MCHP - Microchip Technology Incorporated
Q2
Q2 2026
2025-11-06
+
Operator: Greetings, and welcome to the Microchip's Q2 Fiscal 2026 Financial Results Conference Call. [Operator Instructions]. As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Steve Sanghi. Thank you, sir. You may begin.
Steve Sanghi: Thank you, operator, and good afternoon, everyone. During the course of this conference call, we will be making projections and other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution you that such statements are predictions and that actual events or results may differ materially. We refer you to our press release of today as well as our recent filings with the SEC that identify important risk factors that may impact Microchip's business and results of operations. In attendance with me today are Rich Simoncic, Microchip's COO; Eric Bjornholt, Microchip's CFO; Brian McCarson, Microchip's VP of Data Center Business Unit; and Sajid Daudi, Microchip's Head of Investor Relations. I will provide a reflection on our fiscal second quarter 2026 financial results, then Brian will provide an update on our data center business, and Eric will go over our financial performance. I will then provide an overview of the current business environment and our guidance for third quarter of fiscal year 2026. We will then be available to respond to specific investor and analyst questions. I will now highlight a few salient points of our financial results. 6% sequential sales growth, net sales were up sequentially in Americas and Asia and flat in Europe, which is not bad for a summer quarter in Europe. Sales from our microcontroller and analog businesses were up sequentially. Specifically, our MCU business grew 9.7% sequentially with strong contribution from 32-bit MCU, while our analog business increased 1.7% sequentially. Our Gen 4 and Gen 5 data center products are seeing strong sales growth, albeit from depressed levels as customers seem to have finished their inventory correction. In the new products area, our blockbuster product announcement came on October 13 when we announced the industry's first 3-nanometer-based PCIe Gen 6 switch to power modern AI infrastructure. Brian McCarson will comment on this later in today's call. Our non-GAAP gross margin was up 236 basis points sequentially. Incremental non-GAAP gross margin was 95% sequentially. Non-GAAP operating margin was up 364 basis points sequentially. Incremental non-GAAP operating margin was 84.6% sequentially. Our incremental gross and operating margins are very positive. Inventory went down by $73.8 million sequentially. Calendar year-to-date reduction in inventory is $261 million. Inventory days were 199 days. Our inventory over 3 quarters has gone down from 266 days to 251 days to 214 days to 199 days. Underutilization in our factories in September quarter was $51 million. The product gross margin in the September quarter was 67.4% due to a rich product mix driven by data center products. We added $71.8 million of a new -- we added $71.8 million of new inventory write-off and $51 million of underutilization charge makes a total of $122.8 million of charges. Divide that by the net sales of $1.1404 billion and you get a non-GAAP gross margin impact of 10.8 percentage points. Subtracting it from the product gross margin of 67.4%, we got a non-GAAP gross margin of 56.7%, which is what we reported. So the product gross margin remains very healthy. We still need to bring down inventory write-offs and underutilization charges. We are pleased to announce that we have entered into a purchase and sales agreement to sell our Fab 2 wafer fabrication facility located in Tempe, Arizona to a third party. The sale of this facility is part of our previously announced plan to restructure our wafer fabrication operations. Under this restructuring plan, Microchip completed the closure of Fab 2 in May of 2025 and begin to transfer the process technologies from Fab 2 to Fab 4 in Gresham, Oregon and Fab 5 in Colorado Springs, Colorado, both of which facilities have ample clean room space for expansion. The transaction is subject to closing conditions and is expected to be completed in December 2025. Now we have a special guest for you today. Let me introduce Brian McCarson, Corporate Vice President of our Data Center Solutions business unit. Brian will speak about our recent announcement of industry's first 3-nanometer-based PCIe Gen 6 switch. Brian?
Brian McCarson: Thank you, Steve, and good afternoon, everyone. I'm the Corporate Vice President and Leader of the Data Center Solutions business unit at Microchip. And today, I'm excited to introduce you to the latest addition to our Switchtec family of products. Our new Gen 6 PCIe switch announced on October 13 marks a significant milestone in Microchip's technological leadership within the AI and enterprise data center infrastructure markets. The build-out of AI data centers continues to accelerate with hyperscalers committing to gigawatt scale deployments. Some recent announcements have outlined single infrastructure projects in the 5 to 10 gigawatt range, targeting completion between 2026 and 2027. These developments are driving our current design engagement cycles. It is critical to understand that regardless of whether our customers deploy NVIDIA, AMD, Intel or custom ASICs, all require high-performance PCIe switching infrastructure. This is where Microchip's Gen 6 Switchtec products are designed to excel. Last month, at the Open Compute Project Global Summit in San Jose, California, we introduced the industry's first PCIe Gen 6 switches manufactured using 3-nanometer process technology. These new devices deliver 4 distinct competitive advantages. First, PCIe 6 doubles the bandwidth to 64 gigat transfers per second per lane compared to PCIe 5.0, eliminating GPU to storage, memory and CPU bottlenecks that constrained previous generations. Our new Gen 6 switch features an industry-leading maximum of 160 lanes per device, significantly increasing total data transfer capacity. Second, our 3-nanometer implementation provides 15% to 20% power per lane advantage over competitors' products developed on 5-nanometer and older technology nodes. This is critical when deploying hundreds of thousands of GPUs and switches in multi-gigawatt data centers. Choosing Microchip's devices enables customers to lower total power consumption without compromising performance. Third, all our Gen 6 Switchtec devices offer advanced device telemetry and multicast capabilities, allowing a single GPU data packet to be transmitted to multiple devices simultaneously, thereby improving GPU efficiency. And fourth, we have implemented a secure boot-based hardware root of trust that supports post-quantum cryptography and is CNSA 2.0, the commercial national security algorithm suite compliant, meeting or exceeding both government and commercial security requirements. This represents industry-leading device security. We are now sampling these products to qualified customers and recent engagements have been validating both our technical approach and our market timing. From a financial perspective, we believe this represents a significant growth opportunity for the company. AI servers require substantially more PCIe switching infrastructure than traditional servers to enable resource pooling and the composable architectures that hyperscalers demand. Our total addressable market encompasses the entire data center PCIe fabric, not just a subset. We are vendor agnostic, selling into all data center and AI architectures. Design win cycles typically span 12 to 18 months from initial engagement to production, aligning our current sampling activity with initial production starting in June 2026 and volume ramping towards the end of calendar year 2026. Looking ahead, our Gen 6 Switchtec devices position us to capture a meaningful share of the committed AI infrastructure build-out across 3 growth vectors: hyperscale training infrastructure, enterprise AI training and inference deployments and high-performance computing applications. I will pause here and turn the call over to Eric for comments about our financials. Eric?
J. Bjornholt: Thanks, Brian, and good afternoon, everyone. We are including information in our press release and in this conference call on various GAAP and non-GAAP measures. We have posted a full GAAP to non-GAAP reconciliation on the Investor Relations page of our website at www.microchip.com and included reconciliation information in our earnings press release, which we believe you will find useful when comparing our GAAP and non-GAAP results. We have also posted a summary of our outstanding debt and our leverage metrics on our website. I will now go through some of the operating results, including net sales, gross margin and operating expenses. Other than net sales, I will be referring to these results on a non-GAAP basis, which is based on expenses prior to the effects of our acquisition activities, share-based compensation and certain other adjustments as described in our earnings press release and in the reconciliations on our website. Net sales in the September quarter were $1.14 billion, which was up 6% sequentially and $10.4 million above the midpoint of our September quarter guidance provided on August 7. We have posted a summary of our net sales by product line and geography on our website for your reference. On a non-GAAP basis, gross margins were 56.7%, including capacity underutilization charges of $51 million and new inventory reserve charges of $71.8 million. Operating expenses were at 32.4% of sales and operating income was 24.3% of sales. Non-GAAP net income was $199.1 million and non-GAAP earnings per diluted share was $0.35, which was $0.02 above the midpoint of our guidance. On a GAAP basis in the September quarter, gross margins were 55.9%. Total operating expenses were $549 million and included acquisition intangible amortization of $108.1 million, special charges of $6.3 million, which was primarily driven by our activities associated with our closure of Fab 2, share-based compensation of $53.3 million and $12.3 million of other expenses. The GAAP net income attributable to common shareholders was $13.9 million or $0.03 per share and was positively impacted by our settlement of an audit with the IRS dating back to fiscal year 2007. Our non-GAAP cash tax rate was 9.5% in the September quarter. We expect to record a non-GAAP tax rate of about 10.25% for all of fiscal year 2026, which is exclusive of the transition tax and any tax audit settlements related to taxes accrued in prior fiscal years. Our inventory balance at September 30, 2025, was $1.095 billion, which was down $73.8 million from the balance at June 30, 2025. We had 199 days of inventory at the end of the September quarter, which was down 15 days from the prior quarter's level, driven by our inventory reduction actions. Included in our September ending inventory was 16 days of long life cycle, high-margin products whose manufacturing capacity has been end of life by our supply chain partners. Inventory at our distributors in the September quarter was at 27 days, which was down 2 days from the prior quarter's level. Distribution sell-through was about $52.9 million higher than distribution sell-in. Our cash flow from operating activities was $88.1 million in the September quarter. Our adjusted free cash flow was $38.3 million in the September quarter. And as of September 30, our consolidated cash and total investment position was $236.8 million. Our total debt decreased by $82 million in the September quarter, and our net debt increased by $247.7 million. Our adjusted EBITDA in the September quarter was $341.8 million and 30% of net sales. Our trailing 12-month adjusted EBITDA was $1.103 billion, and our net debt to adjusted EBITDA was 4.69 at the end of the quarter. Capital expenditures were $36.5 million in the September quarter and included approximately $20 million for a building purchase in India, supporting our ongoing R&D activities in Bangalore. We expect capital expenditures for fiscal year 2026 to be at or below $100 million, and depreciation expense in the September quarter was $39 million. I will now turn it back to Steve, who will provide some additional commentary on our September quarter results and our guidance for the December quarter. Steve?
Steve Sanghi: Thanks, Eric. As you saw in the last quarter, our net sales continued to grow sequentially. We are continuing to see the inventory go down at distributors, at our distributors' customers, our direct customers and contract manufacturers. The distributor sell-in versus sell-through gap did not shrink last quarter. It was $49.3 million in the June quarter, and it was $52.9 million in the September quarter. The good thing about that is that distributor inventory went down even further. We expect that the distribution sell-in will eventually rise to meet the sell-through over the next couple of quarters. Next is gross margin. As I described in my summary earlier, product gross margins were very healthy at 67.4%, but our inventory write-off and underutilization charges knocked down the non-GAAP gross margin to 56.7%. We still need stronger sales to drive down inventory write-off and underutilization charges. However, our customers and distributors are taking advantage of short lead times and are continuing to drive down their inventory. Now to the market environment. We are seeing some recovery in our key end markets in automotive, industrial, communication, data center, aerospace and defense and consumer. They're all looking somewhat better. The strongest sales performance last quarter was in the data center market, albeit from depressed levels as the inventory at end customers and distributors corrected, we saw a large increase in bookings and shipments of our Gen 4 and Gen 5 products, which included PCIe switches, memory, flash controllers, storage and rate cards. We believe we are extremely well positioned with our Gen 6 PCIe switch with it being the only 3-nanometer-based device currently sampling in hyperscaler and enterprise data center customers, beating our competition in virtually every specification metric. Now let's get into our guidance for the December quarter. Our backlog for the December quarter started lower than the starting backlog for September quarter. The bookings for July were higher than bookings for any month in the last 3 years. August bookings were seasonally low, but better than our expectations and September bookings were quite strong as expected and the best booking month in 3-plus years. Overall, September quarter's bookings were 10% higher than those of June quarter. The book-to-bill ratio for the last quarter was 1.06. October bookings were higher than July, so we have a good start to this quarter's bookings, and November bookings so far are very strong. Embedded in these strong bookings is the observation that customers and distributors are scheduling these for March delivery and are continuing to lower their inventories into this calendar year-end. A comment about lead times. While lead times for our products have been 4 to 8 weeks for some time, we are continuing to experience lead times bounce off the bottom and are experiencing increases on some of our products. We're running into challenges on certain kind of substrates and subcontracting capacity and also some foundry constraints on very advanced nodes. These challenges remain isolated to specific areas. Our customer request for expedited shipments have increased significantly from a couple of quarters ago, pointing to some customers' inventories running low. I also want to remind investors that December is seasonally our weakest sales quarter of the year and is typically down low to mid-single digits sequentially. This is mainly due to a lot of holidays in the quarter and our customers shutting down their factories during the holidays. Taking all of these factors into account, we expect our net sales for the December quarter to be $1.129 billion, plus or minus $20 million, which would be down 1% sequentially at the midpoint. We expect our non-GAAP gross margin to be between 57.2% and 59.2% of sales. We expect our non-GAAP operating expenses to be between 32.3% and 32.7% of sales, and we expect our non-GAAP operating profit to be between 24.5% and 26.9% of sales. We expect our non-GAAP diluted earnings per share to be between $0.34 and $0.40. I want to highlight the operational discipline in our business model. Despite a seasonally challenging December quarter with expected slightly lower revenues, our operational improvements are expected to deliver strong profit performance. Non-GAAP operating profit is projected to increase by over $13 million sequentially at the midpoint of our guidance. This operational discipline is evident in our business model's ability to deliver significant flow-through of incremental revenue to operating profit in normal business environments. While we are only providing 1 quarter of guidance, and that is for this current December quarter, which is seasonally the weakest sales quarter of the year with a lot of holidays and customer shutdowns, we currently expect 3 strong quarters of March, June and September 2026. March quarter backlog is currently strong, and we currently expect March quarter sales to be stronger than a seasonal low single digit up sequentially. Finally, a comment on our capital return program for shareholders. Starting this quarter, we expect our adjusted free cash flow to be roughly even with our dividend payment driven by increasing profitability, low CapEx and liberating cash inventory. In future quarters, as we have excess free cash flow above dividends, we intend to use this to bring down our borrowings. With that, operator, will you please poll for questions?
Operator: [Operator Instructions]. The first question comes from Chris Caso with Wolfe Research.
Christopher Caso: I guess to start, maybe you could characterize what you're seeing now versus what you saw 90 days ago. At that time, you did talk about expectations for better than seasonal growth in both the December and March quarter. It sounds like December is on the better end of normal seasonality. But how do you feel now against what you thought 90 days ago?
Steve Sanghi: So I think as you have seen through a lot of industry announcements of our peers and competitors, the total business environment have taken a slightly softer tone. I think you saw that through most of this earnings season. And ours are really no different. Even though our December quarter guidance is better than seasonal, seasonal would be minus 3% to minus 5%, sometimes minus 5%, and we're only down minus 1%. I think if you go back 6, 9 months ago, I would have expected to continue to have small sequential growth even in the December quarter. But number one, the overall softer tone in the business environment; and number two, some impact of tariffs on customer psyche and people don't know when to make capital investments or not and people are holding back. I think a combination of all those things have led to this guidance we have given. Now if you look at the last quarter bookings, you ordinarily wouldn't think so. Our bookings were 10% higher. And if we had kept getting the turns at the pace we were getting in the prior quarter, September quarter would be -- December quarter would be a lot higher. But as I mentioned in my comments, we observed that customers are scheduling these bookings for March quarter and are continuing to decrease the inventory on their balance sheet, leading up to their year-end distributors as well as the customers. So this is sort of a strange push pull that's going on in the market. So we think we need to just hunker down for this one quarter, which is the weakest quarter of the year. And then we have strong momentum going into March and should be back-to-back from several good quarters.
Christopher Caso: Understood. As a follow-up to that, you mentioned that the product gross margins were holding up, but obviously, the charges are what's weighing on the gross margins. Could you give us an update on what you expect from both the inventory reserve charges and underutilization charges as you go through with presumably those next stronger quarters as you get into next year? How quickly can some of those charges start to roll off?
Steve Sanghi: So we, in general, don't know and don't guide those things for the future quarters. We just kind of look back and guide the actual that we experienced in the quarter. And current quarter is sequentially down 1%. So it's harder to make a whole lot of impact when your sales are actually minus 1%. We are looking for just a few days of reduction in inventory. And then I don't really have a guidance on utilization or inventory write-off in the current quarter. But if I look at it over the next several quarters, leading into stronger quarters of March, June and September, we currently expect to start ramping our factories at some point in time, start hiring people late in this quarter and then start ramping the factories. So as you ramp the factories because the inventory is coming down and as you ramp the factories, that will lead to lower underutilization. And regarding inventory write-off, the products we are writing off now is not from excess build. If you go back a year ago, 1.5 years ago, we were building lots of excess product above the sales. Factory footprint was very large. And we were writing off product because just even the freshly built product was really in excess of demand significantly. That's not what's happening today. Our utilization in our internal fabs is quite low actually and inventory is coming down. But what's happening is a lot of the product we built 2 years ago based on then customer demand, in many cases, the mix has shifted and some of that product is slower moving. So once it becomes 2 years old, we have to write off the rest of the product. So that is what's driving some of these write-offs in every quarter. It's not that the sales are 0 on those products. They will sell off, but they are not selling at a pace which would prevent them some of it from being written off. So I think in either case, we are -- we're really in the eighth or ninth inning on this entire inventory write-off as well as the underutilization. This one quarter, which is the weakest quarter of the year is going to hurt sort of where it falls in the year. But after this quarter, I'm quite optimistic that we're going to have back-to-back good quarters and we will make a meaningful difference in inventory write-offs and utilization.
J. Bjornholt: I think I'll just add a little bit to what Steve said. So we did talk about how our plan was to ramp output out of the wafer fabs in the December quarter. We are still doing that. So it is increasing. The impact on underutilization charges in the current quarter will be modest. They'll be down, but just ever so slightly. And then our assumption is that the inventory reserves will come down. But as Steve said, it is hard to predict. But we are guiding a pretty nice sequential increase in the non-GAAP gross margins to like 58.2% at the midpoint of guidance. So we are still seeing some benefit there in a tough quarter.
Operator: Our next question comes from Tim Arcuri with UBS.
Timothy Arcuri: Steve, so I wanted to understand just kind of what's going on. I know it's sort of this weird environment where December is soft, but people are booking out into next year. And you can kind of see it in the backlog. I know you stopped giving backlog breakouts in the filings. But in March, your current backlog was very low and a lot of it is parked in LTSAs still. So what are these LTSAs? I mean, it's coming down very slowly. Why is it taking so long to bring these down? And I mean, if customers want product, I would think that like what's the point of parking stuff out in March and June and giving you no visibility in the near term?
Steve Sanghi: We don't have any more LTSAs out there. When I came back last year, we actually here this month, very rapidly, we dismantled our program and essentially removed many of the customers' obligations on these long-term projects. And we took some cancellations. We took pushouts. We essentially allowed the customer to reset their backlog. So there's no impact on LTSA today. Customers are not taking any product they do not need. But...
J. Bjornholt: Steve, let me just insert one thing there, Steve. So you're referring to PSP. We still have some of these LTSAs in place, and we've been flexible with customers in terms of pushing out their requirements. They might have put in a 5-year expectation with us, and we aren't holding them accountable for taking that inventory and allowing flexibility to push that out by a year or 2 years, whatever they need to keep that engagement strong. And that's what Tim is referring to that he sees in our public filings on the LTSAs. They are coming down, but coming down slowly.
Steve Sanghi: Okay. So I was meaning that we dismantle the PSP program. And on the LTSAs, we're not forcing customers to buy anything that they do not need. We are basically being very flexible. They can buy what they need, and we're not forcing them to buy what they don't need. So that's not really causing any kind of problem. I think this is just -- I would have hoped that where the customers' inventories are low, they will take substantial intake at direct customers as well as distributors in the December quarter. But what we are observing is we're getting strong bookings, but they're scheduling it in March. And just basically, lead times are fairly short and they're taking chances and driving the inventory lower than I would have thought they would do.
Timothy Arcuri: Got it. So I guess just from the perspective of like what the point is of even having these LTSAs there because they're barely coming down actually. So really, the only thing that matters is kind of the current portion of what's -- I mean, you're not disclosing backlog anymore, but the current portion as of March was actually pretty small. So the LTSAs, like what's the point of even having them if they don't provide you any coverage in a quarter like December?
Steve Sanghi: It basically incentivizes the customer to continue to design with us. If you are sitting on a push where they can use our part or they could use TI or NXP's part and they're equally good and prices are similar, if they have an LTSA with us, I think that breaks the push. So there are those kind of benefits in engagement. But in general, it's no longer providing us any extra visibility, and we're not forcing the customer to take the product they don't need. [indiscernible] What got us into the trouble in the first place.
J. Bjornholt: Right. One thing I want to clarify, Tim, is we have not changed anything in terms of what we disclosed in terms of backlog. That is a requirement that we put that number in our 10-K, so we do that once a year, but it has never been disclosed to my recollection in our quarterly 10-Q filing. So it's once a year thing that we do with the 10-K filing.
Timothy Arcuri: It was in the Q actually last year, but totally get it.
Operator: The next question comes from Vivek Arya with Bank of America.
Vivek Arya: Steve, I'm curious what's driving your confidence to expect the next 3 quarters to be above seasonal? Your lead times are still low. There are so many macro cross currents. And most of your peers, as you mentioned, founded a little more defensive and were hesitant to guide more than the current quarter. So I'm curious, what are you seeing that they are not seeing to suggest that the next 3 quarters would be above seasonal?
Steve Sanghi: So March quarter is driven by just the visibility of the backlog. If you look at our backlog today for March quarter and compare it to what the December quarter backlog was, on August 6, I think that would be a 1 quarter difference, right? The backlog for March quarter today is much higher than December quarter backlog was on August 6. And the bookings that are coming in, its turns component into the March quarter is very strong. So March quarter, my comment is largely driven by visibility and the rate of bookings and turns. Now beyond that, I think customers are stretching their neck a little bit, distributors true by taking inventory down this quarter, which really they should not. In many cases, where the inventory is low enough, but they are basically dressing up their balance sheet for the end of the quarter and want the product in March, with the customers' inventory having come down significantly and distributor inventory coming down significantly, there is still a $50 million gap in sell-in and sell-through, which we think some will correct in March and the balance will correct in June probably. And then the rate of bookings is likely to continue as customers replenish their inventory. So my outer quarter commentary is driven by just the inventory will even get lower and they will need the product for June quarter and September quarter. And June and September quarter are historically our 2 strongest quarters of the year. They were both up nicely this year and many times, they're up usually even in soft years. So I'm less concerned about June and September and March quarter is driven by the visibility.
Vivek Arya: Understood. And for my follow-up, gross margins are going up nicely in December. Is that mostly utilization driven? I think on the last call or before that, you had mentioned that you expect it to take utilization up by 15%, 20%. I'm wondering what level are you taking it up? And do you expect to increase it again in March? And is there like a target level of inventory in dollars or days that we should think about until which point you will be more careful with taking utilization up further?
Steve Sanghi: Eric, can you take that?
J. Bjornholt: Yes. So I mean we're essentially managing our manufacturing output on a weekly, monthly basis. And so not going to break out the percentages that we're going to increase. But we're shipping out of our factories an amount that is significantly higher than what we are producing. And so we can't let that get too out of whack. So -- because we just can't ramp the factories super quickly, so it will be over time. So I would expect that we'll continue to ramp the fabs that those are decisions that we can make as we go through each month and look at the environment and see where inventory and backlog and revenue expectations are. The second part of your question was what, I'm sorry?
Vivek Arya: Sorry, do you expect to increase it again, Eric, in March, right? If you're expecting several above seasonal quarters, then does it mean that there's an expectation that you'll continue to increase utilization?
J. Bjornholt: Yes. We will need to continue to ramp the factories over time. It probably won't be a steady increase. It's just going to depend on the environment. But in the case that March, June and September have revenue growth, we would definitely be doing that.
Operator: The next question comes from Joe Quatrochi with Wells Fargo.
Joseph Quatrochi: I was curious if you could comment, is there any specific end markets that you can point to that you're seeing this kind of push, pull more than others?
Steve Sanghi: I don't really know if there can be an end market-specific commentary on it. I think we're getting bookings across the board on most segments, bookings are fairly strong, but the bookings delivery requested is in the March quarter, leaving the December quarter as per our guidance.
J. Bjornholt: Yes. As you know, we don't break out end markets on a quarterly basis, and it's a little more difficult for us to track it. We break it out once a year, but it seems like this is a pretty broad-based phenomenon that we're seeing.
Joseph Quatrochi: Okay. Fair enough. And then on the Gen 6 PCIe switch offering, you talked about maybe tightness at leading-edge wafers. Curious being at 3-nanometer, what's your kind of line of sight in terms of wafer availability as you look to ramp that in the second half of next year?
Steve Sanghi: Brian?
Brian McCarson: Yes. So we see a really healthy long-term strategic relationship with our foundry supplier in 3-nanometer, which is TSMC. And we believe we have the line of sight to the capacity that's needed to support our customer needs.
Operator: The next question comes from Blayne Curtis with Jefferies.
Blayne Curtis: I wanted to just ask because maybe I had this wrong. I thought that the inventory charges were supposed to go away pretty sharply into the end of the fiscal year. I guess you're guiding to continue in December. So I guess what changed? I guess, as forecasts start going up, I thought that the kind of the level of inventory would match the increased forecast and you wouldn't have this charge. I know you're saying the mix is now different. So is that what's changed? And kind of how far out should these inventory charges extend?
Steve Sanghi: I think with this weak quarter of December, we got to get through. But after that, I think we should in the stronger quarters, start to sell the inventory, significant inventory and have the charges really start to drop. Honestly, I expected charges to drop a little more than they have, and this weak quarter isn't helping. But I still feel that the inventory charges will come down rapidly as the year-over-year sales growth improve. So I think that's the key thing, and I've talked about it before, because you take the prior 1 year of sales and multiply it by 1.5 to get 18 months equivalent, and then you compare your inventory to that number. And if your inventory is higher than that number, then you have to write off the balance. So while our sales have been improving in the last 2 quarters, our year-over-year sales have been negative. So every quarter, the last 12 months sales have been coming down. And therefore, it's 18 months equivalent has been coming down. And that's what kind of has been driving some of the inventory charges. And starting this December quarter, that phenomena is reversed year-over-year starting to grow. When that happens, I think it will have impact on write-offs, inventory write-offs coming down.
Blayne Curtis: And then just to wrap some math behind that, I guess, the current impact is around, call it, 5% percentage points to gross margin. So you said it should come down over the next several quarters. So is that the right way to kind of add back that 4%, 5% headwind to the 58% that you just guided to? And then I'm assuming utilization would help by a few points as well. Is that the right way to frame gross margin over the next couple of quarters, 4 quarters?
Steve Sanghi: So that's exactly the right way to frame. So if you look at for September quarter, add the underutilization and the inventory charges, they added up to $122.8 million and divide it by revenue, that was 10.8 percentage point impact. I think as those charges come down, gross margin goes up dollar for dollar essentially. And that's our path to a 65% gross margin. So the product gross margin last quarter was 67.4%. So from a product gross margin standpoint, we're actually ahead of our longer-term target. But we just got to have -- these charges need to come down. And this current soft quarter isn't helping, but I think we'll regain momentum starting the March quarter.
Blayne Curtis: Thanks, Steve.
J. Bjornholt: And we have said this publicly, but that $71.8 million charge we had this last quarter, that charge never goes to 0. There is always some level of inventory reserves that are taken. And we do believe that at some point in the future, we will start to get a benefit of selling through a higher level of what's previously been written off. We just don't have line of sight to that, and that is hard to predict, but there's always some level of charge.
Operator: The next question comes from Harsh Kumar with Piper Sandler.
Harsh Kumar: Steve, the analog business in September quarter, the analog business was a little bit slow to catch up. Most of the sales came from microcontroller on the incremental side. Is that how you see the December quarter shaking out? Or do you expect a little bit more of an even contribution from -- like a similar percentage contribution from MCUs and analog?
Steve Sanghi: Harsh, if you look at the June quarter, those things were reversed. Analog growth was a lot stronger than the microcontroller growth was. And either you or somebody else asked the exact same question that analog grew more and microcontroller grew less and is that what you want to -- is that what you will see in September quarter. In September quarter, they reverse. Microcontroller did better and analog did worse. So these things go back and forth. I think these are both large product lines and thousands and thousands of customers. And just everything is not perfectly linear. In some quarters, one is higher, other quarters, different one is higher.
Harsh Kumar: Fair enough. And then when I think of -- at least I think of Microchip, I don't think of 3-nanometer leading -edge products as something that you're focused on. You're obviously highlighting it here in the earnings call. Is this a sort of a strategic shift where you will target more of leading-edge data center products? Just some color on your strategy would be helpful here.
Steve Sanghi: Yes, you should absolutely take that as a strategic shift. We hired Brian McCarson into Microchip. After I came back last year, I think, Brian, you joined us in when exactly?
Brian McCarson: January of this year.
Steve Sanghi: January of this year, so it was just a couple of months later. And Brian is focused on getting our data center products to state-of-the-art in market positioning. And this one, the 3-nanometer Gen 6 Switchtec device is the first one. And you'll see a series of new devices coming in from this business unit, all state-of-the-art products to gain significant share in the fast-growing data center market. Now we are -- this is not the only place we're putting attention. We also formed an AI business unit. I think I talked about it some time ago. And this was just a few months ago. So you'll be getting some updates on what's going to come out of that group. We're also putting a lot of effort into our FPGA business unit. You'll be hearing some new product announcements in this coming year on our FPGA products. You heard our high-performance space computing announcement, I believe, where we're doing a -- under a NASA contract, we're doing the next-generation space computer. So yes, it's a strategic shift towards continuing to do what we're doing in microcontroller and analog and all these other products. But in addition, have a component of our business, which is across more advanced nodes, high-performance products, lower power, market-leading products with a much higher growth profile, thus pulling the overall CAGR of Microchip higher than you would otherwise expect. That is the strategic shift.
Operator: The next question comes from William Stein with Truist.
William Stein: Steve, any estimate or best guess as to when the underutilization charges and inventory write-downs get to sort of a normalized level where maybe we'd see the product margins just show on the non-GAAP P&L without a whole lot of adjustments. Maybe I guess that takes us to somewhere between 65 your target and where you're running now a little bit higher than that. Is that something we should expect sort of in the early part of fiscal '27, do you think? Or will it be further out?
Steve Sanghi: I'm not comfortable forecasting at this point in time, especially in an otherwise soft quarter. I think we will make substantial improvement in the next fiscal year. We'll make improvement in March. And then again, June is the start of the next fiscal year. So you're not directionally wrong. I just don't want to put an absolute time frame or an absolute figure.
William Stein: As a follow-up to that...
J. Bjornholt: I think we're confident in saying that the inventory write-offs normalize quicker than the underutilization goes away. Both will be moving in the right direction, but underutilization will take longer is our expectation, but we're not putting a time frame around it.
William Stein: That's helpful. As a follow-up, when I have discussions with investors, particularly ones bullish on Microchip, they look at these 2 charges and they say, well, those will go away at some point. But then also we'll get -- we'll get some leverage on the gross line, some gross margin leverage. So we'll actually see gross margins go higher than that level. And I wanted to check my understanding of this because I think the underutilization charges are designed to sort of simulate a 90% utilization level. So you may get higher margins, for example, from mix. But from utilization, can you correct me if I'm wrong, that we shouldn't see higher gross margins than the product gross margins you're referring to from utilization unless we get above 90% utilization. Is that approximately correct?
Steve Sanghi: Eric?
J. Bjornholt: Yes. So when Steve is quoting the 67% plus product gross margin, we would not recommend that anybody puts that into their models, quite honestly, as that's where we're going to end up. We have a 65% long-term model. We just did whatever it was, 56.2 -- excuse me, 56.7%, and we're guiding to 58.2% at the midpoint. So we've got a long ways to go. Each of our factories has a level of what we would call normal utilization. It could be 90%, the number you used in one factory, it could be 75% in another factory. And it's just going to depend on how the revenue mix comes in over time. Obviously, in the up cycle, we were running at 100% plus capacity in just about every factory. And at that point in time, there was other things that were helping and expedite charges and whatnot, where we got to 68% plus gross margin. But that is not standard by any means. And we're really focused on getting to the 65% -- and when we get closer to that number, we'll obviously reevaluate and provide future guidance to you guys.
Operator: The next question comes from Harlan Sur with JPMorgan.
Harlan Sur: Can you guys give us an update on the Fab 2 closure, rightsizing of Fab 4 and Fab 5 back in the June quarter? You guys were targeting about $115 million in cost savings annually at that -- at the current quarterly revenue run rate, that's about 250 basis points of gross margin improvement. Is all of that now accounted for in your current gross margin profile? Or is there still more on the come?
Steve Sanghi: Eric?
J. Bjornholt: Yes. So for Fab 2 specific -- for Fab 2 specifically, we talked about $90 million of annual cash savings, right? And we're on our way to accomplishing that. But as Steve indicated, with this agreement that we have in place that's still subject to closing conditions, if that closes in December, we'll get some cash from that, which isn't going to be disclosed at this point in time. But those costs really aren't impacting our non-GAAP gross margins today. The 2 biggest factors are the underutilization charges and the inventory write-offs. But it's great that we've got to this point with that factory and getting it completely out of our cost structure, hopefully, by the end of the quarter will be a nice step for us.
Harlan Sur: And then -- thanks for the update on your PCIe switching portfolio. On a segment reporting basis, data center and compute was 19% of your total revenues in fiscal '25. What's the rough mix of data center versus client or PC compute? I assume majority of the segment is data center focused. Any way to quantify it? Is it 60-40, 70-30? I mean these products are more application-specific, so easier to track. I assume you guys track this mix pretty closely, but any way to quantify the mix differences?
Brian McCarson: Yes. We don't break that out to that level of degree to everyone. Majority of that is related to data center more than anything within that bucket of 19%.
Operator: The next question comes from Joshua Buchalter with TD Cowen.
Joshua Buchalter: I also wanted to ask about the backlog visibility. So you mentioned more customers wanting to take orders in March than December, which I understand because lead times are short and they can. But given the backlog is down sequentially and you're expecting an above seasonal March after what seems like an above seasonal December, maybe you could spend a little bit of time talking about longer term, what's driving the confidence after March that you're going to see those 2 to 4 quarters of very strong results? Is it demand signals? Is it expectations of inventory restocking because you're nervous that -- or your customers get nervous that things are too low? Just trying to understand what signals you're seeing as we get into '26.
Steve Sanghi: So I think as we get on the other side of the December quarter, you have several wins on the back kick in. Number one, the difference between sell-in and sell-through has to close. Distributor inventories are going to become normal here. They have come down very significantly. There is maybe a little bit to go here and there, but they're largely getting corrected. So when the correct in another quarter or so, you have a $50 million sell-in to sell-through gap that needs to close. So that is a wind on the back. And it will close by people buying more product on sell-in, which is a GAAP revenue. The second is the same phenomena on our direct customers, contract manufacturers. As their inventory corrects, they will start buying what they're consuming. Today, they're buying much less than what they're consuming. So that's another wind on the back. And third, June and September are seasonally strong quarters. December is our weakest quarter. March is the next better and then June and September usually are strong. So we have strong quarters coming up, plus we have this inventory phenomenon in distributors, direct customers and contract manufacturers. All of that together, I think we'll have a decent financial performance.
Joshua Buchalter: And then I also wanted to follow up on the 3-nanometer PCIe switch. Maybe you could give us a few -- some details on what's your expected go-to-market and sort of how mature the commercial engagements are. You're doing 3-nanometer, your peers are at 5-nanometer. Are you competing on performance versus cost as a result? And should we expect meaningful revenue contribution in 2027 with the part coming out at the end of 2026?
Steve Sanghi: Brian?
Brian McCarson: Yes. Several really good questions there. We are targeting this family of products at the hyperscaler, the enterprise OEM and ODM customers. So we're covering all segments of the -- both AI data center and enterprise data center markets with these products. And in addition to the customers that I mentioned, we see a sizable market. We think the total available market for our Switchtec family of devices that we serve should exceed $2 billion per year today. And while there's a lot of variation in different expectations for growth, we're expecting greater than 10% CAGR on that total available market through 2035. So given that we expect to release to production by June 2026 and typical design win cycles with customers, 2027 and the latter part of 2026 is when first revenue should be appearing.
Operator: The next question comes from Chris Danely with Citibank.
Christopher Danely: So Steve, just going back to your earlier comments, you said that you expected the December quarter to be a little better and things seem to get a little softer at some point in the September quarter. Can you just talk about, I guess, when that happened? And did any particular areas like geos or product lines or anything stand out on the softer side and why you think that happened? Do you think there were some tariff-related pull-ins earlier in the year? Or is something else happening?
Steve Sanghi: So I think our bookings remain strong. July was one of the best booking months in 3 years. And then August was slower, but still better than expected, and I have talked about the August was always slow because of the holidays and vacations and all that, but it was better than expected. And September was very, very strong, best booking months in 3-plus years. And the book-to-bill ratio was positive. The bookings were up 10% sequentially. The only reason why I say December is a little softer and disappointing is the customer decided to give us strong bookings, but schedule them in January and essentially address their balance sheet for calendar year-end. So therefore, the turns component of that wasn't as strong as I would have expected. With the inventory corrected, I thought customers and distributors will restock. Well, they decided to restock 3 weeks later, not restock on December 26, but restock on January 15. And that's kind of the difference we're talking about. Therefore, the backlog for March looks good. On certain product lines, the March quarter backlog is stronger than December quarter backlog. Not across the board, but on some key product lines. So this is the observation which changed what I thought would happen. I thought with all this inventory correction, December quarter would be a lot stronger, low single-digit up rather than minus 1%. But this phenomenon that they decided to buy that product in January rather than buy it in December has changed that equation.
Christopher Danely: Okay. And just my follow-up. So you mentioned some constraints. It sounds like they're still mostly on the back end. Are those constraints getting worse? When do you think you can get them under control? And is it causing you to miss out on any sales?
Steve Sanghi: So, it's not -- nothing is disastrous. We are managing it, but the substrate capacity is the one that was extremely constrained, if you remember, a couple of years ago, 3 years ago. And now as these advanced products have come in, all the AI products and products from all these companies going into high-end data centers and all that, they all require substrate capacity. And in the last quarter, we were also competing with cell phone builds for the December quarter because they all require substrate. So we were competing with some significant demand. Now that demand is for new product launch, and that is over now. So some of those constraints have gotten less so, but it's still really touch and go. I think, yes, in certain cases, customers wanted the product in September or in December, and we're shipping it a quarter later, yes. It's not hundreds of millions of dollars, but it's meaningful.
Operator: The next question comes from Joe Moore with Morgan Stanley.
Joseph Moore: I wonder if you can just help us try to triangulate where the demand actually is. Looking at your bookings levels, 1.06 is pretty good, but your revenue is about half of what it was at the peak. And I guess the real consumption is somewhere between that peak number and where you are now. But just as you look at the bookings pattern, do you have any updated sense on where we're going to get when we get to consumption levels?
Steve Sanghi: I cannot help you or anybody else to figure out where the exact consumption level would get to. I would agree with you that it's somewhere between the peak and where we are. But which one it is closer to where it is, I think that's a million-dollar question. Honestly, this entire recovery has been a lot slower than anybody would have expected. I think all these tariffs and customer concerns about capital investments in automotive to EV to gasoline shift, and there have been a lot of curves that have been thrown at this market. And the overall progress has been less than I personally would have wanted. So we'll continue to make that progress. I think pick up pace in the March quarter, but I'm not willing to guide where we eventually reach equal to consumption. I think that remains a challenging exercise.
Joseph Moore: Okay. Fair enough. And then with my follow-up, the data center products, when you talk about the new switch and things like that, where do you guys stand in terms of hyperscale cloud relationships? Do you have partnerships there where they're coming to you and saying, here's the product we need. Can you help us to develop that? I know you have a lot of data center businesses that were acquired when they were sort of enterprise-centric. Can you just update us on where you are with going to market with these bigger cloud customers?
Steve Sanghi: So Brian, we do business with all of them. Why don't you take that question?
Brian McCarson: Yes. So we have active engagements with all the hyperscalers and OEMs you would expect across this broad both enterprise and AI data center market. We do not work on custom ASIC products as our main business. So we instead focus on understanding the workloads that our customers are most interested in accelerating and optimizing within the data center and build the right competitive features to best meet their needs. And I think this latest announcement around our first-to-market 3-nanometer Gen 6 PCIe switch demonstrates that with industry-leading security features, industry-leading telemetry, industry-leading power and performance per lane. And so we will continue to build our products with the hyperscaler OEM, ODM and enterprise data center markets directly in line.
Operator: At this time, I would like to turn the call back over to Mr. Steve Sanghi for closing comments.
Steve Sanghi: Yes. Thank you very much, everybody, for hanging in there. And as I said, after this quarter, I think we should get back strong momentum, and we'll see some of you on the conference circuit this quarter. Thank you very much.
Operator: Thank you. This does conclude today's teleconference. You may disconnect your lines at this time. Thank you for your participation, and have a great day.