Semtech Corp Q2 FY2026 Earnings Call

Good day, and thank you for being here. Welcome to Semtech Corporation's Second Quarter Fiscal Year 2026 Earnings Conference Call. Please note that this call is being recorded. I would now like to pass the call to Mitch Haws, Senior Vice President of Investor Relations for Semtech. Thank you. Please proceed.

Thank you, and welcome to Semtech's Second Quarter 2026 Financial Results Conference Call. Participants on today's call are Hong Hou, our President and Chief Executive Officer; and Mark Lin, our Executive Vice President and Chief Financial Officer. Before we begin the prepared remarks, I would like to highlight upcoming investor events, including the Deutsche Bank Technology Conference on August 27, The Benchmark TMT Conference on September 3, The JPMorgan Rising Tech Leaders Forum September 4, and The Piper Sandler Growth Frontiers Conference on September 10. Today, after market close, we released the unaudited results for the second quarter of fiscal year 2026, which are posted along with an earnings call presentation to our Investor Relations 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 Slide 2 of the earnings presentation as well as the Risk Factors section of our most recent annual report on Form 10-K for a number of risk factors that could cause our actual results and events to differ materially from those anticipated or projected on this call. You should consider these risk factors in conjunction with our forward-looking statements. We will refer primarily to non-GAAP financial measures during today's call. Please see today's press release and Slide 3 of the earnings presentation for important information regarding notes on 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, Mitch. And good afternoon to all of you joining today. The Semtech team made solid progress again this quarter with sequential increases across each end market leading to record net sales. We also delivered sequential improvement in adjusted gross profit, operating income, and earnings per share, strengthening our financial profile while executing on the R&D roadmap that we believe establishes a foundation for long-term growth. I've completed my 1-year tenure as Semtech's CEO and reflecting on the three priorities I outlined in our earnings call a year ago, we have made tremendous progress. First, on strengthening the balance sheet. At the end of Q2, we have reduced debt by $879 million from the time I started as CEO, resulting in a year-over-year quarterly interest expense reduction of 80% and a substantial net leverage ratio improvement to 1.6x at the close of Q2 '26 compared to 8.8x a year ago. This strong improvement to our financial foundation allowed us to focus on drivers for our business. Second, on rationalizing the portfolio and increasing investment in the core assets. I'm happy to report that the core assets we have delineated, namely data center, LoRa, and PerSe, each strongly contributed to our net sales momentum throughout the year. With increased R&D investment into these core areas, we anticipate further acceleration of our momentum. Third, revitalizing our winning culture. This is an area of progress of which I'm most proud. By strong engagement with employees through frequent site visits, interactive information sessions, small group and one-on-one meetings, as well as regular and transparent communications, we provided much-needed clarity in the company's vision, strategy, and priorities, following a call to action. By instilling a culture of customer intimacy, operational discipline, and strong execution, we believe we have made great progress on achieving roadmap alignments with our key customers through significantly improved customer engagement, securing new product design wins, and delivering strong financial performance. I'd like to extend my sincere gratitude to the senior leadership and all of our fellow employees for their resilience, dedication, and commitment to Semtech's Rising Initiative. Going forward, the priority of portfolio optimization is further elevated. We have managed our noncore assets back to a growth trajectory. And combined with the market tailwinds, we believe these assets represent a very compelling business to the right shooter. We believe we are well positioned to further transform Semtech into a higher growth and more profitable company. Now let me move the discussion to our end markets. For Q2, infrastructure net sales were $73.4 million, up 1% sequentially and up 39% year-over-year. Infrastructure revenue growth benefited from record revenues in our data center business. Net sales for data center reached a record $52.2 million, up 1% sequentially and up 92% year-over-year, benefiting from our broad portfolio. FiberEdge products achieved record net sales, offsetting the CopperEdge air pocket from the initial rack deployment at our anchor customer. Based on Q2 performance, we expect continued strong opportunities for FiberEdge demand for the remainder of calendar year 2025 and beyond from our optical module customers serving North America cloud service providers or CSPs. This conviction is supported by our direct ecosystem engagement, which correlates with increases in the data center CapEx forecast from multiple hyperscalers, sovereign operators, and enterprises. During Q2, bookings and forecast from our optical module customers serving China-based CSPs were generally cautious due to limits on GPU availability. That said, we have started seeing accelerated data center bookings over the past several weeks for this market. Looking ahead to the next several quarters, we expect the data center market to continue its multiyear growth cycle. The market is shifting to higher data rates to support the increased compute and network interconnect bandwidth, resulting in strong demand for FiberEdge 800-gig TIAs moving rapidly from 400 gig. Beyond 800 gig, we are supporting multiple customers on their 1.6T transceiver designs with both TIAs and drivers. We currently expect volume ramps to start in the first half of 2026, commensurate with the broad deployment of 1.6T switches. While the shift to higher speeds to achieve high bandwidth is a given, it is increasingly important to deliver this bandwidth using low power and low latency network interconnect. Semtech's analog expertise allows CSPs to deliver high-performance compute and increase storage capacity while constraining their budget for networking. On the optical side, we have secured several LPO design wins with our TIAs in 400-gig and 800-gig transceivers. We believe we have secured the lion's share of the TIAs in most optical transceivers. Our 800-gig LPO laser drivers were specifically designed to comply with our LPO MSA requirements, and we believe it is the only compliant driver in the market. Several optical module customers are conducting design-in and testing of our drivers on their transceivers. We are engaged with three of the leading hyperscalers with our 800-gig LPO solution and expect revenues to begin ramping in Q4 of this year. We are accelerating our R&D roadmap and are targeting making 1.6T LPO drivers and TIAs available for sampling before the end of the year. Another high bandwidth, low-power solution is the CopperEdge for ACC and onboard linear equalizer. During the quarter, we delivered 800 gig and 1.6T ACC cables to multiple hyperscaler and enterprise customers for testing and qualification. Those customers are seeing benefits of strong signal integrity, lower latency, and importantly, much lower power consumption, as much as 90% below competing DSP-based AEC solutions, while offering lighter and more flexible cables as well as significantly longer reach compared to direct attached copper cables. We continue close engagement with our anchor customer for their future rack platforms using CopperEdge and 1.6T optical transceivers using our FiberEdge product. We are on track and expect to launch ACC with U.S. hyperscaler customers during calendar year 2026. Currently, we are enabling all the major cable suppliers, all of which have begun initial qualification at multiple hyperscalers. As data center topology continues to evolve, we see copper remaining a foundational element of next-generation data center interconnect, particularly for short-reach links where its cost, power efficiency, speed, and reliability are unmatched. With the bandwidth requirements increasing from 400 gig to 800 gig, 1.6T and beyond, advances in active copper technologies are extending the reach and offering significant power savings, making copper an essential complement to optical solutions. In high-performance computing and AI clusters, copper enables low latency, energy efficient connections at the rack and row level where optics address longer reach needs. By leveraging our 20-plus years of experience in analog data center solutions, we are helping our customers achieve the performance, efficiency, and scalability demands of today's and tomorrow's data center with a comprehensive product portfolio, addressing lane speeds from 10 gig to 400 gig with lane counts from one to eight channels. Moving forward, the momentum in FiberEdge, combined with our emerging CopperEdge and LPO opportunities, all supported by the strong data center CapEx spending, positions our data center business for strong growth. Now moving to our high-end consumer end market. Net sales for Q2 were $41.2 million, up 16% sequentially and up 11% year-over-year. Net sales in consumer TVS were $29.9 million, up 22% sequentially and up 15% year-over-year, consistent with the seasonality associated with the smartphone unit ramps and our strong content across multiple customers. This growth exceeds overall growth in the handset volumes, aligning with our belief that Semtech is gaining content and market share, stemming from our market-leading performance and supply chain excellence. Designed for ultra-high capacitance, sensitivity, and fast response times, these devices safeguard displays as well as high-speed interfaces such as HDMI, USB, and display ports without compromising signal integrity or performance. This makes them ideal for use in smart TVs, game consoles, laptops, wearables, and mobile devices. Leading global consumer electronics brands integrate Semtech's TVS technology into their products to ensure device performance, durability, and reliability. In addition, our PerSe sensing technology is being increasingly deployed across a growing range of applications, from consumer electronics to automotive and industrial markets. In devices such as smartphones and laptop computers, where specific absorption rate standards are becoming more stringent, PerSe enables intelligent power management by detecting proximity and optimizing RF performance to meet regulatory requirements, without compromising the user experience. In addition, PerSe enables precise gesture control with ultra-low power consumption, both of which are highly valued for wearables such as headsets and smart glasses. We are actively engaged in design discussions with a broad range of customers in both smart glasses and smartphone platforms, supporting both existing designs and new launches over the coming quarters. Moving towards the industrial end market. Q2 industrial net sales were $143 million, up slightly sequentially, in line with our outlook and up 14% year-over-year. Within industrial, net sales of LoRa-enabled solutions were $36.9 million, down 5% sequentially and up 29% year-over-year, supported by continued expansion across several end markets and in multiple applications. LoRa offers a unique combination of long-range connectivity, low power consumption, and robust performance in challenging environments. Its ability to transmit data over several kilometers while operating for years on a single battery charge makes it ideal for predictive maintenance, asset tracking, energy management, and smart city infrastructure. It also enables cost-effective and secure monitoring and control of equipment, infrastructure, and environmental conditions over large areas. We are seeing growth in applications, including home security systems, smart appliances, pad and personal trackers, and community-based environmental sensors. In addition, our recent generation LoRa chips offer dual-band capability, 2.4 gigahertz and ISM frequencies to enhance bandwidth. This capability is supporting a new generation of connected devices that require reliable low-power communication without the complexity and expenses of traditional networks. U-band capability is facilitating LoRa's adoption in emerging low-altitude economy, including drone delivery, aerial surveying, and emergency rescue. LoRa is especially well suited for this environment as it combines long-range communication, low power consumption, and strong signal resilience—three factors critical for aerial operations. LoRa technology is used to provide reliable telemetry and sensor data transmission even beyond visual line of sight. This allows operators to gather real-time insights without relying solely on high-bandwidth short-reach video links. Our IoT systems hardware business recorded Q2 net sales of $64.8 million, up 2% sequentially and up 24% year-over-year. Bookings in our hardware business continue to be strong, up over 40% year-over-year due to both the broad market recovery as well as our position as a leading North American supplier. We see strong 5G momentum as IoT transitions from 4G with the growth in both bookings and design wins. We believe we hold a leadership position with the 5G RedCap and are progressing well in launching Qualcomm-based platforms in the coming year. We continue to lead in 5G LPWA, advancing satellite IoT through non-terrestrial network or NTN, which opens up new opportunities for global connectivity. For routers and gateways, our partnership ecosystem continues gaining momentum. As announced in June, several of our products, including our flagship XR60 5G router, achieved Verizon Frontline-Verified status. We now support Verizon's Frontline Network Slice, a dedicated 5G highway for first responders. This opened up significant opportunities in public safety where mission-critical connectivity is paramount. In July, we hosted an AirLink Partner Summit in Dallas. We shared our product roadmap and showcased a range of compelling use cases in public safety, public transit, utility, oil and gas, as well as government applications. Our various partnerships represent fundamental steps as we evolve from a product vendor to a solution provider of choice for mission-critical applications. In summary, we delivered another quarter of strong financial performance in Q2, reflecting both the strength of our core business and the disciplined execution of our strategy. At the same time, we continue to invest in our R&D, which will fuel future growth, ensuring our technology remains at the forefront of the market requirements and customer expectations. With that, I will now turn the call to Mark for additional detail on our financial results and our outlook for the third quarter of FY '26.

Thank you, Hong. I am pleased to report that for Q2, net sales were a record $257.6 million, above the midpoint of our outlook, up 20% year-over-year and the sixth consecutive quarter of growth. Net sales trends by end market, reportable segment, and geographic region are included on Slide 16 of the earnings presentation. Adjusted gross margin was 53.2%, down 30 basis points sequentially and up 280 basis points year-over-year and above the midpoint of our outlook. Semiconductor products adjusted gross margin was 60.7%, down sequentially from 63.7% and up year-over-year from 59.2%. Looking at the adjusted gross margin dynamics in more detail, high-end consumer sales were seasonally higher in Q2, which has a modest negative impact on product mix. Product mix within Signal Integrity was impacted by higher sales of telecommunications products as well as the forecasted decline in CopperEdge revenue. IoT Systems and Connectivity adjusted gross margin benefited from higher sales of routers and gateways with Q2 at 39.5%, improving sequentially from 34.4% and up year-over-year from 35.4%. Adjusted net operating expenses were $88.4 million, within our guidance range. Adjusted operating income was $40.6 million, resulting in an adjusted operating margin of 18.8%, up 460 basis points year-over-year. Adjusted EBITDA was $56.5 million, up 39% year-over-year, and adjusted EBITDA margin was 21.9%, up 310 basis points year-over-year. Adjusted net interest expense was $4.1 million, down 80% year-over-year. Annualizing the Q2 amount, adjusted net interest expense is well under a single quarter's expense from just a year ago. This has allowed us to accelerate investment in strategic high-growth areas of our business while driving earnings growth and cash flow. Q2 adjusted net interest expense decreased sequentially from $5 million, reflective of our continued prioritization of using free cash flow to repay debt. Other net nonoperating expenses were $1.3 million, primarily from foreign exchange revaluation losses, reflective of a weaker U.S. dollar during the quarter. We recorded adjusted diluted earnings per share of $0.41, up from $0.38 in Q1 and a substantial improvement from $0.11 recorded a year ago. In the second quarter, we recorded a noncash $41.9 million goodwill impairment charge from our connected services business that is reflected in our GAAP results. While net sales for this business remained stable, down 1% year-over-year and up 3% sequentially, these results did not meet our internal earnings forecast and resulted in a reassessment of this business's goodwill balance. Operating cash flow for Q2 was $44.4 million, sequentially up 60% from $27.8 million and up from negative $5 million a year ago. Free cash flow for Q2 reflected similar growth at $41.5 million, sequentially up 59% from $26.2 million and up from negative $8.4 million a year ago. We ended Q2 with a cash and cash equivalents balance of $168.6 million, up $12.1 million from Q1, while making optional principal prepayments of $25 million on our term loan. At the end of Q2, net debt sequentially decreased $37.1 million to $359.1 million. Along with debt reduction, strong business performance contributed to an adjusted net leverage ratio of 1.6 as of the close of Q2, down sequentially from 1.9 and down year-over-year from 8.8. Now turning to our third quarter outlook. We currently expect net sales of $266 million, plus or minus $5 million, up 12% year-over-year at the midpoint. We expect net sales from an infrastructure end market to increase sequentially, including growth in data center. We expect net sales from our high-end consumer end market to be up, reflective of typical seasonality as well as content gains. We expect net sales from the industrial end market to be slightly up with LoRa about flat sequentially, combined with growth in our IoT cellular business. Based on expected product mix and net sales levels, we expect adjusted gross margin to be 53.0%, plus or minus 50 basis points, a 60 basis point improvement year-over-year at the midpoint. Adjusted net operating expenses are expected to be $88.8 million, plus or minus $1 million, resulting in an adjusted operating margin at the midpoint of 19.6%, a 130 basis point improvement year-over-year. Adjusted EBITDA is expected to be $60 million, plus or minus $3 million, resulting in an adjusted EBITDA margin at the midpoint of 22.5%, a 90 basis point improvement year-over-year. We expect adjusted interest and other expense net to be $5 million, benefiting from leverage-based pricing on our term loan that aligns a lower interest rate to a lower leverage ratio. We expect an adjusted normalized income tax rate of 15%, consistent with Q2. These amounts are expected to result in adjusted diluted earnings per share of $0.44, plus or minus $0.03 based on a weighted average share count of 91.6 million shares.

Thank you, Mark. We can now turn the call back over to the operator for the question-and-answer session.

Our first question comes from Harsh Kumar with Piper Sandler.

First of all, congratulations on very solid, very steady results. Hong, I did have a question on LPO opportunity and the timing for it. Not too long ago, a competitor sort of suggested that the timing may be imminent and that they were basically in commercial production. I was curious about your thoughts on the LPO timing. So I was curious about how you see the progression through the year. Is there a possibility that LPO could be preponed, maybe it could come earlier? Or are you just in different sort of customers and maybe with a different, slightly different timeline?

Thank you for the question. Yes, LPO, we have been engaging with a broader customer base for applications for CSPs in the U.S. and in China. And they are at different stages of testing and qualification. But one thing is good that our TIA has been in pretty much every transceiver manufacturer's design and qualification. As for timing, some of them will start deployment in Q4, and some others, I wouldn't say is imminent right now as it’s small volume. Because we have a broad product portfolio and our TIA has already been designed into the DSP-based retime solution, we don't see the incrementally higher demand due to LPO just yet, but we do expect that Q4 will mark the start of deployment. We have the design wins in 800 gig and also 400 gig for DR4. That may start in volume soon.

Understood, Hong. And then maybe I could ask you about the general state of the data center spend. You're sort of a networking player. You have parts and modules and cables. You talk to obviously a lot of large companies. I was curious, I wanted to see where your level of enthusiasm is on the continued data center spend these days as you talk to these large hyperscalers and the large networking players.

Yes. Thank you for that question. Yes, we do engage with our direct customers, which are the module manufacturers, but we also engage with the CSPs in the U.S. and also in China. We all read the same news and the earnings report. The CSPs, they have strong conviction and the forecast to increase the CapEx spending to expand the data center capacity and upgrade for AI capability. And we're seeing from our direct customers, a very strong forecast for 2026 and beyond. So we will be benefitting from this tremendous backdrop. On the other hand, in China, the CSPs, as I discussed in the prepared remarks, the first part of Q2, they tend to be a little bit cautious due to the limits on GPU availability. But in the last several weeks, they have come back, and the booking activity has improved pretty significantly. The forecast for the remaining of '25 and 2026 is very optimistic as well. So we see the market having a very optimistic tone, and we're seeing the results from the new business opportunities and the bookings. And because we supply electronic components, you may hear some pockets of constraint in other areas, but it's not for us. Our PMD, physical media devices, are designed to support VCSELs, EML, and silicon photonics modulators. So we don't see the constraint at this point. But we do plan ahead to add more testers and back-end OSAT capacity in anticipation of a pretty significant ramp for 2026 and beyond.

And our next question comes from the line of Joe Moore with Morgan Stanley.

You talked about your outlook for CopperEdge. Can you provide us with a little bit more color on how confident you are in seeing broader adoption? And when you talk about hyperscale customers, what types of projects are you working on there?

Yes. Thank you, Joe. Yes. We have been talking to customers over the last three to four quarters. We covered a pretty broad ground and have been engaging with over 20 customers in the entire ecosystem. Certainly, the awareness level right now has significantly increased. We work with our cable customers very closely. There are four to five key ones that all have our 100 gig per lane or 800 gig cables and 200 gig per lane or 1.6T cables sampled to different CSPs and enterprise customers. We're seeing strong traction. They definitely see the advantage of low power, more flexibility, higher signal integrity, and longer reach than dock cables. They are designed, in one case, for scale-up similar to our entry customer for interconnecting different processors, ASICs in one cluster with more applications in scale-out connecting from the servers to top-of-rack and also serve the backplane replacing the dock cables. We see use cases replacing dock cables and AECs. All these applications are leveraging the unique properties such as low power consumption, high signal integrity, and extended reach compared to the dock. We also have customers using linear equalizers for onboard applications to improve the signal integrity and stretch the reach from the ASIC to, say, front of the panel or pluggable ports. We will see probably a couple of hyperscalers drive to the high-volume ramp first either in Q4 or the beginning of 2026. For 1.6T cables, the timing of the ramp will coincide with the switches. If you don't have 200 gig ports, you don't really need the connectivity to connect the ports. But for 800 gig or 100 gig per lane cables, we start seeing some demand and are preparing for the latter part of this year.

Great. And then separately, are you seeing any supply constraints on 1.6T optics?

The 1.6T optics, Joe, at this point, we don't see strong volume demand yet. But every module manufacturer is designing their optical modules using different DSPs and PMDs, which we provide to our customers; they all require different pitches and different configurations for packaging. So we support a wide range of demand. But the volume ramp is expected in 2026. Currently, the ports would require 1.6T connectivity from two major ASIC manufacturers—one is making GPUs and the other is making switches. That makes it difficult to predict the timing of start for their volume deployment. So, bottom line, we don't see the constraint from outside.

And the next question comes from Timothy Arcuri with UBS.

This is Dino on for Tim. So just a question on LoRa. It looks like results came in strong ahead of the $30 million to $35 million range you previously mentioned. Are you seeing significant contributions from other applications of LoRa? And do you expect to see LoRa performing at the same level in the next few quarters?

Yes. Thank you for the question. That's a great question. Certainly, we are very pleased about the demand for LoRa. We think we are on the right track in providing enhanced capability by the new product. For example, the dual band in addition to ISM baseband; we provide the device capability to run on the 2.4 gigahertz band as well. What that does is to provide enhanced bandwidth data rate, but at a trade-off of a transmission distance. The LoRa can already cover hundreds of kilometers, so it's not a big trade-off. By increasing the bandwidth, we've unlocked a range of applications, including the low altitude economy, drone deliveries, and some applications related to edge AI; for example, parking meters can snap a still picture and transmit the pictures back as an archive using enhanced bandwidth. We also had a LoRa Plus that basically combines LoRa with other RF protocols. By combining, you can address different applications that traditionally LoRa alone cannot address. These new technologies really open up new market opportunities. We are very pleased to see that LoRa continues to have very strong demand. As a matter of fact, the number of end nodes we shipped last quarter is a historic record. So going forward, we are confident that we expect LoRa revenue on a quarterly basis to be between $30 million and $40 million. So certainly, it's going to be an increase from our original belief from $30 million to $35 million.

And our next question comes from the line of Quinn Bolton with Needham & Company.

Congratulations on the results and outlook. I guess I wanted to follow up on the ACC opportunity just to understand timing. It sounds like you still expect some ACC revenue potentially in the fiscal fourth quarter ending January. And if I heard your answer to the previous question correctly, it sounds like 800 gig or 100 gig per lane ACCs potentially for scale-up is the first application to ramp?

So Quinn, that may confuse you. For 800 gig, it is for the interconnect in the backplane and also between racks. Historically, they use dock cables that can reach data rates of 100 gigabit, even 200 gigabit per lane. However, the cables were made of 26 gauge, very rigid; using their words, as rigid as a rod. So when you bend it a little bit, you compromise the signal integrity. They still refer to it as scale-out applications for different interconnects and different switches. But right now, the designs for 200 gig per lane and 1.6T application are scale-up, used for interconnecting different racks for ASICs. The customers have found more applications in scale-out, in the backplane for interconnections, for example, from the NIC card to the top of the rack. And in many cases, within the switch fabric's topology, they are simply replacing the dock cable with ACC due to its flexibility and better signal integrity, maintaining an incremental power budget.

Okay. I guess, Hong, just so I'm clear, the 1.6T or 200 gig per lane cables—I thought a lot of those would depend on Broadcom's Tomahawk 6 switch that enables 200 gig per lane. So are there applications for 1.6T ACCs that ramp before availability of that switch? Or do your 100-gig designs ramp in the fourth quarter before availability of that Ethernet switch platform from Broadcom?

Right. The volume demand will start from 100 gig first per lane, I would say, in Q4. The 200 gig first applications to volume will be the scale-up between different ASICs. Then it will be followed by 200 gig per lane scale-out applications, as you correctly pointed out, when the switch dies are more available in volume, and you connect the NIC card to the top of the rack.

Got it. Okay. That makes sense. And then I wanted to switch. I know the data center business is driving a lot of growth, but you mentioned the PerSe business and engagements in new smart glass platforms, as well as smartphone platforms. I'm just wondering if you could give us your outlook for the ramp of PerSe. I think you've got a smart glass platform that you're on today that's already achieved high volume. Do you see continued growth in smart glasses? And any comments you can make on PerSe adoption in the smartphone segment would be helpful.

Yes. So Quinn, PerSe devices have been the industry standard for smartphones, and you know that for almost all smartphone manufacturers, we're there, and we're in the process of getting into the last major one. As for applications in smart wearables and smart glasses, certainly, our lead customer is Meta's Ray-Ban glasses. There are several other platforms using basically the same functionality linked to their own large language model for AI applications, and we are there. The smart wearables continue to evolve and demand more functionalities. We are engaging with a broad range of customers and designing next generations of products. So that's the next-generation product. This area can evolve into a sizable market, and we are at the forefront of it.

And our next question comes from the line of Christopher Rolland with Susquehanna International Group.

So I do know it might be a little early for January guidance, but seasonally, I do believe historically, that's been down. Is there a broad way to think about kind of seasonality or how we should think about January? Like could you outgrow typical seasonality given the LPO ramp or the ACC ramp? Just any broad kind of milestones or things to think about for January?

Yes, Chris, so we provided our outlook for Q3 and our end market commentary should help investors formulate thoughts for growth in the out periods. You're correct, high-end consumer sales do tend to trail down in Q4. We don't really see a change to that particular trend. I mean high-end consumer net sales were $41.2 million in Q2. That's a multi-year high, up 16% sequentially, up 11% year-over-year. While we're gaining design wins and market share, which allows us to grow above market rates, I still view Q4 as just a seasonally trailing down, not a weakness at all in any of the business. Our industrial end market is performing well. We've heard from some industry bellwethers on these trends. ISC business is performing well for the 4G to 5G transition, which is a tailwind there. And in infrastructure, data center is definitely a growth engine in all the commentary that Hong has provided in our prepared remarks and for the Q&A up until now. We do have a very broad portfolio. FiberEdge shipments were up about three times compared to a year ago. Anything that we talk about in terms of LPO would be incremental, and ACC is definitely incremental to that ramp.

That's fantastic. Maybe for my second question, I think you guys said you were down to 1.6x leverage. Pretty incredible from where you guys were just a couple of years ago. But my specific question is, what does this mean for the odds of doing a potential acquisition and/or what does this mean for the odds for doing a potential divestiture?

Yes. So Chris, we are pretty excited about the progress we have made. A couple of years ago, our leverage ratio was at 8.8x. So certainly, that's a huge improvement. This improved financial foundation allows us to be more aggressive in capturing growth opportunities. Through close engagement with customers, we have identified many great growth opportunities, and we have been able to balance the R&D spending with our bottom line over the past year. I think we're striking the right balance. We're increasing R&D spending in the core area by 20% sequentially while still delivering good bottom line performance. Going forward, we'll continue to invest in core areas, and there might be some opportunities when we analyze our portfolio; we're seeing we can apply technology leverage, customer leverage, or operational leverage. We're seeing some capabilities to do small tuck-ins, but the highest priority for us is still the portfolio optimization. We will decide what areas we want to get into with Board support and continue the strategy moving forward. So hopefully, that answers your question. Definitely, the core and noncore delineation is not just a paper exercise; it's really a North Star to set the priority for us.

And our next question comes from the line of Rich Schafer with Oppenheimer.

I have a question about Tri-Edge. I'm interested in the outlook for PAM4 and the Tri-Edge business as the industry seems to be concentrating more on 100G and 200G lanes. What are your plans for that business? How do you perceive the Tri-Edge opportunity? Is PAM4 expected to drive growth for Semtech?

Rich, thank you for your great question. So the Tri-Edge has been our traditional offering. It is basically integrated products with drivers and TIAs integrated with our clock and data recovery ones. You could almost say it's an oxymoron that it's the analog version of the DSP. We offered our product at 50 gig per lane to give, say, four channels, giving a 200-gig aggregated bandwidth, and if eight channels, it reaches 400-gig aggregated bandwidth. Our customers, particularly in China and also one of the CSPs in the U.S., have been using our Tri-Edge in AOC cables for 400 gig and 200 gig. That has been ongoing on a limited basis. However, one big CSP in the U.S. is giving us the forecast and starting the ramp up by using the Tri-Edge in AOC applications for 400 gig. Our plan is to continue to push that envelope to make the PAM4 50 gig move up to 200 gig. We're going to skip 100 gig PAM4 because we're a little too late for that. The beauty of the 200-gig Tri-Edge CDR based is going to continue to deliver low power consumption and have the advantage of both equalizing in the frequency domain and retiming in time domain. This is on our roadmap and under development. We believe when the 1.6T transceivers are launched, the very next thing they will be is driving for volume and then later on driving for cost reduction and power reduction. So we will catch that wave of providing low-cost, low-power versions of Tri-Edge.

And our next question comes from the line of Cody Acree with The Benchmark Company.

Congrats on the progress. Hong, can you just go back to your ACC commentary on the cloud service providers? You mentioned expecting that to begin early '26. Is that any reset of timing from your earlier expectation of ACC diversification in Q4? Or was that always the case, the delineation between CSPs and the cable providers?

Cody, thank you for the question. The timing for the LoRa brand is going to go through the platform architecture. The platform architecture of our customers will go with the appliances they need in there. For example, in this case, the switch—timing of the switch availability, as you probably know and heard, is going to be pushed out a little bit. That is the one thing we found during Q2. However, as I said, there are two other use cases. For example, 100 gig per lane, 800 gig ACC cable, that's independent of that timing. We believe that ramp will start in Q4 for that flavor. Another is a 200 gig per lane and 1.6T cable for scale-up between different racks of ASICs, which will also align along the same timeline in Q4. When starting to ramp back from a very low level to a sizable level, timing does matter a lot because the ramp-up slope is quite steep. The good thing is that we have managed through that type of ramp in the past supporting the anchor customer, and we have the confidence that we'll be able to support the market adequately.

Yes, so we have guided out that quarter. You can just expect us to be prudent and not overspend. And maybe we'll just leave it at that.

And our next question comes from the line of Tore Svanberg with Stifel.

I wanted to just take a step back and ask about the general business environment from a linearity perspective. Is there sort of any color you could share with us on linearity of sales and especially on bookings?

Yes. So Tore, thank you for your question. We are seeing pretty strong booking activities in data centers, LoRa, and PerSe. Even the consumer—high-end consumer TVS is very strong. The industrial modules have a tailwind, and the booking activity is very strong. As for the linearity, quarter to quarter, you always have different product mix. But I would say, it is very rare to have all items lining up in supporting positive momentum at once, and that is currently the case. I feel really good about future quarters.

Very good. As my follow-up, I had a clarification question on LPO. So you said you expect three leading hyperscalers to be in 800 gig production in Q4 of this year. Are those three hyperscalers all U.S.? Or is that U.S. and China?

So three hyperscalers, two are in the U.S. and one in China. But the LPO transition to take over some of the DSP-based transceiver market share is inevitable. The beauty for us is if it stays with DSP-based, we have TIA content. If the transition occurs to LPO-based, our SAM is going to double; we will have the driver content as well. The timing to us is important, but it's not as super sensitive because we are already an incumbent for DSP-based transceivers.

Hong, I wanted to go back to the opening part of your prepared comments where you talked about things accomplished in your first year but really use that as an opportunity to ask you what you would like to see the business accomplish in your second year, especially as you look at the infrastructure business from where you are today. What would you be happy with the business accomplishing over the next four quarters, either from a product standpoint, a scale-up standpoint, et cetera? Just some qualitative color on where we would be from where we are would be really helpful.

Thank you, Craig. If I say largely broad brush, the second year, because of better improved financial foundations, we can go more aggressive. In the first year, we uncovered some opportunities, but we had to balance the bottom line with our investment in R&D. That's one thing. The second thing is some of the R&D spending we have in the first year will start showing results and momentum in the second year. I don't want to miss again. In the first year, we played some catch-up; in the second year, I want to lead the market outright with the solutions we put out. But instead of focusing on each product line, we will focus on the principle the market needs: higher bandwidth, lower power, lower latency, and lower cost. The entirety of our portfolio is focused on these very fundamental attributes that we can offer to customers.

That concludes today's call. Thanks to all of you for joining us today, and we look forward to seeing you at various investor events over the coming weeks.

Thank you. And with that, this does conclude today's teleconference. Thank you for your participation. You may disconnect your lines at this time. Have a wonderful day.