, on energizing our people.
I'm a firm believer in promoting a high-performance culture, and I'm proud that we have launched Semtech Rising, an initiative incorporating employee development, mentorship, recognition, and pay-for-performance elements to bring out the best from our employee base and elevating our determination and drive to win. This initiative also leverages expertise from Semtech's board members. We modified the charter of our compensation committee to become the Human Capital and Compensation Committee. We firmly believe this change elevates the importance of human capital development and aligns with Semtech's focus on creating and fostering a diverse and vibrant workforce. In the coming quarter, these three priorities will continue to be my focus. Moving to end markets. For Q3, infrastructure net sales were $65.8 million, up 24% sequentially and up 52% year-over-year. Net sales for the data center were a record $43.1 million, up 58% sequentially and up 78% year-over-year.
Consistent with our outlook for Q3, shipment commenced on our CopperEdge 200 Gb linear redrivers used in 1.6T active copper cable, or ACC, applications. CopperEdge net sales were in the high single-digit million dollars. We expect incrementally higher contribution in our Q4, followed by a ramp progressing through FY 2026. There have been multiple reports regarding Blackwell GPU rack designs and timing of volume shipment, which could potentially impact the TAM and timing of ACC market where we provide key enabling IC components. That said, allow me to provide some assurances based on our ecosystem engagement. We have invested time with our customer and end users of the racks over the past few months. We reaffirmed our expectation of exceeding the floor case provided a couple of quarters ago based on the first-hand information from the ecosystem.
I connected with many CSPs, cable manufacturers, and ecosystem participants at the very well-attended Open Compute Project, or OCP Global Summit, last month, where Meta presented, contributed, and demonstrated on the show floor its Catalina platform. Catalina is a dual-rack NVL36 design connected with the CopperEdge-enabled ACCs. This appears to be a major platform that multiple CSPs will adopt in the near future for their AI data centers. Several companies are currently conducting design work using our CopperEdge chips in 200 Gb traces on their boards to improve signal integrity. Rack designs have varied and will continue to vary as CSPs deploy various configurations in their data centers, but the OCP Global Summit reaffirmed my belief that a count of 200G ports connected to cables with a length up to 3 meters is a more relevant measure of a CopperEdge SAM.
At 200G and at a cable length up to 3 meters, CopperEdge will meet signal integrity requirements not readily achievable with direct attached copper or DAC cables, and at a lower latency, lower cost, and much smaller power consumption required compared to DSP-based real-time solutions. Semtech's low-power, low-latency CopperEdge solutions have gained positive attention in the data center ecosystem, and our technical collaboration with a number of CSPs and cable manufacturers has accelerated since last quarter. My team's engagement with architecture decision-makers and technical executives in the ecosystem significantly streamlined the time between understanding our customers' challenges and delivering Semtech's proposed solutions. I believe this is proven to be a positive differentiator in CopperEdge proliferation and improved Semtech's NPI time to market. In addition to CopperEdge, our Tri-Edge PAM4 products continue to contribute meaningful sequential and year-over-year growth.
Our lead customer commenced a production ramp in 400 Gb AOC products in Q3. Demand remains robust for our FiberEdge transimpedance amplifier, or TIA, and the laser drivers. Data center deployment at 100 Gb has been ramping up strongly, and we believe we have captured incremental market shares thanks to our closer engagement with our customers and our operations excellence. Moving to linear pluggable optics, or LPO. We have received initial TIA orders from several module manufacturers for test and qualification of both 800G and 1.6T LPO transceivers at CSPs. CSP engagement has proven insightful. It appears that LPO adoptability is meaningfully correlated with a 30 signal-to-noise ratio at the host. Fortunately, both current and future generation switches' supply significantly improved performance, and this enabled easier LPO adoption in many specific use cases.
Our confidence in LPO adoption has increased since last quarter, with meaningful net sale contribution from TIAs and redrivers expected by the latter portion of FY 2026. Within the infrastructure, our telecom business consisting of PON and backhaul reported Q3 net sales of $20.5 million. We have been supporting a pilot build of triple-gen 50G PON with a major Chinese carrier, and the carrier CapEx noticeably for 5G Advanced deployments is expected to nominally improve over the coming quarters. Moving to our high-end consumer end market. Net sales were $40 million, a sequential increase of 8%, reflective of market share gains and consistent with expectations of seasonally stronger Q3. Net sales in high-end consumer TVS grew to $28.3 million, up 9% sequentially and up 7% year-over-year, with a sequential growth in each quarter of the current fiscal year.
We communicated market share growth in consumer TVS through last quarter, augmenting our prior commentary. Our expectation is for continued market share expansion at the world's largest consumer electronics company and at other key North American and Korean companies, based not only on our design activities for future generations of product, but also for Semtech's ability to deliver on time and to meet demand upside. Thanks to our technology leadership, proven quality, and fulfillment capabilities, we are among the first to be added to a BOM for many products in the pipeline, so we have much greater visibility into the design cycles of next-generation products. Our class-leading PerSe products continue to excite the market, with design wins across device manufacturers in smartphone, computing, and wearables.
We believe PerSe's capabilities are also highly valued by device manufacturers to maintain compliance with a specific absorption rate, or SAR, regulations while minimizing effect on device performance. In this area, the ability to detect and measure distance to a human allows the device to optimize RF transmission power and data throughput. We are pleased to be included in the leading smartphone chipset vendors' reference designs to meet SAR requirements. Moving to our industrial end market. For Q3, industrial net sales were $131 million, up 5% sequentially. Within the industrial end market, LoRa-enabled solutions recorded Q3 net sales of $29 million, up 1% quarter-over-quarter and up 104% year-over-year. Encouragingly, consumption for our recent-generation LoRa product has been increasing, which signals market adoption of this enhanced capability.
LoRa Gen 2 offered a smaller footprint and reduced the power consumption, while LoRa Gen 3 delivered improved radio performance and a further simplification of customer development through onboard LoRaWAN provisioning capability. Supporting LoRaWAN remains a key company strategy. I benefited greatly from visiting the SAMS conference in September. As I said during my presentation at the conference, LoRa has demonstrated tremendous potential with a comprehensive ecosystem that pushed the LoRa from concept to product and now to an industry. I gained very constructive input while at the conference on how Semtech can support ecosystem enablement and enhance what I believe to be Semtech's already robust hardware and software roadmap. The infrastructure to cover gapless coverage for LoRa continues to gain momentum. In addition to EchoStar coverage throughout Europe, satellite operators are considering offering a similar service for Americas.
We expect such availability of LoRaWAN will create a boundless potential for asset tracking. Our IoT systems business recorded Q3 net sales of $57.9 million, up 11% sequentially, yet another quarter of net sales growth coupled with robust bookings and backlog. As previously discussed, Semtech started addressing channel and end customer inventories earlier in the cycle, and as the market inflects upwards, our growth is now muted by the channel congestion. Highlighting customer engagement for this business, I believe we held a successful customer roundtable in October where our customers provided valuable insights and helped shape our roadmap. The roundtable also allowed us to further highlight our in-house Trade Agreements Act, or TAA, capabilities, something we expect to bring significant value as a North America supplier to the critical infrastructure markets we serve.
Our in-house TAA capabilities have contributed to funnel orders, backlog, and sales to federal agencies, an area where we have identified an adjacent market and renewed our focus. In Q3, we launched the Australian instance of AirLink Management Services. Australia is one of our stronger markets outside of North America, and this instance meets ever-stringent local data resiliency requirements. Q3 net sales for IoT connected services were $24.6 million, with a slight sequential and year-over-year growth for this recurring revenue business. AirVantage allows our customers a single panel to monitor lifecycle management, consumption, and SIM management, all packaged with a robust API. Our future growth expectations are bolstered by additional customer adoptions of AirVantage, including AirVantage data insights and its AI-enabled features to facilitate analysis and decision-making on connected equipment.
We also expanded our smart connectivity platform to add voice-over LTE coverage, which is now available in over 40 countries and territories. In industrial TVS, net sales in Q3 were $10.2 million, up a robust 7% sequentially. We have noticed the current market sentiment in the industrial market, but we remain confident in Semtech's growth with our product offerings. Our industrial products address protection for electronics deployed in increasingly harsh industrial ESD environments. As factories increasingly automate, our customers are increasing their reliance on Semtech's solutions, where severe ESD challenges abound. In the automotive market, demand has increased sequentially with a growing prevalence of wired and wireless networks in automobiles. Also, Semtech has been a leading supplier of advanced display protection solutions in the high-end consumer market. As our display customers diversify into the automotive sector, we believe we are seeing a disproportional benefit in the automotive display market.
In summary, I'm very pleased with Semtech's execution and performance across our businesses in Q3, and now turn the call back to Mark for additional details on our financial results and our Q4 outlook. Mark?
Thank you, Hong. For Q3, we recorded net sales of $236.8 million, up 10% sequentially. Net sales trend by end market, reportable segment, and geographic region is included on Slide 16 of the earnings presentation. Gross margin was 52.4%, up 200 basis points sequentially and up 110 basis points year-over-year. Operating expenses were $80.6 million, slightly below the midpoint of our outlook, increasing 3% sequentially and representing what we forecast to be prudent investments in the business to accelerate realization of growth opportunities. Operating income was $43.4 million, resulting in an operating margin of 18.3%, up 410 basis points sequentially and up 810 basis points year-over-year.
Adjusted EBITDA was $51.1 million, and Adjusted EBITDA margin was 21.6%, up 280 basis points sequentially and up 760 basis points year-over-year. Gross margin, operating margin, and EBITDA margin all sequentially increased in each quarter of fiscal year 2025, demonstrating our operating leverage. We expect to finish the year with sequential improvements in each of these metrics based on the midpoint of our Q4 outlook. Net interest expense was $18.4 million, and we recorded net earnings per share of $0.26, up from $0.11 in Q2 and up from $0.02 in Q3 of last year. Operating and free cash flow for Q3 was $29.6 million and $29.1 million, respectively. We ended Q3 with a cash balance of $136.5 million, which included a principal payment of $5 million on a credit facility.
Subsequent to the end of the quarter, we made a further principal payment of $10 million, and these payments are consistent with our previously stated capital allocation priority of reducing debt. Now turning to our Q4 outlook, we currently expect net sales of $250 million, plus or minus $5 million, a 6% sequential increase at the midpoint. We expect net sales from the infrastructure end market to increase sequentially, with data center applications leading the growth. Infrastructure is expected to provide the strongest near-term tailwind. We expect net sales from the high-end consumer market to be down, reflective of typical seasonality. We have gained market share in this end market and do not believe channel inventory is a headwind to Q4 expectations. We expect industrial net sales to be up, with increases across LoRa and our cellular IoT portfolio.
Based on expected product mix and net sales levels, gross margin is expected to be 52.8%, plus or minus 50 basis points. At the midpoint of our outlook, this would be a 40 basis point sequential improvement. Operating expenses are expected to be $82.8 million, plus or minus $1 million, growing at about half the rate of net sales growth at the midpoint and resulting in operating margin at midpoint of 19.7%, a 140 basis point sequential improvement. Adjusted EBITDA is expected to be $56.9 million, plus or minus $2.8 million, resulting in Adjusted EBITDA margin at the midpoint of 22.8%, which would equate to a sequential increase of 120 basis points. We expect net interest and other expenses to be $19 million and expect an income tax rate of 15%.
These amounts are expected to result in a net earnings per share of $0.32, plus or minus $0.03, based on a weighted average share count of 80 million shares. Our outlook at the midpoint contemplates another quarter of growth in net sales, improving gross, operating, and Adjusted EBITDA margins, and higher diluted earnings per share. With that, I'd now like to turn the call back over to the operator for Q&A.
Thank you. And at this time, we'll be conducting our question-and-answer session. Please limit yourselves to one question plus one related follow-up question for each time that you queue. If you would like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate that your line is in the question queue. You may press star two if you would like to remove your question from the queue.
For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. One moment, please, while we poll for questions. And our first question comes from Quinn Bolton with Needham & Company. Please state your question.
Hi, Hong and Mark. Congratulations on the nice results and outlook, especially in the data center business. I guess, Hong, I wanted to start with the data center business. It sounds like you continue to be very comfortable with the floor TAM for ACCs, but hoping you might be able to elaborate on that. You mentioned the Catalina rack is an opportunity, but, wondering if the ACC TAM that you see now starts to encompass other opportunities at hyperscalers or CSPs.
And so just wondering if you might be able to talk about some of the use cases you see starting to contribute to that $100 million-plus opportunity. Thanks.
Thank you, Quinn. Yeah. So as we discussed at OCP, you participated as well. The Catalina is going to be the main platform. We know one major CSP is going to be used at the baseline for deployment in 2025 and beyond, as long as they use GB200 GPU processors. And we know that standard is getting traction, but I haven't got the detailed confirmation on which CSPs are using it and what proportion.
Another major progress, I would say, since the announcement of ACC is the awareness level by the CSPs on this elegant capability and solution to improve the signal integrity and improve the link budget by adding a very small fraction of power consumption, no latency, and very little, well, incremental cost as well. So now we are seeing many several CSPs and other companies are considering using the linear equalizer in their entire trace design. It can be on a board, it can be in the connectors, it can be in the cables as well. So that gives me the confidence that after the OCP, that our floor case guide d a couple of quarters ago, is indeed a floor case.
Perfect. Maybe just a quick follow-up there.
Just for the CSPs looking to use linear drivers in cables or PCBs, does that ramp in calendar 2025, or do you think that that's a longer-term opportunity? And then my follow-up is you seem to be more upbeat now about the LPO opportunity. I think you said you've got initial orders for TIAs for LPOs starting to ramp in fiscal 2026, the back half of fiscal 2026. Can you talk about what types of applications you're seeing LPOs starting to be deployed this year? Thank you.
Yeah. Let me just follow up on the first one on the ACC. So the applications on the board and in the cable and even in the connectors by the multiple CSPs are in the qualification phase and in the demonstration phase. So the typically good thing about the copper-based solutions is the qualification cycle is relatively short compared to the optical-related products.
So I would say probably from the mid of the 2025 calendar year, the other opportunities on the linear equalizer will start contributing to the revenue. As for the LPO, as I mentioned in the prepared remarks, that we have already got orders and shipped in low-volume a bit. There's a 800 Gb DR8, 100 Gb per second per trace, and 1.6T side, primarily on TIA. So they are building transceiver modules and shipping to different customers for qualification. And the applications are for both scale-up and scale-out. So the scale-out is primarily connecting the NIC card to top of the rack or end of the row switches. And scale-up is to increase the cluster size for the GPU compute cluster or ASIC compute cluster by some CSPs.
Perfect. Thank you for that additional color on.
Thank you.
Our next question comes from Harsh Kumar with Piper Sandler.
Please state your question.
Yeah. Hey, Mark Lin, Hong, and the entire Semtech team. Congratulations on a very strong quarter, very strong guide. I had one quick one on data center, another one on LPO. So Hong, for you, could you just give us an idea of what you're thinking about the growth for ACC in the January quarter as you're guiding? And then when do you think it'll steady out to a sort of a steady-state sales format? Is it second half, or will it continue to build through the year? And then I have a follow-up.
I will let Mark answer the ACC question.
Harsh, we said this in Q3 was high single-digit millions. In Q3, it's a nominal ramp in Q4, and then it progressively ramps through FY 2026, Q1, Q2, Q3, and Q4.
So we've been pretty consistent with that messaging, and we don't really see a change in that timing.
Great. Thank you. And then for my follow-up, if I can talk about LPO, maybe a very theoretical question, Hong. So a lot of debate on LPO. There's a lot of non-believers that think it'll never happen because of interchangeability. Is it possible in theory to have a compatible interchangeable module using LPO, or is that just theoretically impossible? Could this eventually just take over the whole optical business?
Yeah. Harsh, that's a good question. And I think the debate is still ongoing even after two years.
And what I have found out is that the early sentiment developed by the impossibility of the LPO for scale-out was largely done on the previous switch version, where we understand that signal-to-noise ratio from the host is not as good as the current generation, say, for example, Tomahawk 5. So if they did the LPO and tested in the previous version series, and there might be very little margin, and that makes interoperability very difficult. But in the current and future generations with switches, fortunately, the host has demonstrated really pretty superb signal-to-noise ratio. And so the LPO is like ACC. A linear equalizer is an essential part of it at the start. So with a good signal-to-noise ratio, and the current and future switches can use LPO.
But as the adoption timeline is getting closer, and we noticed that many CSPs, they started pivoting to LRO, so basically using one side and the linearized, the other side retimed solution to be safe because they do need a low-power 1.6T interconnect solutions. To us, we have arguably the best TIA on the receiving side. So we'll benefit from either LPO or LRO. And as industry progress and getting better understanding on the compatibility of different types of hosts with LPO and LRO capabilities, I do believe this type of transceivers can chip away a sizable total addressable market currently served by the DSP retimed s olutions.
Congratulations, guys. Thank you so much.
Thank you.
Thanks, Harsh.
Your next question comes from Cody Acree with The Benchmark Company. Please state your question.
Yeah. Thanks for taking my questions. And congrats on the progress.
Guys, could you, Hong, could you talk about specifically to the Blackwell opportunity? Obviously, there's been talk about the NVL36 platform discontinuation of development support. It doesn't sound like that that's happening at a certain CSP that you're working with. I want to know if that's more of a one-off in the industry. Is that something that you're seeing more broadly continue to be adopted, or has there been a change at NVIDIA that is impacting the broader ACC opportunity at that customer program but may not be impacting the long-term opportunity beyond them?
Yeah. So Cody, I definitely have talked to many CSPs, and everyone, they have their views, but largely based on how much compute power they wanted to have in that cluster.
Right now, 72 GPUs seem to be an ideal cluster size, but it's also bounded by their ability in the data center infrastructure to cool the racks. 36x2 seem to be pretty ideal. For one leading CSP, we know that is their baseline for 2025 and 2026 and for as long as they use Blackwell. I have heard some others using NVL72, and where we don't have the contribution for backplane, but at the front end, they either need to connect 1.6T ports or 800 gigabit ports to top of the rack or end of the row switches. I think that's where LPO can really have a good provide a very differentiating solution because of the low power consumption. I think that is probably why the industry is pushing very hard on the LPO solutions.
But as for the different variety of racks, we're going to be continuing to hear reports in the future just for the aspiration to continue to increase the number of GPUs in the cluster to be able to handle the model with more parameters. So it's going to be dynamic. That's why I wanted to really believe that counting the number of ports of 200 Gb per second will be a more relevant measure going forward rather than just stare at how many of the NVL36 racks out there.
All right. Thank you very much for that color. Can you maybe just help frame your overall data center opportunity? How much of that is fiber and Tri-Edge in the future and delineate that from your ACC and your LPO opportunity?
Yeah. Typically or historically, we haven't provided that delineation where we gave a high single-digit million figures for ACC.
But realistically, the entire data center market is growing quite robustly for all of our products: ACC, PMDs, Tri-Edge. And we expect when LPO does ramp, it'll continue to kind of support that growth level.
Thank you.
And our next question comes from Tore Svanberg with Stifel. Please state your question.
Yes. Thank you. And congratulations on the strong results. Hong, the CopperEdge revenue, high single million for next quarter, sorry, for this quarter, what's the mix there between 200 Gb versus 100 Gb? It sounds like it's primarily 200 Gb. And as we think of fiscal 2026, will again the majority of the mix be 200 Gb per lane?
Thank you, Tory. Yes. The revenue in Q3 on CopperEdge is primarily 200 Gb. And going forward, the 200 Gb is going to be the primary driver as well.
We do have some contribution for 100 Gb, especially after the industry is aware of the capability that the linear equalizer is providing to improve the signal integrity. Some CSPs are revisiting their architecture in the data center and start considering using a 100 Gb ACC instead of DAC cables. But the contribution we count and we guide is primarily 200 Gb.
Tore, let me just double-click on that. In our Q3, the 200 Gb was substantially all of our CopperEdge shipments.
Perfect. Thank you for that, and let me move on to a non-data center question, so you guys have done a good job to climb back to this $1 billion run rate that you've guided for the January quarter. As we just think about strategically, because Hong, you mentioned you're still reassessing everything.
So how much of that $1 billion should we think of as true strategic revenue versus perhaps segments of the business that are requiring less investments at this point?
Tore, that's a good question. For the Q4 guidance, we broke down the different industrial segments we have right now, and you can use that as a guideline. Thank you.
Fair enough. Thank you.
Thank you. Our next question comes from Tristan Gerra with Baird. Please state your question.
Hi. Good afternoon. Now that CopperEdge is ramping, how should we look at the free cash flow generation over the next few quarters and any update on the type of debt leverage ratio you expect exiting next year?
Yeah. That's to say, I'm quite pleased, Tristan. Q3 operating cash flow was $29.6 million. Free cash flow was $29.1 million. Cash flow definitely we've inflected consistent with the business.
And I'm pleased that cash flow generation is broad-based across our businesses. We may have to build a little bit more inventory supporting demand, but we continue to generate cash. And as you see that, as soon as we generate the cash, we delever, right? We pay down principal. In terms of where we're going to exit, we only guide out one quarter, but you can maybe take a look at our EBITDA that we reported this quarter kind of annualized. That might give you an indication of potentially where our leverage ratios are heading.
Great. This is helpful. And then if I look at the coherent announcement of the NVIDIA-based DSP back in September, does that create opportunities for your TIA in terms of content?
And is that a solution that eventually gets available outside of NVIDIA and will allow you in that particular transceiver to reach additional customers?
Yeah. Tristan, so for every DSP retimed transceivers, we should have content in there. And definitely, our baseline data center revenue growth has been primarily driven by the DSP-based transceiver growth. And Q3, in the first quarter, we had some meaningful revenue from the CopperEdge for ACC. But that DSP-based transceivers will continue to grow, will continue to benefit from that.
Great. Thank you.
Thank you. Our next question comes from Scott Searle with Roth Capital Partners. Please state your question.
Hey, good afternoon. Thanks for taking the questions. And congrats on the data center performance. Maybe I'll shift away from the data center since it seems like it's been covered.
Hong, on the PON front, it seems like it was another decent quarter, still really China-focused. I'm wondering if you could give us some thoughts in terms of the incoming administration, any sort of impact on the PON business in China, if there is any, kind of how you're thinking about growth there, and then specifically how you're seeing design activity outside of China, particularly as we start to talk about 10 Gb and above.
Thank you, Scott. So the PON business up to this point has been primarily in China. And we expect another tender offer over the next quarter or two. And we have the product ready for it. We have a triple-gen demonstration in there. And it can operate at 10, 25, and 50 Gb. And so that is really setting the benchmark of the capability. We're the leading provider in that market.
As for what is the future political situation there, we are absolutely very mindful about that. But so far, we have not seen any impact. As for the U.S., the report about the Calix and Louisiana, they got the BEAD funding. So we have been engaging with all the leading equipment manufacturers in the U.S. for the U.S. PON opportunities as well. So we're really well positioned for the global infrastructure upgrade. And beyond China and the U.S., we're trying to, through our module manufacturers, address other parts of the market. But those are the two major markets for the PON opportunities.
Great. Very helpful. And if I could follow up just on the LoRa front, continuing to stay away from data center for a minute. A nice recovery quarter.
I think at $29 million, I think it's still below where you'd peaked a couple of years ago in the $40 million or so range. I'm wondering if you could talk a little bit about some of the applications where you're getting the traction. I know there have been some continental-wide buildouts being driven by guys like Netmore in Europe. It sounds like there are some opportunities in the U.S. But just kind of give us some high-level thoughts in terms of where you're seeing that design traction and how we should think about that growth rate going forward from these levels. Thanks.
Thank you, Scott. So yeah, previously, a couple of years ago, the LoRa market had some noise about Helium. So I think that noise dissipated. And right now, we're seeing a clear signal of the continued growth. From Q1 to Q2, we had a strong growth.
In Q3 to Q2, we have marginal growth. But year-over-year basis, we had a strong growth. Our use cases right now, I have spent a lot of time talking to ecosystem partners. I'm just totally fascinated. Traditionally, it has been smart metering. But it's connected spaces and for automation, for asset tracking. And now I see more and more use cases. And you mentioned about Netmore. They come out very strong. They are aspiring to consolidate and to have multiple territories and multiple countries' presence to proliferate their successful and rewarding use cases. I think that is really showing that the industry is starting to mature. And so far, the major applications are smart metering and asset tracking. I think the asset tracking is going to be having a lot more applications. And then the smart homes, smart factories, and the use cases have been more known right now.
There are a lot of well-established solutions for those. And I talked to several companies. They are seeing tremendous growth in their business in those use cases. So we're very excited about the LoRa growth perspective. And we will be having new products coming out, Gen3 and Gen4, make the development of downstream integrated solutions easier. And we plan to use the LoRa Alliance to provide more enablement to the ecosystem and raise the level of awareness of the successful use cases.
Okay. Great. Thanks so much. And congrats on the quarter and outlook.
Thank you.
Our next question comes from Craig Ellis, B. Riley Securities. Please state your question.
Yeah. Hong, Mark, congrats on the execution, especially around growth and margins. Hong, I wanted to go back to data center, but maybe approach it in a more longer-term way.
I think it was at least three quarters ago that we started talking about what seemed to be a single company, more single product opportunity as having a $100 million base opportunity to it that would be in the 2025, 2026 timeframe. The question is this: Is the business looks like it's gotten significantly broader customer and application-level exposure and design potential? How do we think about the size of this business two to three years down the road?
Craig, that's a great question. I think the opportunity started with this single company, single platform. And that is a great trailblazer for this new product that's accelerated our time to market. But right now, as you mentioned, as we observed, this capability is broadly recognized and beyond that single company, beyond that single platform.
That's why we have been thinking about and to qualify the opportunity by counting the number of 200 gigabit per second ports. The reason for that is everywhere you have 200 Gb per second transport, you have the same challenge. You need the same solution for signal integrity. With the Tomahawk 6 rolling out right around the corner, well, maybe six to 12 months, and all the ports, it's going to be 200 Gb. It's only increasing our opportunities. So far, the application has been for scale-up. But with the scale-out added into the opportunity pool, we got a tremendous opportunity in there. Probably a better, a more relevant measure is to qualification of a number of 200 Gb ports. Then you can put a multiplier on adoption ratio, 30%, 50%, or 70%.
So we are in the process of establishing that model to quantify the opportunity so that people will not just stare at how many NVL36 racks are being forecasted and being built. So we will report that progress as we're making good progress.
That's a helpful summary. Thanks very much, Hong. Mark, I wanted to cycle back to you. You talked about things happening. And Hong talked about where we are on the calendar planning cycle with expense management. As we think beyond the fiscal fourth quarter and think through 2025, are there any discontinuities coming with expense investment, whether it be mask set costs or other things? And should we continue to expect that R&D and OpEx would grow at some small coefficient of revenue growth? Is that reasonable, or are there one-time things coming?
Yeah. Craig, we don't see any one-time things coming.
We've gone through our strategic planning, so we do expect incremental investments, but I believe those incremental investments will have near-term returns, and the other area is Hong has absolutely stressed there's no evergreen R&D projects, so if something were to turn and the market moves away from that particular application, we don't mind just discounting the sunk costs and redirecting our efforts. In terms of growth, we've reiterated a number of times healthy growth is probably OpEx grows at half the rate of revenue growth, and I think that's what we're sticking to. Now, just some of that operating expense investment will be redirected, as you heard from in our prepared statements, probably a little bit more towards data center, but those are some near-term growth opportunities there.
Got it. Thanks for that, Mark. Thanks, Hong.
Sure thing, Craig. Thank you.
Our next question comes from Christopher Rolland with Susquehanna International Group. Please state your question.
Hey, guys. Thanks for the question. I want to echo my congrats on a fantastic quarter and guide. So my first set of questions are around ACC. And I guess maybe first to level set, how much of CopperEdge were you shipping in Q2? Or was that high single digits, million, all incremental? And then the bulk of these shipments, is this all to the big AI GPU guy directly? And they're going to be using it for branded cables? Or are these going to third-party cable OEMs for channel sales or to the CSPs direct via EMS? How is the go-to-market here for these shaking out? And how might you expect that to change through the next 12 months?
Yeah. Great question, Chris.
We're looking at when we mentioned that it's high single-digit millions of CopperEdge. Effectively, you can consider that 200 gig active copper cable. And it's really directed. Our customer is cable suppliers. But I think we get the demand really from that large company. So at this point, just with the demand for the cables, I don't think really anything's going into channel. It's really going to supply rack shipments. And then just to add to that, Chris, you asked Q2. Q2, basically, it's just a very minimal sample quantity. So the Q3, the high single-digit million dollars is the first volume. Yeah. You consider that for 1.6T applications there, Chris.
Excellent. Thank you very much. And then just a quick clarification was PON. I guess PON was down, if I'm doing my numbers right. But just wanted a clarification.
But a better question, perhaps, bigger picture is around divestitures. Where do we stand there? Are you guys still interested in selling parts of the business? And conversely, what is the interest in perhaps buying parts of your business as well?
Yeah. That's an area, of course, where we're a little bit sensitive. We did, in our prepared remarks, we did keep kind of balance sheet and portfolio as the number one priority. And we do believe order matters. So it is the number one priority. At this point, I think all of our businesses have inflected to growth. So as Hong mentioned in his prepared remarks, we believe that should help valuation. But that in no way will delay or maybe impede our desire to potentially divest these non-core businesses.
Excellent. Thank you, guys.
Thank you, Chris.
And our next question comes from Tore Svanberg with Stifel.
Please state your question.
Yeah. Thanks. I just had a quick follow-up for Mark. Mark, so 40 basis points base plan to improve gross margin for January. How should we think about gross margin for fiscal 2026? Is it mainly mix at this point that will drive the gross margin? Or are there other things, maybe scale or anything like that, that could potentially also lift it as well?
Yeah. Scale definitely helps. But definitely, the primary driver in our guide is mix. So it's a 40 basis point improvement. But we did get a little bit of a tailwind from the CopperEdge shipments this quarter. So that was definitely a tailwind. But as other portions of our business inflect upward, there's a little bit lower margin in IoT, our systems hardware business. So that's a little bit lower. We'll definitely take the gross profit.
But definitely, that business doesn't contribute quite the percentages, let's say, our infrastructure business.
Sounds good. Thank you.
All right.
Thank you. And there are no further questions at this time. I'll hand the floor back to Mark Lin for closing remarks.
Great. Thank you for joining. And please visit our investor website at investors.semtech.com, where we post upcoming investor conferences where Semtech will be in attendance. Have a great day.
Thank you. All parties may now dismiss.