芯源系統 (MPWR) 2022 Q2 法說會逐字稿

Welcome everyone to the MPS Second Quarter 2022 Earnings Webinar. Please note that this webinar is being recorded and will be archived for 1 year on our Investor Relations page at www.monolithicpower.com. My name is Genevieve Cunningham, and I will be the moderator for this webinar. Joining me today are Michael Hsing, CEO and Founder of MPS; and Bernie Blegen, VP and CFO. In the course of today's webinar, we will be making forward-looking statements and projections that involve risk and uncertainty, which could cause results to differ materially from management's current views and expectations.

Please refer to the safe harbor statement contained in the earnings release published today. Risks, uncertainties and other factors that could cause actual results to differ are identified in the safe harbor statements contained in the Q2 earnings release and in our latest 10-K and 10-Q filings that can be found on our website. MPS assumes no obligation to update the information provided on today's call. We will be discussing gross margin, operating expense, R&D and SG&A expense, operating income, other income, income before income taxes, net income and earnings on both a GAAP and a non-GAAP basis.

These non-GAAP financial measures are not prepared in accordance with GAAP and should not be considered as a substitute for or superior to measures of financial performance prepared in accordance with GAAP.

A table that outlines the reconciliation between the non-GAAP financial measures to GAAP financial measures is included in our Q2 2022 earnings release, which we have furnished to the SEC and is currently available on our website.

Now I'd like to turn the call over to Bernie Blegen.

Thanks, Gen. MPS achieved record second quarter revenue of $461.0 million, 22.1% higher than the first quarter of 2022, and 57.2% higher than the second quarter of 2021. This broad-based year-over-year revenue growth was a result of consistent execution against our strategies.

Turning now to the second quarter 2022 revenue by market. In our enterprise data market, second quarter 2022 revenue of $65.2 million increased 53.4% from the first quarter of 2022, primarily due to an accelerated ramp in our data center and workstation computing sales.

Second quarter 2022 revenue was up 117.9% year-over-year. Enterprise data revenue represented 14.2% of MPS's second quarter 2022 revenue compared with 10.2% in the second quarter of 2021.

Storage and computing revenue of $122.3 million increased 26.6% from the first quarter of 2022. The sequential revenue improvement reflected higher commercial notebook and storage sales.

Second quarter 2022 revenue was up 111.6% year-over-year. Storage and computing revenue represented 26.5% of MPS's second quarter 2022 revenue compared with 19.7% in the second quarter of 2021. Second quarter consumer market revenue of $97.3 million increased 21.7% from the first quarter of 2022. The sequential quarterly revenue improvement was broad-based with particular strength noted in home appliances and gaming.

Second quarter 2022 revenue was up 27.9% year-over-year. Consumer revenue represented 21.1% of MPS's second quarter 2022 revenue compared with 25.9% in the second quarter of 2021.

Second quarter 2022 industrial revenue of $55.9 million increased 15.1% from the first quarter of 2022, reflecting increased sales of products for power source and security applications. Second quarter 2022 revenue was up 28.9% year-over-year. Industrial revenue represented 12.1% of our total second quarter 2022 revenue compared with 14.8% in the second quarter of 2021. Second quarter automobile revenue of $61.0 million increased 11.9% from the first quarter of 2022. Primarily -- due primarily to increased sales of applications for advanced driver system systems, the digital cockpit and lighting products.

Second quarter 2022 revenue was up 25.3% year-over-year. Automotive revenue represented 13.2% of MPS's second quarter 2022 revenue compared with 16.6% in the second quarter of 2021. Second quarter 2022 communications revenue of $59.3 million was up 6.7% from the first quarter of 2022. Most of the sequential revenue increase was related to 5G infrastructure.

Second quarter 2022 revenue was up 58.3% year-over-year. Communications sales represented 12.9% of our total second quarter 2022 revenue compared with 12.8% in the second quarter of 2021.

Moving now to a few comments on gross margin. GAAP gross margin was 58.8%, 90 basis points higher than the first quarter of 2022 and 280 basis points higher than the second quarter of 2021. Our GAAP operating income was $141.9 million compared to $96.1 million reported in the first quarter of 2022 and $60.6 million reported in the second quarter of 2021.

Non-GAAP gross margin for the second quarter of 2022 was 59.0%, up 70 basis points from the gross margin reported for the first quarter of 2022 and 270 basis points higher than the second quarter from a year ago. The quarter-over-quarter and year-over-year increases in both GAAP and non-GAAP gross margin is attributed largely to operational efficiency gains and a more favorable sales mix. Our non-GAAP operating income was $179.4 million compared to $133.6 million reported in the first quarter of 2022.

Let's review our operating expenses. Our GAAP operating expenses were $129.1 million in the second quarter of 2022 compared with $122.7 million in the first quarter of 2022 and $103.6 million in the second quarter of 2021. Our non-GAAP second quarter 2022 operating expenses were $92.7 million, up from the $86.6 million we spent in the first quarter of 2022 and up from the $70.3 million reported in the second quarter of 2021. The differences between non-GAAP operating expenses and GAAP operating expenses for the quarters discussed here are primarily stock compensation expense and income or loss on an unfunded deferred compensation plan.

For the second quarter of 2022, total stock compensation expense, including approximately $1.2 million charged to cost of goods sold was $42.9 million compared with $39.8 million recorded in the first quarter of 2022. Our second quarter 2022 GAAP other income -- other expense was $5.1 million compared with $634,000 in the first quarter of 2022.

Our second quarter 2022 non-GAAP other expense was $7,000 compared with non-GAAP other income of $1.6 million in the first quarter of 2022. The decrease is due to a $2 million increase in charitable contributions, partly offset by the favorable impact of currency exchange rates. The difference in non-GAAP other income and GAAP other income is the income or loss on an unfunded deferred compensation plan.

Switching to the bottom line. Second quarter 2022 GAAP net income was $114.7 million or $2.37 per fully diluted share compared with $79.6 million or $1.65 per share in the first quarter of 2022 and $55.2 million or $1.16 per share in the second quarter of 2021.

Second quarter 2022 non-GAAP net income was $157.0 million, or $3.25 per fully diluted share compared with $118.3 million or $2.45 per fully diluted share in the first quarter of 2022 and $86.5 million or $1.81 on a per share -- per fully diluted share in the second quarter of 2021. Fully diluted shares outstanding at the end of Q2 2022 were 48.3 million.

Now let's look at the balance sheet. Cash, cash equivalents and investments were $814.1 million at the end of the second quarter of 2022 compared with $775.9 million at the end of the first quarter of 2022. For the quarter, MPS generated operating cash flow of approximately $105.2 million compared with Q1 2022 operating cash flow of $107.4 million.

Accounts receivable ended the second quarter of 2022 at $125.5 million, representing 25 days of sales outstanding, which was 4 days lower than the 29 days reported at the end of the first quarter of 2022 and 1 day higher than the 24 days in the second quarter of 2021.

Our internal inventories at the end of the second quarter of 2022 were $359.6 million, up from the $311 million at the end of the first quarter of 2022. Days of inventory of 172 days at the end of the second quarter of 2022 were 6 days lower than at the end of the first quarter of 2022. Historically, we have calculated days of inventory on hand as a function of current order revenue.

We believe comparing current inventory levels with the following quarter's revenue provides a better economic match. On this basis, you can see days of inventory of 162 days at the end of the second quarter of 2022, which were 13 days higher than the 149 days at the end of the first quarter of 2022 and 44 days higher than the 118 days at the end of the second quarter of 2021.

I would like now to turn to our outlook for the third quarter of 2022. We are forecasting Q3 revenue in the range of $480 million to $500 million; GAAP gross margin in the range of 58.4% to 59.0%; non-GAAP gross margin in the range of 58.7% to 59.3%; total stock-based compensation expense in the range of $42.8 million to $44.8 million, including approximately $1.3 million that would be charged to cost of goods sold; GAAP R&D and SG&A expenses between $136.2 million and $140.2 million; non-GAAP R&D and SG&A expenses in the range of $94.7 million to $96.7 million; litigation expense in the range between $2.3 million and $2.7 million; interest and other income in the range from $1.3 million to $1.7 million before foreign exchange gains and losses; fully diluted shares in the range of 47.9 million to 48.9 million shares.

In conclusion, we are continuing to execute on our growth strategies, including expansion and diversification of our R&D centers and manufacturing partnerships in multiple countries.

I will now open the webinar up for questions.

Thank you, Bernie. Analysts. I would now like to begin our Q&A session. (Operator Instructions) Our first question is from Rick Schafer of Oppenheimer.

Congratulations to you guys, another great quarter. This may seem like a silly question given the guidance. But I'm just curious, are you seeing any impact from the delayed launch of Sapphire Rapids? I mean I know your accelerator content is a lot higher than your CPU core power content.

I think 48-volt, and please correct me if I'm wrong, I think our math shows 48-volt track into sort of $100 million this year, so I guess I'm just looking at sort of the puts and takes. I know it's -- this is a pretty much a monster guide. But I'm just a little curious if you were seeing any drag there from that delayed launch?

So far, we see it for the next year or so. All our growth is -- these are -- as you know, these are all growing fields. I mean we don't own the products that were designed in the last few years, like I mean, there for whatever the version of it, okay?

And so if there's a delay, the 1 thing's that we actually care less. And it's out of MPS control. And -- but overall, we have a new -- has a higher power processor and MPS can provide a much higher benefit to those market segments.

So it's off for about 1.5 years and 1 years , and we don't notice it as in this period, and where we gain more market shares. And in a way, we grow from this existing business, we have plenty of it to growth.

And if I could follow up maybe on the supply question. Majority of your wafer supply is still in China, plus you've got to pay your big Chengdu back-end facility. So I guess, how concerned are you if with trying to derisk supply chain as we keep watching headlines with the U.S. government trying to tighten restrictions, et cetera, on equipment and everything in China? So just curious, your thoughts there. And maybe as part of that, if you could talk about where things stand now with TSMC? And give a sense, maybe, of timing and capacity plans there?

Yes. Just like any other companies, okay? Clearly, we're transitioning out from the last 20 years of manufacturing from China to other places. And all these infrastructure had to be built up. And we're just like other companies. We're in a transition, so actually, we started transitioning earlier. In wafer, started from our engineering manpower. So that came in and we transitioned out of 6, 7 years ago -- 6 years ago. And as manufacturers, we started like 2 -- about 3 years ago. And as you know, we always use a trading edge of a DRAM fab, okay? or a digital fab. And 3 or 4 years ago, clearly, okay, these are higher node, okay? 60 nanometers and 40 nanometers fabs. They're all engaged with MPS, okay?

And as we have a reputation, we will fill up all these fab. And now it opens up in careers and I mean Taiwan and - these are the places now, okay, we're only talking about a fab this one now. And in the next few years will be, again will be -- I can't say it's a move out of China soon, okay? We -- and a still bigger capacity, okay? And we have a large market segment in -- 30% of our revenues still from China. So in the next year or a couple of years later, next year, we probably will be very diversified.

And just to add to that, that as the capacity restrictions are becoming less of a concern for the market in general, customers are asking for diversification as in a China plus-1 strategy. So we're working along those lines in companion to expanding our overall capacity.

Yes. They all require -- each region center require their local supplied. That's where we're playing the game. I mean that's our customer request. And so we are fully aware of that. And so we engage with all these fabs. And okay, across South Asia and Europe and Korea and Japan. And so that's how I see it.

Our next question is from Matt Ramsay of Cowen.

Congratulations on the results. So I had 2 questions, and I'll just go ahead and ask both at the same time because I think we've got multiple calls going tonight. The first one, Michael, both in the server business, so the new enterprise data segment and in the PC business, could you give us some indication of how far ahead of the unit sales to your power -- your core CPU power or accelerator power products actually sell in versus the unit shipments that they get reported by the end customers?

And burning the second question, completely unrelated, like $800 million in cash, give or take, if you could kind of walk us through some of the uses of cash there. I'm sure Michael would like to build some inventory, which is kind of always the case. But if you have any new things to say there, that would be really helpful.

Well, first, your first question, honestly, I don't really know. I mean it's difficult. And as you know, we sell -- these are building blocks. And for more or less in a server and data center areas, and these are more generic parts, and they can be used multiple awake. And it's hard to track.

And frankly, we don't really care. And as long as we -- our revenue goes up, right? And for very high powers, and these are power like a 48-volt power, okay? We do have pretty good dominant players in that segment. And so the -- I think a ramp hasn't really just started recently. I mean we'll be -- in the future, there will be a lot more.

But mentioning about -- you mentioned about notebooks, I mean we're mostly in high-performance gaming or mostly in the -- in commercial notebook. And number of assets and versus the CPU -- versus CPUs, it's also hard to match in because we're selling these power devices, they can be through 2 phase, they can be 3 phase, They can be 4 phase. I mean we don't quite note. And also we care less. And so it's difficult to answer it. So all these notebooks are all the high end, the high-end gaming side and gaming notebooks and commercial. And so these are the ones that they have a variety of a use.

And on the issue as far as our cash position, which, let's put it on the table, it's sort of an enviable position to have over $800 million of cash and cash equivalents. And there's probably 3 things that we look at. The first is we've been consistently paying out a routine dividend. This year, it's $3 -- $0.75 per quarter, $3 for the full year. and we're evaluating the sustainability at a even higher level. We generally announced dividend increases in February, comparing with our Q4 operating results and we'd expect to do so again this year.

Another area that we found is very key and strategic to us is building out capacity. And we're looking, as Michael said earlier, for different avenues in order to build out additional capacity. And in some cases, that may require additional investment.

And then finally, as you also added is working capital, making sure that we have adequate inventory on hand in order to sustain our customers' demand profile. And so right now, we're still below our target of 180 to 200 days of inventory. So we'll continue to be investing in inventories as well.

Yes. Might as well, all these are Bernie [said] all the expenses, that they are small compared -- relatively compare the cash that we generate every year. And so what we want to do is that's probably MPS know the best, be consistent. And we give -- increase our revenue yearly. And that's what we have been doing in the past few years, and we'll continue to do so. And also, we don't want -- we will do some acquisitions, but not acquisition for revenues.

Our next question is from William Stein of Truist.

Congrats on the great results and outlook. Something I tend to ask about is traction in the module business because I know that this is something that's helping boost the revenue growth and the margins. I'm hoping you can maybe update us on the traction in that business, please?

Yes. The module is doing well, I mean, but the revenue is still quite small. It's a hundred-some million dollars, but in -- but I know in the next few years will grow double or triple it. And that's what we've seen in the pipeline, in the design win activities.

And so as we said a few years ago, we do e-commerce and okay, we do programmable modules these show true benefits to our customers truly realize it in the end.

Yes. I'd say that particularly as we had this supply/demand imbalance and our customers are also looking for enabling technologies that the decision process for earning a design win where historically had been just on the lowest cost provider. Now things like the programmability, the flexibility, the time to market, the total cost of ownership are taking a larger weighting in the decision process, and we feel like we're very strategically positioned to take advantage of that change in the market.

That's great. And as a follow-up, if I can. I'm wondering if you can talk about your lead times now and how they might have changed in the last, I don't know, a couple of months. And to the degree to which that's been a competitive advantage. My understanding is that you're offering lead times that enable customers to switch away from competitors' products that have lead times that are so long that it suddenly makes sense to switch. Any sort of comments or update on that would help a lot.

Our own lead time really hasn't changed that much. And we still -- we're still in a lot of delinquencies, okay? And that's a good thing, okay? Because as Bernie said earlier, in all the benefit of our product technologies. Okay, finally, our customers realize it, and it became a high demand. And even though and I think it's due to the new design win activities. And they all switch to this type of a new technology or new design methodologies. And now that really benefit MPS. And at the same time, we had to increase our capacities. I mean not only from China and globally, that's what our customer demands.

Yes. And just to top off that answer is that I don't think our lead times were necessarily different from anybody else necessarily in the industry. But we have the advantage of having so many new products coming on the market, greenfield opportunities, that we invested ahead of the curve, and that's where we were able to have product availability when others didn't.

Our next question is from Quinn Bolton of Needham.

Congratulations on the nice results. I guess I would love to ask, obviously, we've seen some in the compute space Intel and Micron offering much more subdued outlooks for the second half. I think there's clearly some inventory purge going on in the channel. And I'm just wondering, obviously, your September guidance is very strong. But as you look at the order book, have you started to see any changes in the notebook or computer and storage and enterprise data center segment that might echo some of the comments we've heard from folks like Intel and Micron in the broader market? And then I've got a follow-up.

We -- for the notebooks, okay. And for our side, still -- demand is still pretty good and still good, like I mean, and I think as the orders and the orders slow down than before. And but we still have delinquency. And for memory side, for memory side, and there's a lot of new format, okay, starting, we're still facing shortages now.

Yes. And if I could just add to that is that we are looking at any areas of our business that might be susceptible to either backlog that is canceled or is pushed out. And keep in mind that cancellations are always a part of all semiconductor companies. It's not just a onetime or a new event. And the fact of the matter is, is that any influence relative to the size of our overall backlog today is very minor. And so that's what's given us very confident outlook for the second...

The rate of a booking and the rate of cancellations and I would say they're very similar in the last few years. I mean last few quarters, yes.

Okay. So no noticeable uptick in cancellations or pushouts, it sounds like is what you're saying?

Yes.

Got it. The second question is just enterprise data showing very strong growth, and I think that's going to be one of the biggest drivers for the business over the next couple of years, given the greenfield opportunities. And just -- can you guys size for us, is the accelerator card and 40-gig volt opportunity larger than the share gain on CPU power? Do you think they're both equally sort of driving the growth? I'm just trying to sort of -- what's the biggest driver, do you think, for the enterprise data set?

I think it's both. And not only the CPU size for the servers, we gain -- and we gain the market share. We stepped into -- in that game a couple of years ago, but a very small percentage, and we start to ramp. But we're still far distance, then they need a bigger suppliers. And that's one of -- from the server side. And the 48-volts, as you know, we talk about it. And we talked about since 2017 or '18. And we said that this is the inevitable and a trend. You had to go move to 48 volt.

And we are now in the forefront of, and we're not replacing anybody, okay? We set up this market trend. And so -- and also in the data center and the rack itself is a big opportunities and MPS are ramping revenue from there, to from the AC power, and these are 380 volts and I guess, 240 volts input convert into 48 volt. And also the battery backup, and we provide the same -- we provide the solutions for battery management and also cooling side. And so MPS is almost one-stop shopping place for data center.

Our next question is from Tore Svanberg of Stifel.

First one for Michael. Michael, you're known for driving a lot of new innovative business models. And as we now start to think about adding capacity to get to $4 billion, given the geopolitical dynamics and so on and so forth. I know you've talked a little bit about this, but are you thinking or working on new business models to try and secure more capacity for your continuous growth?

Well, these are one thing, right? We do -- of course, we have the fab, we have the mobile. And we will have increased more capacities outside of China. We started about a couple of years ago, also more than a couple of years ago. And now we engage with some bigger DRAM fabs, okay? We will fill that up in the next couple of years.

And -- but going to futures, MPS actually require less semiconductors. And because a lot of products are more programmable and we can use for multiple use. And at the same time, we're growing our revenues. We're selling a lot more than semiconductor. We're selling power solutions, more modules. And -- so we kind of move up the food chain as the new requirement comes in, we play in the market, we're not competing with anybody. And we just provide the solutions.

And so those are manufacturer. These are clearly -- they're different. They're building modules. And we sign up partnerships, okay, for assemble these new modules. And it's unprecedented. And a lot of testings, a lot of qualifications, MPS designed the whole system from a ground up. And so these are the -- there is no such a facility out there. So we had to invent it, that work.

Yes. And I think just to echo a point there is, if you look at our revenue for the quarter, we grew 57%. And keep in mind that we've had 1 price increase, and that was in February this year for 5%. And so if you actually look at where the source of revenue growth came from, about half of it is volume related, and the other half is higher ASPs because the higher value products are selling. So that means that we've differentiated our supply chains. So we're not just dependent on silicon-based products. It's making total solutions.

But these are -- the bigger effect hasn't really taken place yet. That would be a couple -- a year or a couple of years down the road, you will see much bigger effect.

And -- so there is a lot of the new -- a lot of work to do ahead of us. And these are -- particularly, these are new types of modules. Nobody else build the kind of things, okay? We're ground up, we developed and we even develop a manufacturing and as well as testing the qualifications part of it.

That's great, great perspective. Michael, as my follow-up, and this is related to what you just mentioned there, Bernie, which is pricing. I know you're obviously growing by adding more value and higher ASP products. But you mentioned that 5% price increases. I think it's well documented that your competitors have raised prices by more than that. So I guess my question is just simply that, are you gaining a lot of share because you didn't engage in aggressive price increases as some of your customers? I mean, your competitors, sorry.

Gaining share, it's -- gaining share is a difficult to count because gaining share, if the equals product is gaining share and going to be in a similar product. And -- so you -- our price betters, our performance better, we're gaining shares. Okay. Now we're talking about completely different things. And those market segments we want to get in there is less price sensitive, but quality, performance a lot more important.

And now okay, we offer something that is a lot more than that. And 10 years ago, we said who do we compete with? And not a company sell controller and power MOSFET, they're separate, MPS is integrated.

And now you're talking about MPS, we -- our product even on the silicon side. And we have microcontroller, we have a memory, we have digital, we have power. And very unique how we -- what we -- who do we display. It's difficult to say it. Now it's getting to use these type of silicon-based technology, we migrate out. We became a provider of total solutions. So how do we -- who do we display, and okay, how we gain market share? It's very difficult to say.

Yes. No, it sounds like you're displacing more and more companies than before. Just one last quick one for you, Bernie. Most companies' DSOs were up this quarter because of the China lockdowns, shipments being later in the quarter. Your DSOs actually went down this quarter. So can you just talk a little bit about how the China lockdown impacted you? I mean, obviously, it didn't impact it the same way, but any other color you can share with us, that would be great.

No, we really didn't have much impact from the China shutdown. Obviously, we were able to record revenue growth that is well above the industry average. And the concern we might have had is if our customer supply chains got impacted. But we continue to hit our delivery schedules. And if there was an impact, I'm sure we felt some of it, but it was very marginal.

Our next question is from Ross Seymore of Deutsche Bank.

First question is really kind of a high-level one. Over the years, you guys have generally outperformed the analog market by maybe 10 or 15 percentage points of growth and gain share, et cetera, et cetera. Even last year wasn't terribly outside of that range. But this year, it seems like that delta probably doubles at least, maybe something more. So I do get investors that are concerned that your increased availability versus the peers allows you to ship and then it can be double shipping in response to the double ordering. And so it's a cyclical phenomenon that's widening that gap. Can you just talk about the reasons you think that gap is sustainable? And then looking forward, do you think that gap will continue to grow despite the fact that you're operating off of a larger base?

Yes. I think what you say partially is correct. I mean the other company couldn't ship, okay, we have the inventories and again, we have a capacity to ship. But this year, particularly, we see things are very different. And a lot of products, especially we ship to these Tier 1 companies from the industrial side from automotive side and in even the data centers. And it didn't intend to have MPS has a big -- has a majority as a bigger supply to where they give us as a backup and to test a lot.

And in the last years, we shipped all these units. Our PPS failures, okay? We're far better than everybody else, and that's one thing. The quality of -- quality is everything. And -- but they took a chance when they have a shortage. And so we proved that, we give us a bigger opportunity. We proved it, these products as better. Okay, as good as better than our -- whatever the parts they design them out.

And in the last year, this year, all the new product or the new segment start to grow out. And as our product, we can change it, we can reconfigure it. And that will continue we continue to grow. And as we see it at, we cannot handle all the projects. And so in the foreseeable features and these products will continue to grow.

And keep in mind that we're still facing large delinquencies ourselves. And so we've had to be very cautious and opportunistic as far as how we allocate our wafer starts in order to meet real customer demand. So I think we've been clear that during the first half of the year, we did a cleanup of double orders and have confirmed, as I said earlier, that the MPS' backlog still remains very healthy.

And then when you look at the inventory in the channel, we're at lows. It's very lean. And we believe we don't have perfect insight, but that the inventory on our customer shelves is likewise very lean because we've only been doing partial shipments there.

So as far as the markets we feel reasonably confident, obviously, notebook or some of the consumer could give us, we're trying to stay very close to that and evaluate its impact. But as Michael said, a lot of our new growth opportunities are in these large Tier 1 opportunities. And it's that secular growth that's really driving it. And that's different from just building up in the channel or on customer shelves.

Yes. I'll come back to your questions. Okay. Now you said that you're growing a large basis, so you grow, okay, it's difficult to grow. That's kind of a built in everybody's mind, okay? In my mind, I don't have a limit. And the limit is within yourself, what to do. And people told me, okay, early on $200 million, it's your barriers, like $500 million of barriers. $200 million at the time, it was a barrier at one time. It happened. And $500 million, it wasn't okay, then people tell me, okay, you're going to grow $1 billion, you need to slowdown. And at the time seriously, that was 2017, 2017 or '18, yes, '17, I actually said it, okay, when we get to $1 billion, we're going to accelerate it. And that's at the time, that's how I see it.

And now, okay, MPS, we're not selling silicon anymore. We're selling a lot more than the silicon and why not accelerating the growth. So it's -- of course, I'm not saying that now okay, a lot of things still depend on our execution. But only thing is the mindset when not dwell on selling semiconductors. And selling semiconductor is limited, but you have a lot of service engineering manpower, we can -- our customers can benefit to it. That's not -- that's unlimited almost.

And supporting Michael's point there, you might remember 6 or 7 quarters ago, we made the statement that by the end of Q2 of '22 that we would have capacity to support a $2 billion revenue run rate. And I think that the key there is the execution and the focus, and that's exactly what we've done.

I guess, hopefully, a quicker follow-up to all of this as you expand beyond the semiconductor side, you get into -- I don't want to say systems, but more solutions in general, some of the stuff you talked about with the entire rack, the AC to DC, the cooling, all the above. What do we look at the gross margin doing in that? That sounds like higher gross margin business. And I know consistency of gross margin expansion is the mantra that you guys have lived by and delivered on.

But it seems like mix would go in a big tail or would be a big tailwind for you going forward from a gross margin perspective. So just talk about what this changing in your revenue mix means, whether it's end market or system solutions versus chips to your gross margin?

Yes. That's kind of things we're going to -- of course, I can all say okay, we can charge as much as our customers available, okay? And that's kind of a half bs, okay? And -- but the reality is I think that we're comfortably stay around this mid- to high 50s and 55% to 60% in the range. I think that's a sweet spot for us. We're not going to print the corners for us to grow to over 70%. And when we get that, we get that, okay. So far, we don't have headwinds. And I think the opportunity drives the model itself.

In the next couple of years, I think that we're going to stay around this as we are now. So okay, maybe move up a slide later. We don't have -- at least we don't have the headwinds. And so after 3 or 4 years we'll see how we change our models.

I would just add that I think that we reported a gross margin of 59.0%. And as Michael said, somewhere in that area is sort of the sweet spot for our model that allows us also to accelerate our rate of revenue growth. So it's something that we'll continue to evaluate. But I think right now, we're very comfortable with this being our model.

We're not chasing the volumes going down. We're not -- we're not -- actually MPS is not good at chasing our volume. We're not doing these manufacturing actually, MPS doesn't manufacture anything. And the high-volume things, that's not MPS forte.

Our next question is from Melissa Fairbanks of Raymond James.

I will echo the congratulations on another great quarter. I just had 2 really quick ones for you. First, could you remind us what your inventory target or ideal levels of inventory would be in order to maybe clear some of the delinquencies?

And then second, on a somewhat related note, what should we be thinking about for CapEx this year, either as a percentage of revenue or absolute investment. And then what's the longer-term requirements you need in order to meet your demand or your growth plans?

Good questions. So as far as inventory, keep in mind that being so much of our positioning is around growth that as sort of a risk management decision. We believe that 180 to 200 days of inventory is what's needed so that we can manage an upside in customer demand, but also if we end up in an unfortunate situation where we have lots of inventory that aren't sellable that we can compensate for that without having any disruption to our customers' production lines.

So I think that it's been difficult to manage delinquencies while increasing inventories in order to support our model. But I think we've done a pretty good job in these really unusual times.

And then as far as the capital requirements, I think we've talked in the past as far as what our spending rate is. It tends to be on a quarterly basis, can be somewhere between $14 million and $18 million per quarter with a lot of that being testing equipment or even if we are purchasing buildings, we purchase our own office space. And the first half of this year was a little bit lighter than our normal run rate. But I think 14 to 18, absent a big building purchase, is probably a good run rate.

Our next question is from Alex Vecchi of William Blair.

Sorry, I was muted. Apologies. I'll start over. Bernie, can you hear me? Michael, can you hear me?

Yes.

Yes.

Apologies about that user error here. I was saying apologies again, if someone has already asked this question, but I wanted to expand a little bit on Ross's question, just in terms of the competitive dynamic in your products being superior to those of the competition as well as more of the solution sale.

How do we think about power management, in particular, and your positioning within the customer as these products become more complicated, you take up more space on the board, I would assume that those conversations are becoming more tightly coupled and that you guys are becoming more important to the customer in terms of conversation?

Yes. Look, I mean we -- and as a matter of fact, the form of MPS is from the beginning, so we don't sell pin-to-pin comparables, okay? We -- in a similar product, okay, we offer the -- actually, we're always a far better product. And much compact, much higher efficiencies. And as an also cost to play without the armor charge on the late. And that is known for MPS, okay, and a onetime it's early period of time, the MPS is like a price (inaudible), kind of companies. And we actually -- we didn't, that means that we left a little lot of dollars on the table.

And of course, we're not competing in that market segment anymore. Now okay, we offer either total solutions. And if you do -- if you mention -- if you're talking about any applications, they need power. And we're talking about electrical car, MPS can build a whole car and use electronics. And you want to build the data centers, MPS provides the entire product for data centers, and we are mostly there. And that's how we sell value.

And we're not competing on this product competes with that product. We have all software behind it. And user interface, okay, software, that changes the game. We are not competing with the product per se anymore.

I think an interesting dynamic that we've been observing is that power management was always an afterthought. It was the last thing after you designed your board and you came up with the least bad solution. Now we are introduced at the very front end of the development of an application. And the reason is, is because our power solutions enable our customers to be able to develop higher-power solutions than they would otherwise. So that's an interesting twist in the relationship where we're being introduced more earlier to the process and able to jointly come up with the development of shared solutions.

As you see -- as you remember, MPS like 4 years ago, we actually built a car. Built a very advanced car and far advanced than any car that you see in the market. And with battery management and same time, all the motor controls, a lot more complex than the existing EV. And so just for the purpose, we can demonstrate we can do it. And now we can do it -- and 4 years later, actually, we can do a lot more on that. And so that's kind of examples.

That's extremely helpful. And then, Bernie, just 1 last quick question. In terms of the guidance, any end market that we should think of? Or how should we think about the end markets in terms of strength versus -- strongest versus weakest? Or any notable things to call out?

Yes. I think that the themes that you're going to be seeing for the next 2 to 3 years are the enterprise data and automotive. Automotive had a relatively slow first half, but that was exactly what we had in expectations. There was no new surprise there. And we believe that in the second half, it looks very healthy as does the data center.

Yes, as we see it. And okay, we don't want to -- okay, what will we provide, not customer asked for it. Yes, we will do that, a customer do that. We should lead to customers. So what you should need. That's the game we are really playing, okay? We're playing ahead of game. And I think that all the power stuff like 48 volts, we said that this is like -- we said that in 2017, okay, this is the futures, and we anticipated that. Electrical cars, we anticipated that. And so now we can, in the next few years, and we'll see the very similar things that we'll see all of these will happen.

(Operator Instructions) As there are no further questions, I would now like to turn the webinar back over to Bernie.

Great. Thanks, Gen. I'd like to thank you all for joining us on the webinar and look forward to talking to you again during the third quarter which will likely be at the end of October. Thank you. Have a nice day.