Microchip Technology Incorporated (MCHP) Earnings Call Transcript & Summary

Welcome to the UBS TMT Conference. My name is Pradeep Ramani, and I cover Microchip at UBS. We are delighted today to have with us Ganesh Moorthy, the President and CEO of Microchip; and Eric Bjornholt, Senior Vice President and CFO. Welcome, Ganesh, and Eric, and maybe I'll hand it over to Ganesh for some introductory remarks.

Right. Thank you, Pradeep. During the course of this discussion, 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 the actual results or events may differ materially. We refer you to our recent filings with the SEC that identify important risk factors that may impact Microchip's business and results of operations. So with that, a quick summary of where we are from a business standpoint. So business remains strong with the demand increase continuing to outpace the supply increase. We expect the demand-supply imbalance to persist through much of 2022, possibly into 2023, and we expect to be in a position to support revenue growth for at least the next 4 quarters. With strong debt paydown, we've dropped our net leverage ratio at the end of the September quarter to 2.99% so just under 3%. With that and a strong business outlook, both Moody's and Fitch upgraded our debt to investment grade, and this has enabled us to commence our [ buy and ] stock buyback activity, adding to our capital return strategy of strong dividend growth. Our Microchip 2.0 strategy over the last decade employed serial acquisitions to give us a solid foundation to build scale and breadth of products and solutions and significantly improve our gross and operating margin model. We have transitioned to the next phase of Microchip strategy, which we call Microchip 3.0, and the essential elements of Microchip 3.0 are organic growth at 2x the industry growth rate, focusing on total system solutions and the key megatrends, operating margin target of 44% to 46% and free cash flow target of 38%. An increase in the capital return to shareholders to starting at 50% of free cash flow this quarter and then to further increase it to 100% our free cash flow as the net leverage drops to 1.5x. And so it will be a continuous every quarter increase in the capital return to shareholders. We want to increase the CapEx investment to 3% to 6% of revenue as we see opportunities to deploy that to improve both the financial performance but also our customer support we can have and to invest 130 to 150 days of inventory over the business cycles to provide that capability. And finally, to maintain a strong company foundation that is built on culture and sustainability. With that, thanks, Pradeep, and we are ready for your questions and any audience questions.

Great. I want to start off with essentially what is Microchip 3.0 and the growth rate where you're laying out the target for growing at 2x the industry over the longer term. When I go and look back at historical, you've grown both via acquisitions, and there's been a component of organic growth. But if I look at it on a like-for-like basis comparison, how would you say your 2x industry growth target compares to what you've done over the last 3 to 5 years, so to speak?

So the last 3 to 5 years includes a period where we were doing acquisitions. The last acquisition was about 3.5 years ago. So the rate of growth with acquisitions is obviously higher than just the organic only. But I think the target to grow at 2x of industry is built on 2 primary strategies. The first is for all designs we pursue in all markets a total system solutions approach, taking advantage of the broad portfolio that we have built in the last 10-plus years, maximizing the content we get designed in into each design that we pursue. And then the second is a -- to exploit the 6 mega trends that we have talked about. And we have -- we want to basically ensure that we are well represented with our total system solutions in these 6 mega trends so that we can grow at 2x the rate, we expect Microchip to grow at. If you look at a longer-term view of Microchip over the last 15, 20, 25 years, we've been gaining market share every year just about in the last 30 years to micro-controllers. It's a Marathon, we are every year are taking a little bit of share, a little bit of share more. We've done the same thing on Analog. We started about 10 years later about the last 20 years. We've used different techniques and tools over the years on how to achieve that. But it is very much in our DNA to find ways to gain market share of different techniques at different times. Today, it's TSS and mega trends.

Yes. And I want to sort of touch on that a little bit more because, I mean, do you -- when you talk about TSS and mega trends, how do you, sort of, look at it from a competitor standpoint? Just -- do you believe that you are maybe better positioned in terms of your leverage to these mega trends than competitors or maybe your product portfolio is better or maybe you're ahead of the curve in investing on supply? Can you maybe sort of parse that out a little bit more as to the mega trends envelop a lot of companies and why you can grow faster than the other companies who are good competitors?

Sure. So we focus on the areas of the mega trends that play to our strengths. And depending on which one it is, there's different parts of our product lines or businesses that we target from there. Not everybody has the breadth of capability that we have. We have a tremendous amount of what I call anchor products in any design. And these anchors are different when you go to a data center, if you go into automotive, if you go into an Internet of Things. And the anchor typically is the brains in the system, and that's micro-controllers, microprocessors, or FPGAs, et cetera, et cetera. And so in these mega trends, we have picked them where we have strength in our capability, where we have an approach to what we can do to get to market. So a TSS approach is what we do with for mega trends as well. We have built domain knowledge in these different areas for what we're trying to do. And we're not trying to get every single application in a given mega trend. There's a space as we play in, that is what we're targeting. And we, kind of, shared some of that in the slides that we went into in the Investor Day about a month ago, which shows by mega trend, where are we targeted with which of our products with what growth expectations we have.

Yes. And following up on that, the TSS presentation on your Analyst Day, well, you had a lot of very interesting examples of how you're positioned. But the one burning question I was left with was when I synthesize all of that and just -- if I simply look at your December quarter guidance, I mean, you were at roughly $7 billion of revenue. How do I think about where you are in terms of TSS today as a percent of revenue? And maybe is there a better framework to think about how investors get visibility to how TSS is progressing as we move forward?

[ Some of this ] is hard to do because of how horizontal in -- many of our markets are. We sell 120,000 customers worldwide. These are designs that have very long lifetime. So in many cases, our designs are in production for 10, 15 years, et cetera. And what we did for TSS at that time may be different because we didn't have all the solutions. So it's hard to give you a quantitative framework for TSS-driven growth. You can see it in our data, you can see our growth rates are, in fact, progressing faster. And you can see that many of the qualitative data points we provide. We have shown examples of how we have progressed. Today with all of the strengths we have, what kind of TSS we're able to achieve in design wins. We've shown how the mega trends intersect with our TSS approach and how the mega trends then on top of TSS add another growth -- 2x of our normal growth rate or our expected growth that we would have. So it's the best way to look at it is on a more qualitative basis. I can tell you this, every single discussion inside the company, every review we have, every customer interaction we have. Basically, TSS is very much built into the DNA of Microchip, and all of the people, processes and systems are all driving there. We don't even think about it as something different from what we do. It is what we do.

I see. And if I ask maybe a little bit more of a nearer-term question that stitches together with TSS angle, you provided a very strong outlook in terms of revenue growth for at least the next 4 quarters. And the pricing environment is good today. Can you parse out one for us, the pricing versus volume assumptions in your near-term outlook, given that we are in a pretty heavily supply constrained environment. And maybe speak about how TSS helps you augment your pricing by bringing more value to the customer as well. So...

Sure. So firstly, what we have said is we expect to be in a position to grow revenue for the next 4 quarters. The confidence in being able to say that comes from the fact that our limitations over the next, what we call, short term, the next year is really all supply driven. We have [ nothing ] more backlog, noncancelable backlog to be able to drive the growth. And so the growth consecutively over 4 quarters, is a function of how we are bringing capacity on. It is not dependent on price. There could be price adjustments for cost increases that we pass on. But really, the growth is largely driven by us bringing on more capacity. The TSS, in this time frame, what it does is, it does what we would expect in any environment as we're winning designs is. It provides a larger dollar value of any application that we're able to design, that we win a design in. And so there's a multiplier every time we win one of our core products, micro-controllers, the anchor products in a design controllers, processors, FPGA. We carry along our power, our timing, our Analog, our mixed signal products, et cetera, that come with it.

Okay. And if I say -- if I try and look at other -- some of the other growth trends you talked about. I mean, one of the new announcements I was looking at from Microchip was around your silicon carbide traction. And even on the Analyst Day, I think you said your SAM would be roughly $1.4 billion by 2025. So I just want to understand the growth strategy behind silicon carbide. And where we are starting from today in terms of your revenue contribution? Obviously, it's probably very small. But what is your expectation for your share of silicon carbide in 2025 of that $1.4 billion? And does it skew towards autos or industrial? Or is it balanced? And how do we think about the strategy in general? .

Sure. So first of all, we're very happy with how the strength of our silicon carbide portfolio is, in fact, you will see there was a press release this morning with one of our partners using all of that in the reference designs that they're going to build. There's our reference design into their products as another example. And it's not just the MOSFETs and diodes, it's also got the drivers that are required to go with it. We have a differentiated performance with this portfolio, building upon robustness requirements and high-voltage environments and the traction we see in these opportunities is in the places where we see good gross margin characteristics as well. At this stage, we're not breaking out our silicon carbide revenue, but it is focused on industrial, automotive, aerospace and defense. Those are the markets we see with the most potential for -- we do absolutely expect that silicon carbide will meaningfully contribute to our growth strategy. And that's built into how we see over time. There's lots of upside options. But we also want to make sure that we are driving growth that has the margin characteristics that we expect out of our product lines. And so there are parts of the market that are not with the margin characteristics we like and those are not the ones we're focusing on, and are parts of the market that have the gross margin characteristics we like, recognize the value that we bring and that is where we're focused.

And would you -- when I superimpose that against supply constraints, would you look to do more of the silicon carbide manufacturing in-house or -- to drive the growth? Or is that part of your CapEx plan? Or is it mostly going to be manufactured outside?

It will be a combination of both. We obviously have the capability to do it. We don't do it today because of the size. We will, in time, as the business grows, be looking at that very carefully to take advantage of our manufacturing infrastructure to build silicon carbide.

Okay. And maybe this is a little bit more of a question for Eric. But in your longer-term model, you're guiding to 68% gross margins. I mean that's roughly 300 bps above where you are today and puts you in a very, very elite category for -- in semis with respect to gross margins. Now sure, I understand there's going to be a bridge from where you are today, and there's going to be some in-sourcing of front-end assembly and test. But can you help us parse out that bridge a little bit more with respect to the other contributors, maybe on a relative basis with respect to, say, mix shift, pricing, manufacturing, in-sourcing and maybe even just a change in your product portfolio and growth rates?

Sure. So actually, the midpoint of our guidance on non-GAAP gross margin in the current quarter is 66%. So from there, we've got about 200 basis points of improvement to make. And I would really describe it as there's a lot of blocking and tackling that will help us to get there, right? And we've laid out some of the things that you've mentioned in terms of in-sourcing some more of our manufacturing in assembly and test, we look for a 2-year or less payback on capital equipment from a cash flow perspective. We appreciate equipment over 7 years and generally use it longer than that. And so those are gross margin accretive transactions and also give us more control over our supply chain, which is important, particularly in times like this in times of constraint. On the wafer fab side, we are looking to increase the percentage we do internally. I'd say it's more modest than what we're doing in assembly and test. There's probably not the investment that we'd like to see and what we would refer to trailing edge technology. And because of that, Ganesh recently mentioned that we've taken a license to a process technology from one of our foundry partners and are bringing that inside in our Oregon factory, but that doesn't really come online until 2023. And then it's just we're continuing to introduce new and innovative products and so product mix over time can help. I wouldn't call out pricing specifically other than to say, times right now are not normal from a pricing and input cost perspective, but we've been very successful over the last few years, kind of, pre-COVID in keeping ASPs flat year-over-year with customers and selling on value and that process has gone quite well from a customer perspective. So lots of various factors that go into gross margin, but I think those are the main things to highlight.

Okay. And if I asked maybe a little bit more of a zoomed-out question here. You get a significant portion of your revenues from China. And there is a view amongst some investors that are talking to us that MCU pricing, and to some extent, margins could see downward pressure given competition in China. What are your thoughts on where China is with respect to its MCU market and your competitors? And how it really impacts both your growth and margin targets?

Yes. No, no. Good question. And when you look at our China revenue, you have to look at -- there's two parts to it. There's a part of it, which is designs done elsewhere and then manufacturing taking place in China. And there's another part of it, which is designs done in China for manufacturing in China. We report all of it as China because that's where our sell-to point is. So a lot of the design influence of what they use for a big portion of the business in China is actually done outside of China. That said, we have successfully competed with Chinese MCU competitor for over 25 years. This is not new, right? They didn't just come out in the last year or 2. They've been there in different sizes and shapes for a long, long time. And we have duked it out, and we've had to win on the merits of what we bring, what the total solutions are in our software and our tools and our products and the cost of ownership, et cetera. And the results at the end of the day, over that time have all shown in market share. And that includes people who are from China, from other places, and we're more than holding our own with that. So the growth strategy in China takes advantage of the same two trends of -- same two strategies, which is the mega trends and total system solutions. So no, we're present in a big way in China. We're on the ground with designs, working hand-in-hand with customers. We're not just shipping product. There's somebody who designs, and we don't have much influence on. So we're happy with where we're at. We don't see any major issues with competition in China.

Okay. And on the R&D front, I guess, a little bit more in terms of how you're thinking about R&D. I mean the MCU universe has a broad range of competitors. But if you look at it a little bit more holistically. Some of them have a higher R&D profile. And yet, they are levered to similar growth markets. How do you think about from an R&D perspective, the allocation between, say, your mega trends and the broader markets. Is your R&D leverageable between these end markets and trends?

They absolutely are. And there's a part of the R&D that we do, which is baseline capabilities that feed into multiple businesses, multiple markets and multiple market trends in and of themself. Over the years, we've been fortunate, we've been very efficient with our R&D investments. And I think a lot of that comes because we have a highly collaborative culture that enables us to share IP, right? We don't try to reinvent the IP 4x in 4 different businesses and not only the IP but even the methods and expertise that we build. So a lot of work is used and reused inside of the company across businesses, across regions of the world just because of the systems that we have set out. We also find that operating our R&D at the end of the day, what I would call a [ tight ] R&D budget, finds the best highest ROI projects. Those are the ones that rise to the top and get funded. And very little R&D gets wasted or get -- results in terminated projects as it happens quite a bit in other parts of the industry. So we think both how well the R&D we spend is used and reused inside of the company, but also how efficient we are with the R&D that we spend are important in where we are in our spending for R&D to create innovation.

And dovetailing into that question, I mean if you look at what some of your competitors are doing, there is a sense that 32-bit MCUs will ultimately cannibalize the 16-bit MCUs as pricing converges and many competitors have actually stopped investing in 8-bit MCUs even as you remain pretty committed to doing that. What are your thoughts in terms of growth prospects for 8, 16, 32-bit MCUs? And do you buy the cannibalization point of view? .

We think of the microcontroller market as a continuum of needs and solutions. This 8, 16, 32, yes, there are technical ways in which to go look at it. But I think if you only have 32-bit micro-controllers, you assume that, " Hey, 8 and 16 can all be done with 32 bits. If you only have 8, you assume you can do everything". What we look at is we have a range of solutions, 8, 16s and 32s. We invest in innovation. The market needs all the different kinds of options that we have. And we let the best architecture win from within the portfolio that we have. And with that, we're setting records in each of the 8, 16 and 32-bit micro-controllers as we are in the current environment. And so while there may be some amount of ongoing cannibalization on the edges of what takes place, there is also a significant new component where systems and applications that never used micro-controllers begin to put intelligence in the systems. And often, that ends up being an 8-bit microcontroller or a 16-bit microcontroller because you're starting with something simple. And so it is not just what do you lose on one end of the spectrum to maybe a 32-bit or something else, it is as much what do you gain at the other end of the spectrum from a lot of new applications that are coming in that never used a microcontroller in the past. Take a great contemporary example. You got all these COVID testers that are being out there. The quick testers and all that. Many, many, many of them use low-end and mid-range kind of micro-controllers to be able to get started with what they're trying to go do. And like that, whether it's light bulbs, whether it's faucets, whether it's pregnancy testers, a whole bunch of things that never had any kind of intelligence in them get started on that journey using 8-bits, often 16-bit and then sometimes 32-bits. So we're strong in all three. They all help in our growth model, but we're agnostic to trying to say push people on one direction or not.

Okay. And maybe I'll shift gears a little bit to talk about your -- the bigger picture stuff on the cycle and where we stand and so on. So maybe we can start with your PSP program. You said autos, if you look at the continuum of end markets or auto is probably where you're seeing the highest level of participation in PSP. Is there a geographical distribution in terms of PSP participation? And do you worry about the enforceability of PSP at least across -- throughout the world where the standards are different? And do you think about just PSP, I mean, during the course of the cycle. Sure, customers want PSP today, but what are your thoughts for how PSP works as we move through the cycle?

Sure. So PSP participation is quite evenly distributor across the geographies, like the business problems that PSP solves they're same in all geographies. There's been a matter of are there markets and customers who see the value in doing that. And so largely, what we have seen is within a small percentage plus or minus PSP participation is about the same level in all the geographies that we operate in. The end markets, to some extent, have a say, so you mentioned automotive. But I would say it's not only automotive, it's industrial, it's some of the infrastructure people, et cetera, who all see more durable demand trends. And so that plays a role in this as well. And finally, I think the ratio of the end product that the OEM is making to the semiconductor content also drives a big role into participation. So when you make a $50,000 car with a $400 of semiconductor content, it's a no-brainer to be able to protect the $50,000 revenue with what you're doing in PSP and the $400 of semiconductor with some subset of that $400 and where they're at. As far as enforceability of the PSP, we're not worried at a global level. We have run smaller scale NCNR noncancelable, nonreturnable programs in the past in specific geographies, specific customers and all that and managed them all successfully. This has a larger scale to it, but we're not worried about our ability to enforce what we need to as part of the customers' responsibilities in PSP just as we have our responsibilities that are in. And then as far as what happens with PSP in the future, it's a little early to tell. It was designed as a solution for the problems of what we had a year ago, beginning of this year time frame. And it will probably evolve. It has evolved even in the time that we have been running the program, and it will probably evolve more as time goes on. If it serves the purpose to solve a business problem, then it will continue to be there. if it outlasts its use, then we'll probably have something else at that point, that is more relevant in terms of the business problems our customers are trying to solve.

I'd like to add one more thing on PSP that I think is an important benefit. Obviously, this is a program that was introduced based on what customers were asking for and customers have been very responsive and like the program. But it helps our manufacturing operations teams tremendously, right? If we had a 60-day or 90-day cancellation window, we'd have so much backlog that was just piled up right outside of that cancellation window that customers would kind of continuously push to secure their spot on the line. But with 12 months of noncancelable non reschedulable backlog in PSP, we can make thoughtful decisions on adding people and raw materials and capital and really plan the business so much better than we could otherwise. So there's benefits for both Microchip and the customer, and we're reaping the benefits of that today.

And we think backlog from customers is more responsible when they know that there are consequences, right? If you have no consequences, then there's no problem, just pile on the backlog and see what happens and cancel everything you don't need. This requires them to be as thoughtful about what backlog to place because it is noncancelable, there are financial consequences for.

Yes. And that dovetails nicely to my next question. That, again, from a cycle standpoint, many would argue that this is the time for double ordering and distributor inventories remain low. But -- and I get the argument that PSP helps you, and it gives you some visibility into what customers are really thinking. But how much visibility would you say you have with OEMs and especially the large OEMs in automotive? And I understand the inventory -- the industrial piece is a little bit more fragmented. But how much visibility do you have on the industrial side as well with respect to OEM inventories? And maybe, can you give us some color on whether you see them actually building inventories or being in a position to build inventories anytime soon, especially in industrial?

So the fact is we don't have visibility into OEM inventories because they don't have to report it. And unlike distribution partners who report it to us on a weekly basis where we are able to see it. We do spend time looking at inventory days on the balance sheet of some of the key OEMs who are public companies. So that gives us one view of how inventory is trending. We do hear anecdotally about customers who sometimes have challenges to complete their bill of materials. But also, we have the largest ever imbalance between demand and supply this quarter. And we're going to exit this quarter with a humongous shortage to what customers want. And it's what they really need because the level of escalations we see, the number of customers and [indiscernible] we're trying to help our has not abated, has continued to be high out there, and we would not be in the middle of those conversations if they had inventory and they were just trying to add on to their inventory there. So those would be some of the factors that as we look at as to where we are in terms of inventory building or where it might be building.

And when you look at the lead times, I mean, there's two sides of the client for a very large lead times. I mean, have lead times improved over the last month or so for 32-bit MCUs? And maybe a little bit more of a longer-term question here as a follow-up is, are customers contemplating a redesign of [ sockets ]? And how are you thinking about that?

So firstly, there is no change for the better in lead times within the last month or within the last quarter or any time, right? So lead times remain extended and constraints remain, as I mentioned, for all of 2022. In terms of customers contemplating change, some are. Our first obligation is to make sure we protect existing customers, but we are seeing many who are fleeing from competitors who often sacrifice small- and medium-sized customers. That's just the way the business model works in constrained times. Where we can help them, we are trying to help them and to win them over to Microchip. But at the same time, there's many that we cannot help because we are constrained. And within that, we want to make sure our existing customers are well supported.

Yes. And I was curious, maybe along the same lines on the automotive side, I mean, we're hearing auto OEMs announcing their intention to design chips, right, and get into silicon of late. And just given the MCU, it's probably one of the most important parts in terms of silicon and heavily impacted by constraints. Are you concerned that maybe MCUs are at the top of the list in terms of auto OEMs trying to get into some of this diversification of supply?

I have no concerns whatsoever. Auto OEMs may plan to build a few of the large processors by themselves. The skills, the experience, the DNA required to build these are not at all easy to do. Some of them have vestiges of semiconductors they may have done 20, 30 years ago. The ROI to build a custom product just for yourself is extremely difficult to justify. So I think that I would say we are no more worried about automakers building semiconductors to compete with us, than they should be worried about us building cars to compete with them.

Okay. All right. And maybe the last question on probably the cycle piece. But the industry is in a wave of adding supply. And you have pretty much each player approaching it in what they think is a rational way. And yet, sometimes or at least in the past, it doesn't work out to be rational in hindsight. So I mean, if I look at your strategy in terms of how you are planning to add supply, and this is more on the front-end side, not so much on the back-end side. But in terms of front-end manufacturing capacity, what safeguards are you taking to assure yourselves that supply addition is rational, not just from your perspective, but also from the industry perspective and even foundry partners?

Sure. So when you look at the front-end for Microchip, vast majority of what we build requires what I call trailing edge technologies. Trailing edge on 8-inch is typically 130-nanometer or larger in nodes. Trailing edge at 12-inch is somewhere between 40-nanometer and 130-nanometer in that range. So those are 2 buckets of trailing edge technology. On the 8-inch side, there is no foundry investing in 8-inch. And we are investing at a very measured rate as we go along. There is no overbuild of 8-inch that is going to happen because there are 10 players all trying to build capacity and at some point, it's going to be overcapacity. On the 12-inch trailing edge, we don't build it ourselves. We work with our partners to be able to get it done. But it is significantly under-invested today. Most of the investment that is going from foundries, from IDMs and others, is all going at the bleeding edge of technology. It's going in at 16, 10, 7, 5, 3 at those nodes. Those are not the nodes where we have any substantial volume of revenue in. And so I can't quite speak to how much that investment is and what it might have in term of overhang. But I do know that in 8-inch, there's no investment that anybody else is making. We're making it and we're making it in measured terms. In 12-inch, our partners are making it, but they are making far less than what the industry needs and what we need. And so I think the likelihood of all this resulting in a significant overinvestment or overcapacity, we don't think is there. If you read just the top headline of how much capital is being spent, it might appear that way. But 95% of that capital is going to the bleeding end, which is not where we are.

Okay. and maybe shifting gears to capital allocation because that strategy has -- is going to change a lot. So in the past, I mean, when I look through what you've done in the past, you have been able to modulate the amount of share buybacks. Especially after special situations, for example, like the Atmel deal, depending upon your own internal view on where the stock trades versus your view of the valuation. I mean -- so when I look at your Analyst Day Presentations, and the sense I got was it was more a programmatic nature of the buyback. And my question really is that is there a shift in the mindset from being a little bit more valuation conscious buyback programmatic? Or is there room for a more opportunistic buyback as well? And maybe does this strategy evolve over time, can you paint the picture around that?

Sure. So I'll kick this off, and then I'll let Eric speak to some of this as well. So we are early days of a programmatic stock buyback. We outlined that at our Analyst Day, but essentially a meaningful part of the free cash flow that we dedicate to capital return will go into buyback, taking the combined dividend plus buyback this quarter to about 50% of our free cash flow and then growing every quarter until we get to 100% of which the dividend will be the fixed part and the stock will be the remaining part where we're at. Our buyback is programmatic as we get this thing kicked off. We want to establish a pattern for what we're doing. But also because we don't know -- well, we don't necessarily think corporations are all -- are good at timing the market. So we're going to buy every quarter. We think, as we speak, our valuations are below where they should be. So we are a buyer and have been since we got our investment grade. But that said, we do have financial firepower in our credit line, which will enable us to also opportunistically buy shares if the market temporarily loses connection to what we believe our long-term valuation should be. So programmatic will be the standard approach we take at least for some time. Opportunistic would be, we're ready to do it if we need to do it, but not something we're thinking of at the moment. And there is another component, which many people don't recognize, and I'm going to let Eric speak to it, which is we have synthetically been buying back shares through what we have done to buy back our converts. And so let me hand it to Eric and speak about that.

Yes. So I'm going to sum that up nicely. And just I think the opportunity for investors here is, when you look at our capital return today at 50% of free cash flow rapidly increasing over time each quarter and moving towards 100% as we get to 1.5x leverage. That's something that is in the foreseeable future. So I think that paints a good picture there. We were heavily levered to convertible debt back in time. I think just back in early 2020, we had $4.5 billion of convertible debt outstanding. And as the company has performed well, that has got very expensive in terms of the dilution that has come from that. And so we have reduced our exposure there significantly. And absolutely, that has acted as what I'd call a synthetic stock buyback. And as our stock continues to appreciate, that benefit is going to be with us forever. We laid out a slide in our Investor Day that I think highlights that very well. So we've been focused on capital restructuring of the balance sheet. We are extremely excited to now have that investment-grade rating and start this next journey in Microchip 3.0 with capital returns.

And that's an interesting segue maybe into my next question is [ costs ]. Even though an MCU is inherently a digital component, you have the analog piece and you are compared to analog peers by investors. And so when you look at -- I mean, there is a valuation discrepancy between, say, the high-performance analog guys, ADI versus Microchip. And I get it that some of it is because of the leverage, and that's coming down. But do you feel, even from a competitive standpoint, it is much easier, not much easier, but it is easier for an MCU company to do analog versus an analog company to execute on MCU from a competitive standpoint? And maybe that's something -- that angle is something that's not well appreciated or understood. So maybe reflect upon that for us.

So firstly, I want to correct the perception that microcontroller are not digital product, right? 99% of our microcontrollers are mixed signal product with a small amount of digital. The analog actually is a much larger portion of these products, which is what gives us the opportunity to make these highly differentiated, very sticky, high gross margin products and where they're at. And they are like the analog product lines. They're highly fragmented, right? The vast majority of our microcontroller products if you go in any given mass, it's well under a percentage point of revenue -- percent of revenue that comes from any one mass. So it's highly fragmented, many different permutations and combinations of where we go. As far as -- which is easier if you're a microcontroller company to go to analog or an analog company to go to microcontrollers. I think the results speak for themselves. I think we have shown that as a microcontroller company, integrating analog on our chips as well as providing analog around our chips has been a very successful strategy. And we have the dual engines of both microcontroller growth and analog growth for a number of years that we have demonstrated. And so we do believe that there is an advantage to it. There is also an advantage when we go to market because every system design starts with figuring out what are the brains of that system going to be? What are the anchor products in that system going to be? And those anchor products tend to be the ones that are aligned with our microcontrollers, our microprocessors, our SoCs, our FPGAs because they define how does the rest of the system need to look and they give us substantial insight as to how to bring a more complete solution to the customer, how to get reference designs done? And what analog integration onto the next generation of product would make sense. So I think there are many advantages in a microcontroller processor oriented company [ and then ] take a more encompassing and better control solutions role on what we do.

Great. I think we're out of time. So I would like to thank both, Ganesh and Eric, for spending time with us. And I hope the discussion was of interest to a very large number of people because I certainly enjoyed our conversation.

Thank you, Pradeep.

Thank you.

Thank you.