Greetings and Welcome to Microchip's Business Update Call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. If anyone should require operator assistance, please press star zero on your telephone keypad. As a reminder, this conference is being recorded. It is now my pleasure to introduce Mr. Steve Sanghi, President and CEO. Thank you. You may begin.
Thanks, Operator, and good morning, everyone. Before I begin today, I wish to remind you that during this presentation, I will be making a lot of projections and forward-looking statements. These statements are predictions, and the actual results may differ materially. I refer you to Microchip's filings with the SEC that identify various risk factors in our business. In this presentation, I will share a fair amount of data and information that we have not routinely shared with investors. The reason for sharing this data is to validate our confidence on the path we are on for recovery in our business. I have named our recovery from the depths of this prolonged inventory correction as Up and to the Right 2.0. If you read my book, Up and to the Right, then you would know what it means.
This next slide is a longer safe harbor statement that you should read before you review the rest of the presentation. The next slide, I'll begin with giving you a business update. Our March quarter net sales are on track to our guidance. We believe that our distribution sales will bottom this quarter, and the distribution inventory correction is nearing completion. The direct customer inventory is a bit harder to pin down. However, we believe that we are at or near the bottom of this prolonged inventory correction. Our bookings have seen a meaningful improvement, and I will show you that in a graph. Our total backlog has also stabilized and is showing a slight upward slope. I will also show you that in a graph. So this first one is net bookings history.
Sorry, I hit the scale, but it is clear that January and February bookings shown by the two green bars on the right are meaningfully higher than any other months in the last year. This next slide shows our total backlog. This backlog has been declining for over two years, but in the last six weeks, it has shown stabilization and then a slight slope up. There's no hockey stick here, but stabilization is the first phase, and we are there, and we expect that we will continue to see improvement in the total backlog from here. With that summary, I will give you an update on each of the nine points of our recovery plan that I highlighted to investors in early December at one of the sell-side conferences.
These 9 points are first to resize the manufacturing footprint, second to bring the inventory down, third is to review our growth initiatives, fourth is to do a business unit by business unit deep dive to see what needs to be changed, number five is to review our channel strategy and programs worldwide, sixth was to strengthen our customer relationship, seventh is to confirm or change our long-term business model, eighth streamline operational expenses under the new model, and ninth review our CHIPS Act. So first is to resize the manufacturing footprint. Our manufacturing footprint is too large for the size of our business, resulting in significant underutilization charges every quarter. So we are right-sizing our manufacturing facilities, and we're closing one of the fabs called Fab 2. We announced the shutdown of our Fab 2 in early December.
When we initially made this announcement, we projected the Fab 2 closure to be in September 2025. But we have pulled the shutdown into May of 2025. Many of the technologies from Fab 2 that we have been bringing up in Fab 4 and Fab 5, those efforts have gone extremely well. In some cases, the very first checkout run has yielded near perfect. Due to such efforts, we have pulled the shutdown into May of 2025. The last wafer starts in Fab 2 were made almost two weeks ago. The closure of Fab 2 will save approximately $90 million annually, and the Fab 2 facility and equipment will be an asset held for sale. In our other two fabs, Fab 4 and Fab 5, we are right-sizing them through a layoff. We will size down both fabs below the near-term demand rate, thus allowing the excess inventory to correct.
Once the excess inventory is bled out, we will ramp the fabs to their demand rate. In Fab 4, we're doing a slightly smaller layoff and then keeping a two-week rotating time off for two more quarters. This will allow us to ramp the fab back rapidly as needed by removing the RTO. The layoff in Fab 4 and Fab 5 will save us approximately $25 million in employment-related costs annually, but in total cash charges, it saves more than that because you're making less wafers and other operational expenses are lower. In our back-end facilities, we have two back-end facilities in Thailand. In these facilities, we stopped replacing turnover in fiscal Q1 of 2024. With normal attrition, the headcount has come down significantly. We expect that with projected turnover, the headcount will be corrected by the end of June 2025.
Therefore, no layoff is needed in this facility because the headcount will correct in the next three months. Our two other back-end facilities are in the Philippines. Here also, we stopped replacing turnover in fiscal Q1 of 2024. However, the headcount here has not corrected due to lower turnover. Therefore, we are planning a layoff to right-size this facility. The second of the nine points recovery plan was to bring our inventory down. This slide shows our inventory in days. The red bar was the last quarter actual number of days of inventory. It was 266 days. It was an all-time record number of days, not the record that we are proud of. Our normal days of inventory would be 130 to 150 days. The bars after the red bar show the forecasted decline of inventory in number of days over the next three years.
We have made certain assumptions for production output and revenue built into these projections, and actual results could differ materially. We are forecasting more than a $300 million reduction in inventory by March of 2026. As the manufacturing sites are right-sized and the revenue assumption in this graph may be conservative, we may be able to decrease the inventory faster. Decreasing the inventory is very important to liberating cash. Foundry long-term agreements. So we also have high inventory on wafers that we purchase from our foundries. Part of the reason for high inventory is our commitment under the long-term agreements called LTAs. Microchip signed several of these LTAs at the peak of the supercycle, and there are shortfall penalties in these agreements.
With the demand down significantly, we have renegotiated the cancellation or modification of some of these LTAs, and we expect to take approximately $45 million charge in the March quarter for cancellation or modification of these LTAs. We do not expect any large cash outflows from these changes. It will be mostly a portion of the forfeiture of amounts paid in advance. The third of the nine-point recovery plan is the review of our growth initiatives, Total System Solutions, Megatrends, and artificial intelligence. We'll begin with Total System Solutions. We call Total System Solutions TSS, in abbreviation. In our TSS strategy, we have anchor products and then the products that attach to our anchor products on the same board. Currently, MPU-32, data center products, networking products, microcontrollers, and FPGAs are the anchor products. We have recently added two more anchor products. The first is PIC64.
It is a 64-bit microcontroller with $5 billion of TAM, with a 10% compounded annual growth rate. The second is Ethernet 10BASE-T1S with $4 billion TAM by 2030. This next slide shows some data on our TSS on our anchor products. This is the data we are sharing for the very first time. I'll be sharing significant more data in the upcoming slides. Please don't make me regret it. The slide shows by year our TSS indicator, which is devices per project for various anchor products. If you look at it closely, then MPU-32, which is a first set of bars, MCU16, which is a fourth set of bars, and UNG, which is a third set of bars. Those are more established Microchip product families.
While achieving high TSS in the past in new designs, the level of TSS achievement on those particular product families is somewhat saturated since they're many years old and a high TSS has been achieved. The other two set of bars, which are the data center products and the FPGAs, came to us from prior acquisitions, and with limited Microchip TSS at the time of the acquisition, our average TSS level on those products continues to grow significantly. As you can see in the second set of bars, which is data center products, and fourth set of bars, which is FPGAs, the TSS is rising dramatically. On the average, we're achieving about 4-5x TSS, which means in addition to the anchor products, we're adding an additional four to five products around those anchor products.
As we add PIC64 and 10BASE-T1S Ethernet, these products should go through significant TSS score increase in the future. And I'll give you a little bit more detail on these products. The first of these new anchor products is the 10BASE-T1S Ethernet. Now, Ethernet is rapidly becoming the choice for connectivity. 10BASE-T1S Ethernet will be replacing CAN Bus, RS485, RS232, FlexRay, and many other wired protocols in cars, industrial, and defense networks. We expect that our Ethernet anchor products will drag along significant attach of our other catalog products. The second anchor product is PIC64, a 64-bit microcontroller. On upper left, it shows several PIC64 products with scalable compute performance. On upper right, it shows the six end markets where this product will be designed in and utilized. On lower left, it shows scalable radiation performance. Some product versions are radiation tolerant. Others are even stronger.
They're radiation hard. On the lower right-hand side, it shows where the scalable radiation hardness products will go. The amount of hardness required depends on whether the products are going into low-orbit satellites or spacecrafts or other deep applications like Mars rovers and moon landers. PIC64 as anchor product will drag along significant attach for other catalog products. Continuing on the revenue growth initiative, the next growth initiative is a focus on Megatrends. The sales growth rate in our Megatrends is more than 2X the growth rate in our non-Megatrend applications. Just reminding you, our six Megatrends have been IoT, data centers, 5G, sustainability, electric vehicles, and ADAS, which is advanced driver assistance. Going forward, we have made two changes. First, 5G is nearly saturated, so we have replaced 5G with artificial intelligence and machine learning.
And the second change is that we have replaced ADAS with connectivity. ADAS was only in automobiles, but connectivity is a broader Megatrend. Ethernet is making inroads in more than just automobiles. It's getting designed into industrial, automotive, aerospace, defense, and home applications. And here is some more data. This slide shows a compound annual growth rate for Megatrends versus non-Megatrends. Again, we are sharing this data for the very first time. From FY 2021 to FY 2024, our total revenue grew at a compound annual growth rate of 11.95%. In the same period, Megatrend revenue grew at 25.78% compound annual growth rate, which is more than 2X the overall growth rate. And the non-Megatrend revenue grew at only 3.51% compound annual growth rate. Our FY 2024 was a down year.
FY 2024 ended in March of 2024, and it was a down year. In that year, total revenue declined by 9.4%. In fiscal 2024, the non-Megatrend revenue declined by 14.6%, while the Megatrend revenue declined by only 2.9%. So even in a down year, the Megatrend does very well and declined only slightly by 2.9%. So your obvious question likely is, how does FY 2025 look like? We do this breakdown once a year. I will have FY 2025 data in May of this year and will share with you. My estimate is since 2025, or fiscal 2025, is a significantly larger down year. The non-Megatrend revenue would have declined much more than the Megatrend revenue. So this slide just shows our Megatrends in a more pictorial format.
Again, we have replaced 5G with AI/ML, and we have replaced ADAS by networking and connectivity. ADAS remains a part of this network and connectivity Megatrend. We have also changed the name and the label of electric vehicles Megatrend to e-mobility. E-mobility, of course, includes all the electric vehicles, but it also includes trucks, buses, trains, scooters, motorcycles, and three-wheeler vehicles in various countries. It also includes all the charging infrastructure, so e-mobility is a slightly broader Megatrend, and on the right, it shows the 6N markets where we are focused on. Continuing on our growth initiatives, this slide is about AI Coding Assistant in which we are very first to market. On February 19th, Microchip released an artificial intelligence Coding Assistant to our customers. It is inside Microchip's firewalls, so you have to log in to get access.
This coding assistant has access to all of our reference designs and application code that we have written over years. A customer has to just type in a few commands to tell what he wants to do, and the coding assistant will generate an executable code in less than a minute. It can automate several manual and labor-intensive tasks. It can also get access to Microchip's data sheets without actually leaving the editor, which, if you ask an embedded designer, it's extremely valuable. We believe that this AI coding assistant can increase the productivity of an embedded designer by more than 40%. And we are very first to market in this one among any of our competitors. I also want to update you on our China for China strategy. As you know, the Chinese government has been pushing the local companies to use more of the locally available semiconductor products.
This initiative, we call it China for China. We are partnering with a Chinese company that we have selected, and we'll call this company ChinaCo till later down the line when we can release its name. We will sell our integrated circuit die to this company called ChinaCo. ChinaCo will get it assembled and tested at Chinese assembly and test subcontractors. ChinaCo will then sell that product to the Chinese market with a Chinese brand, Chinese owners, Chinese website, Chinese data sheets, Chinese logo, and Chinese development tool, and local Chinese support in Chinese language.
We will do it mainly on our catalog products and the products where our Chinese customers are requiring a local source, but this can easily grow into any other products at the request of a customer that we want to flow through and have look and feel of a Chinese product. The fourth point in the nine-point recovery plan is a business unit by business unit deep dive to see what needs to be changed. So we had 22 business units. They're really essentially product lines. They all build different kinds of products. We have completed a deep dive of all 22 business units within the 90-day period that I had promised. It resulted in several organizational changes. There are three business units that have been combined into other business units. So now we have 19 business units, but we also added a new AI/ML business unit.
It is just getting started, so it's not included in the 19. So here are the organizational changes that have resulted from deep dive reviews of the business units. First, we have combined our 8-bit MCU and 32-bit MCU business units together into a general-purpose MCU business unit. This will streamline MCU strategy and will accelerate the conversion of our 8-bit MCU customers to 32-bit for whoever wants it. This will also offer a new generation of AI development tools because we have also moved the development tool group into this 8 and 32-bit MCU business units. Second, we have combined our Power over Ethernet business unit into our Ethernet networking connectivity business unit. It adds sufficient critical mass to address a $4 billion TAM in 2030. Power over Ethernet was a very small business unit, not growing, but it had some very good resources in it.
We'll continue to milk the products and continue to have the revenue, but all development resources really would now be working on Ethernet network connectivity. Third, we have moved our 16-bit dsPIC business unit with the analog power business to address the intelligent high-power market. The 16-bit dsPIC and analog power, they often worked on the same board in the analog power business. Fourth, we have formed a new business unit. It is called the Central AI/ML business unit, which will drive common architecture and tools among all of our other Microchip business units. The fifth point in the nine-point strategy plan was a review of our channel strategy and programs. I have completed the review of the channel strategy and implemented the changes to our distribution program listed on this slide. We have made four changes to our distribution program.
First, we are auditing the actual customer ASPs, and we're comparing them to the ship and debit requests. If the actual customer ASP is higher than the ASP our distributor used in getting a ship and debit request from us, then we will claw back the extra margin from the distributors. Second, when we gave a demand creation registration to a distributor in the past, the demand creation flag was for the life of the design socket. We are now limiting the demand creation flag to a certain number of years, and after that, the flag reverts to a fulfillment flag. This gives the distributor a push to design our new products at that customer instead of sitting on a high-margin socket, which exposes the socket to competitors. Third, we had industry-high fulfillment margins.
We have lowered them to be on the high end of competitors but not well above the high end of competitors where we were previously, and fourth, we are reviewing all of our distributors for performance, and in some cases, we're terminating some small distributors for their lack of performance. Overall, distribution continues to be a very important piece of our overall go-to-market strategy, and it is about 45% of our business. Our distribution business is very complex with different distributors in many, many offshore countries. Obviously, you'll have a lot of questions on our distribution strategy, and I'll take those questions in the Q&A session. The sixth point in the nine-point program was strengthening our customer relationships, so we conducted a large-scale survey capturing insights on 6,775 customers, a very large number of customers.
When you look at that chart vertically, it shows a number of customers where Microchip is preferred supplier or approved supplier or not preferred supplier or blocked. So it breaks down the number of customers where we are preferred, approved, and not preferred or blocked. When you look at it horizontally, it shows the number of customers where our relationship going through the supercycle has either improved, unchanged, or declined. So several observations from this chart, and I've highlighted them in color so you can easily go to it. First, the green square. You can see that at 24% of the customers, our relationship actually improved. Most likely, at our competitors, it may be declined. So 24% of the customers, our relationship improved. At 64% of the customers, our relationship was unchanged.
At 12% of the customers in the small red square, at 12% of the customers, our relationship declined. Now, out of customers where our relationship declined, we are still preferred at seven customers, and we are approved at 396 other customers. That leaves out 348 customers where we are not preferred and 49 customers where we are blocked. That makes up only 6% of the customers where we have significant work to do. These 348 plus 49 equal to 397 customers are our priority to improve our relationships, which is 6% of our central customers. In strengthening client relationships, we first identified 293 high-priority clients. 30 days later, via the survey, we identified an additional 400 clients, which were just below that tier. These are the clients where our leadership is engaging with our customer's leadership.
We are applying first principles of empathy with the customer where they went through a difficult time during the supercycle, having candid conversations, taking ownership of the problems, and showing speed of decision-making and speed of working with them to get their new designs to production, and we're using a five-step conversation guide that I won't go into to strengthen our client relationships. The job is to complete the first 293 customer engagements in 100 days, which ends on March 31st, and in the second group, to complete the 400 by April 30. Now, as of just a few days ago, February 27, we had engaged with 235 customers in the first group and 166 customers in the second group. The result of these engagements is quite positive so far, and we are engaging with these customers worldwide.
There's a further breakdown of these customers by geography, which will be true detail for this presentation. The seventh point of the nine-point plan was to either confirm or change our long-term business model. After significant review, we find that the old business model had remnants of peak performance with significant expedite charges, higher prices, and peak factory utilizations. We find that that model is not sustainable. Therefore, we are updating our long-term business model. So starting with the revenue growth, we're giving guidance for our long-term revenue growth to be industry growth plus. We have undergone significant inventory correction. The baseline of our revenue growth is quite difficult to decipher at this point in time. Therefore, we're simply saying industry growth plus with the help of our Megatrends and TSS and other growth initiatives.
As we complete the inventory correction and have more visibility into our business, we may be able to provide further clarity of our long-term growth rates. On the gross margin, we are updating our long-term gross margin target to 65. Our last quarter's gross margin was 55.4%. And one of your obvious questions might be, "Well, how do we get to 65?" Well, last quarter, our gross margin was 55.4%. Last quarter, we also had $82 million of inventory write-off and $43 million under utilization charge. This totals $125 million. Even if we capture $100 million of this $125 million back with lowered inventory, no longer requiring write-off, and refilling up the factories with the right size, we get approximately 1,000 basis points of gross margin back, bringing this gross margin to 65%. So it is really quite doable.
On the operating expenses, we are updating our long-term model to 25%. Last quarter, it was 34.9, and on operating profits, we're updating our long-term model to 40%. The eighth point in the nine-point plan was streamlining our operational expenses towards a new long-term business model shared on the last slide, so this is our OpEx reduction strategy. These are some of the tools by which we will correct the non-GAAP operating expenses from 34.9% of our net sales last quarter to 25% long-term. As our revenue grows, we will continue to monitor operating expenses so operating expenses as a percentage of revenue can trend down towards our long-term model. We're also conducting a corporate-wide layoff starting today, targeting about 10% reduction from office employees. The layoff of office employees will save approximately $90 million-$100 million per year in operating expenses.
The final ninth point in the nine-point plan was to review our CHIPS grant. Soon after I returned as CEO of Microchip, I paused the CHIPS grant negotiations. The new leadership at the CHIPS office is just forming. Some people are leaving, and other people are coming. We have agreed with the CHIPS office to remain paused till the end of April as we complete our own restructuring and then review as we engage with CHIPS grant again. Let me summarize this presentation. In summary, the first 100 days since I came back as CEO have been quite hectic but also very productive. We have right-sized the factories, put inventory in a downward trajectory to liberate cash, reviewed and adjusted our growth initiatives. We did the deep dive of 22 business units and reorganized where needed. We have adjusted our channel strategy.
We have launched a program to strengthen customer relationships, unveiled a new business model, and adjusted operating expenses through a broad corporate layoff, and all of that in the first 100 days. We have also modified or canceled the long-term agreements with foundries to meet the business needs. All in all, I'm pleased with our progress, and I look forward to reporting more progress in the second 100 days. Thank you. Operator, will you please hold for questions?
Thank you. We will now be conducting a question-and-answer session. If you would like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star two to remove yourself from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. As a reminder, we ask that you please limit yourself to one question. Our first question comes from the line of Vivek Arya with Bank of America Securities. Please proceed with your question.
Thank you for taking my question. I appreciate the detailed plan. Steve, I guess the central question still remains, what is the right way to model revenue growth for Microchip? I think you are suggesting it's a little bit over industry. If I look at industry over the last five or 10 years, growth has been sort of in the mid-single digit. If I look at the average growth rate for your peers that Consensus has for this or next year, it's somewhere in the 5%-10% range. So I guess the question is, given all that information, how do you think about Microchip's revenue growth? And then specifically, what revenue has been assumed for that reduction in inventory chart that you showed? Thank you.
So Vivek, thanks for your question. We're not able to give you further details on assumptions we use in the revenue chart. That would be giving guidance too far out. We are still in the middle of an inventory correction and looking to form a bottom here. And as our business becomes normal, whatever that means, it simply means that there's no excess inventory in the channel or at customers, and we're doing a routine normal business with the customers, then we'll be able to more establish an ongoing growth rate. I just think that at the current point, we're not able to give you any more information on that. Thank you.
Thank you. Our next question comes from the line of Chris Caso with Wolfe Research.
Please proceed with your question. Yes, thank you. Good morning. I guess the question is with regard to some of what you talked about in the near-term booking trends. And if you could give us some color about why you feel that there may be some sustainability there. I know during this downturn, it's been really challenging that there've been some head fakes in the past. Why do you think this one's different? And if you could also comment on the proximity of the Chinese New Year holiday and what effect that may have had on some of the bookings since the beginning of the year.
So if you look at the bookings for January and February, Chinese New Year hurt the bookings negatively, not positively, because they're gone for a period of a week to two weeks. So the bookings are actually overall booking. They may book a little early, but then the bookings are lower usually during the March quarter, and March quarter usually is down in China because of less working days. The bookings increase very well correlates with the correction of the inventory in the distribution channel where we essentially feel that that correction is very much needing completion. The inventory at some direct customers is still there, and it's working its way down, but more and more customers are completing inventory every week or every month and are joining the fray where they are giving us normal amount of bookings, so the data is really not a head fake.
I am and I will remain conservative in giving you guidance, so when I say and share the actual data of bookings, it's meaningfully better than the history we have had, and there are good reasons to back that up.
And this is Eric, just a note. When you look at that chart, right, we had nine months where bookings were essentially flat in kind of in a range, and we've seen a meaningful increase here in both January and February.
Right. Helpful. Thank you.
Thank you. Our next question comes from the line of Tore Svanberg. Proceed with your question.
Yes. Thank you, Steve, for all this detail and great job the first 100 days. My question is on your China business. I was just hoping you could clarify that a little bit more. So I do understand the ChinaCo and the relationship and so on and so forth, but perhaps you could add a little bit more information on how big is that revenue today? And I think you mentioned that initially it's catalog, but eventually it can move beyond catalog. So any more color you can add on that would be really helpful. Thank you.
Certainly. So let's break down our China business. China business is about 18% of our total business. Now, out of that 18%, half of it, which is about 9%, is a business that simply is assembled or manufactured in China by our customers, but it goes out to U.S. or Europe. And those customers really want the Western product. They want our product. They don't want the local Chinese brands to be going into that product for export. Much of that business is already trying to move out of China, going to Vietnam or going to India or going to Malaysia because of the tariff threat. Also, our customers are moving the assembly of that business to somewhere else. So that business just moves with us. The other 9% of the business where it's made in China stays in China.
Half of that business is made up of products which are very complex, advanced Ethernet products, data center products, high-end FPGAs, and other highly analog integrated mixed signal high-voltage products where that's not where the focus of many of the local Chinese brands is. So if you remove the other half, which is 4.5%, which are more complex products which will stay with Microchip, then it leaves the other 4.5%, which is more of a commodity. And those are lower-end microcontrollers, analog parts, commodity parts, catalog kind of parts. So that's where a lot of the local brands are appearing. So that's about 4.5% of our business. And if we don't do anything, that business over time would be vulnerable.
Now, we make very good products, and we have significant market share, but that will gradually erode if we didn't do anything. And this will allow us to participate and not allow that business to erode.
Very helpful. Thank you.
Thank you. Our next question comes from Christopher Danely. Proceed with your question.
Hey, thanks, Dan. I guess just a modeling question, guys. So if we look at the previously announced and the currently announced restructurings, should we just take that December quarter OpEx and are you going to cut out maybe, I don't know, $50 million a quarter out of that and then a little bit out of COGS? Just any sort of clarity on the number impact here to the current business model?
So. there's a lot of the on the OpEx side that comes out very quickly. We're starting communications today on those changes. In some locations, it takes a little bit longer to work through those things just depending on country and statutory requirements, but we really expect that the full impact on the OpEx side, that we'll be realizing that as we get through the June quarter, so over the next two quarters, those should be the end of this quarter, and in the next quarter, those savings should be realized. The Fab savings are obviously a little bit different because we've got to kind of sell through higher cost inventory, and it takes longer for those savings to show up in the P&L, but we talked about that before when we discussed the Fab 2 closure,
so I will be a little careful starting with the December expenses because through middle of December, all our worldwide employees were still on a 10% pay cut, so December didn't model the full expenses. March does. And then in March and the middle of March, we're doing those cuts. So there are a lot of moving parts, and got to be a bit careful. Can we guide them a little better?
Not at this time, right? W e're, like I said, implementing some of these changes starting this week. It takes a while to get those communicated and realize the savings. So I really wouldn't change March at this point in time. There'll be some impact, but you'll start to see a bigger impact in June. And by the end of June, we'll have that full run rate coming through.
Is that run rate about $50 million per quarter in savings? I just want to get the amount correct.
No. We've kind of quantified that in the $90-$100 million range in terms of the annually. Annually. Annually. So it's roughly $22.5-$25 million a quarter.
The changes we're making on the manufacturing side don't show up on the OpEx. They show on the COGS line where capacity is lower.
Got it. Okay. Thanks, Dan.
Thank you. Our next question comes from the line of William Stein with Truist. Please proceed with your question.
Great. Thanks for taking my question. Perhaps I'll touch on the channel inventory comments. Steve, you said that the distributors are near the end of the inventory correction. We've heard this from them for certainly over a year, maybe longer. But if you just look at their financials, i t's pretty easy to see that inventory is still extremely elevated for them. If it's not elevated for you, it suggests you're in a much better position than their average supplier, which is just a little tough to understand given ESP and the changes you're making today, right? So how can you help us reconcile that distinction, please?
Yeah. So let me explain that. So for the last seven quarters or so, every quarter, our selling revenue to distributors has been declining. They have been buying less and less and less because they had inventory of products, and then they wanted to buy less and less. So what I'm seeing is now it's no longer declining. That doesn't mean the inventory has gone to perfect, but the inventory has gone low enough, mostly corrected, that they are now either out of mix or need the new products or need a mix of products where their current customers are. That revenue is no longer declining. But overall, if you see the difference between sell-in and sell-through, the sell-through is still higher than sell-in. So they're continuing to lower their inventory by having sell-through higher than sell-in. But our sell-in to them is no longer declining. Does that make sense?
Yep. That helps. Thank you.
Okay.
Thank you. Our next question comes from the line of Joshua Buchalter with TD Cowen.
Please proceed with your question. Hey, guys. Thank you for the detailed presentation and for taking my question. I wanted to ask about the competitive environment. Obviously, a lot has changed over the last few years, but how are you viewing the competitive environment as you sort of, it sounds like, start to lean a little bit out of the distribution channel? And also, most of your peers now, everybody has parts versus coming out of the shortages. And in particular, you mentioned pricing concessions, your willingness to take pricing concessions on your last earnings call. How has that factored into your long-term model? Thank you.
So I would think competitive environment is really what it has always been. It's not excessive. It's not benign. The competitive environment has gone benign during the shortages because all customers worried about is acquiring any parts that they could get. And now lead times are short. Parts are available. So the competitive environment has gone really what it was pre-COVID. And any kind of price decreases are very normal where you have a price concession on the new design, on the new design, and a new product. But we're not giving price concessions on existing designs where the customer is buying product on a run rate basis for years. Any kind of price concessions are built into our long-term model.
Just maybe to clarify that first part of your either statement or question, we aren't de-emphasizing distribution or not leaning into distribution in any way. We've just made some changes in the compensation structure, but distribution is still a really important part of our business and how we reach the long tail of 120,000 customers that we serve.
Got it. Thank you both
. Thank you. Our next question comes from the line of Harlan Sur with JP Morgan. Please proceed with your question.
Good morning. Thanks for hosting this event. On the lower turns business, that was responsible for the weakness in the December quarter relative to your original guidance. On the March quarter guide, I assume that the Microchip team was a bit conservative on percentage turns business required to hit the midpoint of the revenue range. Can you guys just give us an update? First two months, March versus first two months, December quarter, what's the trend on terms? I assume it follows the overall booking trend, which is up quarter over quarter in dollars. T ypically, when the supply demand starts to kind of normalize, the first thing you would see is a pickup in terms, but wanted to get the team's view.
I think, Harlan, you can overly analyze monthly data and sometimes can lead you to maybe not good erroneous results. December quarter is totally different than March quarter. December quarter has two holidays, Thanksgiving and then Christmas. Christmas is almost in the U.S. and Europe, and Thanksgiving is only in the U.S. So the December quarter can be front-end loaded because of those holidays. March quarter very much depends on is not front-end loaded. People are slow to come back from the holidays to revamp their factories in the current environment.
A lot of our customer factories had a shutdown over the holidays. So they're coming back slowly on the 3rd or 4th or 5th of January, and then the March quarter often depends on where the Chinese New Year will fall, which doesn't fall on a given day like Christmas does. It's on a lunar calendar so it changes a lot based on that, so rather than trying to analyze month versus month behavior where you cannot do that for those two quarters, what we're telling you today is based on where the backlog is and the turns we have taken and the turns we think we will likely take. The current quarter is very much on track. Yeah.
Just on turns in general, when lead times are so short and we don't expect that to change, obviously, we're going to change the slope of our inventory graph, and dollars and days will come down here over the coming quarters. But turns are going to remain in a very important part of our business. With short lead times, our customers are used to that in our business over many years. And as they start to see their inventory drop, they know that we can respond on short notice because we've got inventory and capacity. So turns will remain important. We likely will not be breaking out a specific percentage of turns needed in any particular quarter, though.
Okay. Thank you.
Thank you. Our next question comes from the line of Joe Moore with Morgan Stanley. Please proceed with your question.
Great. Thank you. I wonder what you see when you get to steady state, the balance between internal fabs and foundries, and has that changed at all? And kind of can you give us a sense with these capacity shutdowns? What kind of throughput utilization are you looking for? I know sometimes it'll be higher and sometimes lower, but how are you optimizing that?
So I think when we reach steady state, inside-outside business should still be 60% outside on fabs, right?
Yeah. I think over the last few quarters, it's been about 37% internal, 63% external. I don't expect significant changes in that. In shutting down Fab 2, we will be transferring all the process technologies that aren't already qualified in our other two fabs, and that process is underway.
Those still stay inside, plus we're ramping some products that we're bringing from outside. So it's going to be somewhere in that range between 37% and 40% in the U.S., rest from outside. The other part of your question was the utilization inside. W e are sizing down the factories, so we are fully utilized, right?
Yeah. Yeah. And that doesn't necessarily mean that all of our underutilization charges go away. We got to ramp back up to a certain level. But we're reassessing all that as we go through this process as Fab 2 shuts down and the changes in headcount are made in the other factories to get us right sized, as Steve says.
Great. Thank you.
Thank you. Our next question comes from the line of Christopher Rolland with Susquehanna. Please ask your question.
Hey, guys. Thanks for the question. My question is really around inventory. One of your competitors has moved to a consignment model with their customers. Just given your inventory position, does that make sense? Maybe, Steve, if you could talk about the puts and takes on that kind of model, and on inventory, Eric, could we unreserve that inventory at some point in time, just given the long life of your products? Thanks. Maybe I'll take that one first. So that answer is easy, so once you reserve a product, you don't unreserve it. It stays reserved, and you can get the benefit from that written-off product if and when the product sells through in the future, but you don't reverse those charges. Those are permanent in nature.
On the first question.
Yeah. Previously unreserved.
Sorry. Go ahead, Steve. Maybe the follow-up point on that, and Steve kind of alluded to this in his gross margin discussion in the new model, is we've had huge charges for inventory reserves for the last year plus. Steve mentioned that last quarter, December quarter, was about $82 million of charges related to new inventory reserves. Now, as our inventory on the balance sheet comes down, and hopefully we're in an environment where revenue is growing, backlog is improving, business confidence is improving, those are the first charges that will be reduced and hopefully eventually almost entirely go away. That will be a nice initial pop that will get in gross margin improvement prior to increasing utilization in the factories as we need to start ramping the factories. So hopefully that helps. Go ahead, Steve, on the consignment piece.
Well, the first part of your question was one of our competitors has moved to consignment with the customers. I think you probably mean they have moved to consignment with our distributors. I don't think anybody has moved to consignment with the customers. We have looked at consignment with the distributors in the past, and with our vast distribution network, with lots and lots of regional and catalog distributors around the world, how our product is distributed, it is a very, very complex equation regarding who imports it, who owns it, what are the return rights? And so in our system, it just does not work. It'll take an hour to explain it fully, but we are not moving towards consignment of inventory at our distributors.
Thank you, guys.
Thank you. Our next question comes from Craig Ellis with B. Riley Securities. Please proceed with your question.
Yeah. Thank you for taking the question, guys, and appreciate the expansive breadth and depth of information. I wanted to go back to the table that showed Megatrend revenue and its performance over time versus non-Megatrend revenue. Understand, Steve, how the mix shift in the 2X with potentially Megatrend revenue now greater than non-Megatrend factors into your objective for above-industry growth. Then related to that, and maybe more for Eric, to what extent are Megatrend products' gross margin better than those for non-Megatrend, and to what extent, if that is the case, is that factored into the gross margin target? Thanks, guys.
Well, t he first part of your question is trying to decipher the growth rate. Now, obviously, if the Megatrend keeps going at that 2X rate and Megatrend is more than half of our business, it points towards a positive growth rate trend. So I'll just leave it at that. You're right. The second part of your question was the gross margin for the Megatrends. I don't think we break it out by that, but I would think that the Megatrend products have pretty good gross margin.
They do, and most of our products really do. There's a couple of Megatrends that are pointed out there. T he investor base understands that FPGA has high gross margins, right? So that's a piece of our strategy and the growth factors, and that will factor into that, but all of our products in them have very high gross margins, and that's built into how we get back to 65% over time. If I can, Steve, most of our anchor products that we include in those Megatrends are our highest proprietary products or most proprietary products, and so they typically have higher gross margins.
Yeah. Easier to say that on an anchor product, gross margin is much higher. It's less easier to say that gross margin because you could have a commodity catalog product with a heavy gross margin get designed into a Megatrend application.
Yeah. Got it. Thanks for the additional call, guys. Appreciate it.
Thank you. Our next question comes from the line of Vijay Rakesh with Mizuho Securities. Please proceed with your question.
Hi, Steve. Just to clarify on the CHIPS Act , you said you paused it. Is that coming from Microchip, or are you seeing a pause on kind of the disbursement of that? I had a follow-up.
What's the question?
So the question is on CHIPS Act . The pause is as we wait for the new administration to be in place and working through that. And obviously, we're working through our own kind of restructuring actions at this point in time.
Got it. Then on the cash return to investors, how do you view the dividends and debt paydown and buybacks? How are you prioritizing that, I guess? Thanks.
Currently, based on the cash, the free cash flow that we're producing, we're giving 100% of that, if not more, back to the shareholders. Every alternate quarter, the quarter in which we pay the taxes and pay the interest on the bonds and all that, in fact, the free cash flow is slightly lower than the dividend. Therefore, we actually, in those quarters, we have to borrow the money to pay the dividend. There's no plan to cut the dividend. We'll continue to pay the dividend. We believe that this issue goes away in a couple of quarters. It's not a long-dated issue. That's why we don't want to cut the dividend.
After that, the cash flow will cover the dividend. And when our cash flow is above the dividend payment, we'll use the rest of our cash to pay the loan, the money that we borrowed to cover the dividend in the past few quarters whenever we were short. Once we are back to the original debt level before we started borrowing money to pay the dividend, which is a little bit away, then we'll have another conversation with the street. What do we do with the excess cash?
Got it. Thank you.
Thank you. Our next question comes from the line of David O'Connor with BNP Paribas. Please proceed with your question.
Yeah. Great. Thanks for taking my questions, guys, and for the presentation. Very helpful. Just a couple of clarifications on my side. Just going back to China, firstly, that 4.5% of the business, that's that kind of China for China in China that's going through now, ChinaCo, is the expectation that that disappears over time, given the competition that you have on a localized basis there in China? And the rest, just to be clear, the rest outside of that 4.5%, the rest of it is business as usual. That business is moving out of China to Southeast Asia, talked about, Steve. There is no change to that side of things. That's my first question. And then just going back to that slide on the client relationship, you talked about where the relationship, 6%, I think, was the number where it kind of declined.
I know it's a survey, but is there any way kind of how we should think about that number in terms of kind of revenues, if there's any way to think about that? Is that 6% like a proxy kind of for revenues, how to think about it? And then lastly, one final one, if you could fit it in. Just, Steve, you just mentioned on one question earlier, shutting Fab 2 and internally now you will have high utilization. Can you just talk around kind of how much capacity there is, your ability to respond, to bring back online, and the kind of the long-term CapEx to sales, any kind of change in thinking there? Thanks, guys.
So start with sort of China for China. If we don't do anything, then 4.5% out of the 18% China business we have will be seriously at risk. Now, it will happen very slowly. It may take a decade to get there. If you don't do anything, then the 4.5% business would disappear. And what we're trying to do is compete in that 4.5% business by having a Microchip product, local brand, also available to compete with the other local people. And our feeling is that our customers who are buying today our product, if they could check the box that they bought the local product, then they'd rather buy our product, a product with our die in there, than some local design product that is poor quality or is not a broad product base.
And as they go up and down in their applications, our product gives them a more holistic solution. So that's the answer to your first question. The second one was.
The second question related to client relationships on the chart for the survey and revenue dollars. That's not something that we break out, but I think what's important from that is that it was a very broad survey that was done. There is no company in our industry that didn't disappoint customers in some way during the upcycle because lead time is stretched out so far. This percentage, this 6% that Steve talked about, is a pretty low percentage. It's important. We're addressing that with customers and meeting with them and talking about them about how we can continue to partner with them, but we aren't breaking out the revenue dollars. I don't know if Steve wants to add anything else to that.
I don't know if they're competitors with 6% of the customers that are disappointed. That's a pretty low number. Look at that, where in 24% of the customers, our relationship improved. So maybe our revenue is increasing and we're taking share from the others in 24% of the customers, while at 6% of the customers, we may have some challenges. So it's when you do business with 125,000 customers, it's just a very large number, and you can't get your arms around the exact dollar numbers at each customer in the long tail, especially many of them buy from distribution.
So I think what that says is that we're not in that bad a shape on that issue, and we're taking massive efforts to improve further our relationship where it may have declined, and at the end of the day, I think we're going to probably this is going to become positive. It will actually be a positive issue for us rather than negative.
And then David's last question was on Fab 2 and the actions that we're taking there and in other places in manufacturing, and then our ability to respond, and then what CapEx would be on a go-forward basis. So we don't think we've limited our ability to respond in any way. We built out significant cleanroom space at our other two internal factories during the upcycle. We've got a ton of capital that we can tap into that has not been deployed for manufacturing yet, but it's been paid for. We call that sitting on our projects in process account. And because of that, CapEx for the next several years, in spite of what hopefully is a much better revenue environment, should be very low.
Very helpful. Thanks, guys.
Thank you. Our next question comes from the line of William Stein with Truist. Please proceed with your question.
Great. Thanks for taking my follow-up. In the China for China strategy, Steve, what about the development tools? Are those going to be sort of rebranded or presented as coming from another vendor? How are you managing that, please?
So we already have a version of Chinese-built development tools available to our Chinese customers. That problem was solved years ago. We needed a very low-cost Chinese development tool. So we provided an emulator chip to a company who actually builds a development tool, sells them with Chinese support and all that. So the development tool is already available.
Thank you. There are no further questions at this time. I would now like to turn the floor back over to Steve Sanghi for any closing remarks.
Well, we want to thank everyone for being patient until we get to this March 3rd date. Everybody wanted the answers that we provided today on our earnings call in February, and they just were not available. I wasn't done. So I'm glad to share all this with you. And the next time we'll talk to you will be when we announce our March quarter earnings in early May. So thank you very much.
Thank you. This does conclude today's conference call. We appreciate your participation. You may at this time enjoy the rest of your day.