Good morning. Welcome back. Really happy to have the management team from Texas Instruments join us today. Rafael Lizardi, Chief Financial Officer, and Dave Pahl, Head of Investor Relations. I'm Vivek Arya from the BofA semiconductor team. Typical fireside session. I'll have my questions, but please feel free to raise your hand if you would like to bring up anything. Really delighted to see you, Rafael and Dave.
Thank you.
Thank you for joining us.
Good morning. Happy to be here.
Maybe let's start, Rafael, with the state of the union. You know how you're seeing the demand environment shape up. A lot of macro cross-currents, obviously. How are you seeing the demand environment shape up right now versus what you thought at the start of the year?
Yeah. No, as we said at the last earnings call, we're starting to see a broad recovery in the space. This is what appears to be the beginning of a semiconductor cyclical upturn. It is coming in at a time when we are very well prepared. Bless you. Very well prepared for it. We've been investing in inventory, but more importantly, on CapEx over the last four or so years. We have RF-2. We have SM1, SM2. We have L5.1 and L5.2 at different levels of completion. We are ready for what the market throws at us.
Okay. When you came up with the CapEx plan, Rafael, a few years ago, at that time, the industry looked very different than what it is today. What would you have done differently if you had a better sense of how this market was going to go through? Or maybe a better question to ask is, what would you like to do differently now, given where the industry is?
No, absolutely. It is what we are doing, which is the capacity expansion. To do that well ahead of demand and to do that in the phases that we are doing it. Phase one, we have described our plan as having three phases. The first phase is transferring products that run externally, internally, and growing our base. The second phase is to build buildings that will be required for growth and to qualify those new factories. Once we get to the third phase, it is all about incremental capacity. The first two phases require a significant amount of CapEx. Once we get to the third phase, it is just incremental CapEx to match that capacity to expected demand.
I see. Now, you know a skeptic would say the industry, right, not just TI specifically, if the industry was not very good at spotting when it was overshipping, why should we believe the industry now when it says it's been undershipping? What checks do you have in the system to make sure that what you are seeing as a recovery is not just pull-ins and some temporary because of the macroeconomic issues, that this is true recovery?
Yeah. No, in the big picture, we look at macro trends. And by I say macro trends, I mean semiconductor macro trends. Go to slide 19 of our capital management presentation, and you'll see what I'm talking about. It's over the long term, this industry is very cyclical. It goes up, and it goes down. It's pretty clear that now it hit bottom a couple of quarters ago, maybe last quarter. Now it's on a recovery. That's one. Frankly, more importantly, we're just prepared for whatever the market throws at us. If it happens to be that this recovery is going to delay some time, then it's okay. The inventory that we build is very safe. It's very long-lasting. We hardly ever scrap parts. What we do is very little. It's a percent of our revenue.
Our business model is such that it protects the downside in those situations.
Okay. You know, if you look at that chart, what in your capital management presentation and what I've seen across with the SIA and other data, it has shown that in the last three cycles, semiconductor sales have kind of gone up for like 10 quarters, then they come down for four quarters. It is like two- and- a- half years up, one year down kind of situation. If we are kind of starting on this up cycle, do you think we are kind of at the start of what is normally a two- and- a- half year up cycle? I know there is no cycle that will be the same. Where would you describe where we are, in very early stages, middle stages, or how would you describe where we are?
Yeah. Of course, to your point, you kind of alluded to it, but it's very difficult to draw conclusions from three or four data points, which is all we have in the last decade in terms of how many cycles we've had. It's really not enough for a distribution. Our sense is that we're early in the upturn, so probably in the first or second inning. We still have ways to go. History seems to, if you look at the previous peaks, 2014 was a peak, 2018 was a peak, 2022 was a peak. It's kind of every four years. Potentially 2026 is a peak, a peak year. I wouldn't be surprised given the dynamics of the last downturn that could extend to 2027. At the end of the day, we'll be prepared for any scenario.
I see. One of the interesting analysis TI has shown in that kind of peak- to- peak is that across peaks, you have grown, what is it, like 40%, right, or so?
7% CAGR.
7%.
That probably comes out about 40%, right?
Exactly. So you think we should expect the same from 2022 to whether it happens to be 2026 or 2027?
Again, unclear. It depends. But we want to be prepared for that. So that's why when we put the framework of the various revenue levels that we could get to in 2026, one of them was it was as low as $20 billion and as high as $26 billion. And again, that could happen in 2027 instead of 2026. But the $26 billion, that was representative of a 7% CAGR over four years.
I see. What does $26 billion mean, Rafael, in practice? Does it mean that you will have deployed capacity that can support $26 billion in sales, or you have the capability to be up to $26? I.e., let's assume that in 2026 or 2027, your top line is, I don't know, $20 billion, $21 billion, $22 billion, right? Some number lower than that. How b that reflect in your financials? What is the cost of getting that?
Yeah. We are going to be prepared for that level of revenue in terms of capacity, in terms of inventory. If instead we hit a, and we, as you know, later this year, we are going to tell the market if we are going to target $2 billion or $5 billion for CapEx next year. That plays into it. We still have some flexibility on that front. Regardless, if we do not get, I think maybe where you are getting to, if we do not get all the way to $24 billion-$26 billion, we will probably have some excess capacity versus what we could have done. That is the dynamic of this industry. If you do not have the capacity, you are for sure not going to get the revenue. You have to have the capacity in place. The good news is that capacity is long-lived.
It's okay to have it ahead of time. If we don't use it this time around, you'll use it in the future. We're about to close a factory in Texas that's 60 years old. It has equipment that, some equipment is probably 60 years old, but most of it is probably 20-30 years old. That equipment lasts for a long time.
Got it. I guess a different way of asking the question is that even if you are at peak sales in one of those years, hypothetically, but you're not at the installed capacity level, does it mean that you may not be at peak gross margins or peak free cash? How does that delta reflect in your financials?
Let me answer it this way. The way we are modeling, the way we are asking you or suggesting that you model our gross margin, which then falls through to free cash flow per share, which is what we focus on, is to use a 75%-85% fall- through and then adjust for depreciation, which, of course, depreciation is non-cash. You have to adjust for that in order to get to the gross margin. That 75%-85% includes most scenarios. For example, if we are at the higher end of that revenue, you are probably at the 85%. That accounts for the fact that you will be highly utilized, and therefore, you will get better fall- through. If you are at the lower end of those revenue scenarios, you will be at the lower end of that fall- through, and therefore, your gross margin will be lower.
In any case, I think what we model was the $22 billion, $24 billion, and $26 billion scenario would get us back to the low to mid-60% in gross margins. Only the $20 billion scenario would keep us in the high 50% gross margin.
Got it. Makes sense. If we look at the $2 billion-$5 billion CapEx range for next year, based on whatever consensus expectations are, and it does not have to be the right number, it says that your CapEx intensity, even when you tone CapEx down to what may be a more normalized level, is still quite high, right, 10%+ . As long as your top line continues to grow and you are spending at least, if your top line is in the $20 billion-plus range and your CapEx is in the $2 billion-plus range, it says that your CapEx intensity will consistently be double-digit. Is that a fair representation of CapEx intensity?
I think it's factual what you're concluding, or it's a good conclusion. It's only because $26 billion still includes some CapEx that is for phase two, meaning building buildings. It's hard to build half a building. You just build a full building. L5. 2 is a pretty expensive fab. We're going to be spending money on L5.2 in 2026, even in 2027. Once we get to phase three, which is incremental, that's when we're on a more steady state on CapEx intensity. What we have suggested that is a good way to model is pick a long-term growth rate for revenue, say 5%, to make the math easy. CapEx intensity should be 1.2 times that. So 6% CapEx as percent of revenue. If instead it's 7% revenue growth over the long term, then CapEx intensity is 8.4%.
That is before ITC incentives and anything on grants. That is a gross CapEx number that then comes down after we get incentives.
Got it. Makes sense. But $2 billion is kind of the minimum, right, you would say, to support a $20 billion-plus sales kind of margin?
For next year, given those phase- two investments that we still have, given where we think the potential revenue ranges, yeah, $2 billion is the minimum. Whether it's in July or October, we'll give you a better number. Do not expect $2 billion. Expect something higher than $2 billion. That'll be my teaser for today.
Understood. I appreciate teasers. The next question is, the reason that TI has focused so much on kind of made in the U.S. was a right strategy in a situation where there were more government incentives with the CHIPS Act and so forth in place, where there was not really a penalty for producing more in the U.S. Now it seems like with this whole trade and tariff situation, what if China has tariffs in place? What if the E.U. has tariffs in place? Does that not call for a more geographically dispersed footprint that many of your competitors have?
Let me first correct a statement you just made. Many of our competitors claim to have. This is like having Baskin-Robbins in every country, but you do not get the 31 flavors in every country. Fabs are a lot more complex than that. If you have fabs in every country, you actually only have a couple of flavors in each of those fabs. Only their biggest 300-mm factories in Texas have all the 31 flavors in order to ship. Let me get to the heart of your question, which is we have a strategy where we are deploying CapEx primarily in the United States, 300 mm, cost competitive, cost advantage. Those factories then can ship to the world. They are geopolitically dependable from a United States perspective and the western world perspective.
To your point, what happens if China puts tariffs as it threatened to do in April for a couple of weeks? We have four factories and foundry partners outside of the United States. We have two factories in Japan, one in China, one in Germany. We have foundries outside of the United States, foundry partners that can support us in shipping products that are non-U.S.-made into China as needed. There are some trade agreements between Malaysia, Philippines, and China that also allow products that even if they were built in the United States, they were fabbed in the United States. If they are AT'd in those countries, they enter China as country of origin for the Philippines and Malaysia. We have plenty of flexibility to satisfy our customers with parts that they need while minimizing or even entirely avoiding tariffs.
On the competition side, Rafael, I think on the analog side, it's pretty well understood that you guys are the leaders product breadth. The embedded business is a lot more competitive. I think the last count I had was like 30 or 40 companies who can make Arm-based microcontrollers. And many of them are very large established players in Europe, in Japan. China is investing a lot in there. So you think whether it is just that natural competition, everyone using Arm as the architecture, or whether it is just the fact that people would want to, in some situations, avoid buying products that are made in the U.S., do you think that changes the competitive landscape in your embedded business?
Yeah, I can start on that. I think we have actually more analog competitors than embedded. It's more diverse, I would say. Both markets, we've faced competitors for decades. What we've done is to invest into a portfolio that's both application-specific as well as more catalog-like so that we can sell to lots of different customers and build a revenue base that's very diverse. We're also able to build most of those products internally in Lehi. A big part of what we're taking from foundries and bringing it inside is embedded products. There may be some customers that want to buy from manufacturing outside of the U.S. We have foundry partners if they want that. There are a lot of customers that want products not made in China or Taiwan. We're very uniquely positioned to be able to do that.
Again, we do not look at that market as more competitive than analog. It really is about refreshing the products, having a good portfolio, having a broad portfolio, and also choosing where you make application-specific investments so that it is not just one or two customers, that it is half a dozen, dozen customers that we can sell that product to.
Got it. Makes sense. Now, I just wanted to go through first industrial trends and then automotive trends. On the industrial side, when we look at that market from outside, I think everyone gets that, yes, there is an inventory replenishment driver, and then there is an end market recovery driver. When it comes to end market, we do not really see global PMIs bursting above 50 or any. I think there is some skepticism about that driver. How long can just this inventory replenishment driver continue to grow industrial markets like above seasonal?
Yeah, let me try to give you an example. Let's see if this works. If end demand is 200 and you have inventory of 800 and you want to deplete that inventory, you place the order to your supplier, TI, for 50, only 50. In fact, you could go to 0, but let me use the example with 50. That way, you drain 150 every month or every quarter. That demand to us looks like it went from 200 at one point to 50. It got cut by three quarters. We go several quarters at just 50, 50, 50, 50. All of a sudden, that 800 mountain becomes 300 or 250 or 200. The customer says, you know, that's low enough. Now I want to just maintain that. In order to maintain that, they go from ordering 50 to 200. That's the bullwhip effect.
Now we go from 50 to 200. Now our demand quadrupled. Now the end demand all along was 200. Nothing changed. By the way, it quadrupled, and now it's going to stay at that level if the customer wants to maintain inventory. If all of a sudden, end demand, their true end demand is starting to go up to 225, they're going to place 225 almost plus a little more because now they want safety stock. Because now all of a sudden, they don't want one- quarter of inventory. They want a quarter- and- a- half. That's the bullwhip effect. All of a sudden, we're going to see 300. We think up until a quarter, two quarters ago, we were in the 50 level. If you look at industrial, industrial had been running 40% below peak, something like that, 40% below peak.
Some sectors 50% or much lower.
Some sectors even lower. Because there are many sectors, many customers, it's not everybody at the same time. They're going to go through this at slightly different times. We're getting to the point where we're getting to parity with the end demand. We'll see with the end demand that our customers see. We'll see what happens with macro and other things to see where that goes. That pop is not a one-time pop. It's there and it stays and potentially grows depending on what happens with end demand.
I guess that's, sorry, that's really my question. Have you already closed that gap? If we take the situation where end demand, what is that delta right now?
Unclear. It's 100,000 customers, 80,000 parts. They're all in, the permutations are infinite. Where each customer is in their cycle, which is different parts. What we do is we have a very good, robust engine inside of TI to decide what to build based on history, based on orders, based on how many customers buy it, based on how long the product life cycles are. That's why our inventory has gone, when we started the COVID cycle, we had 150 days of inventory, I think it was. Now we have 227 days of inventory.
Shouldn't all that data, Rafael, theoretically give you a reasonable sense of whether you are 10% below parity, 20% below parity? Like where?
I think, yeah, I'd go back to the chart that Rafael referred to earlier.
Slide 19.
Yeah, slide 19. You can see, and we've got red dots at the peak of each cycle. We bottomed sometime last year. That bottom, by the way, is below the previous bottom. It's the first time it's ever happened in 30 years of data. World GDP has grown over the last five years. How can the semiconductor units be so far below even the last trough? It's because we shipped too much stuff on the upside. The delta is inventory burn. That data is through November. We presented it early February. I think we missed December by a week or so. If you update those, you can see that hook that you see, that bottoming is continuing. If that long-term trend line that you see there, pick your percentage, but we're below it.
Yeah, you can eyeball it. It's trailing 12 months, so you got to be careful.
That's right.
I think one of our competitors actually gave a percent. Maybe that's why you're asking. Didn't someone say 10% or something like that?
Yeah. Does that seem reasonable? Low? High?
We just don't know. I don't want to venture a number because we don't know. We're prepared if it's 10% below. We're prepared if it's 30% below, and we're going to have a tremendous spike. Whatever comes our way between inventory and capacity, we'll be ready.
Can we not push back and say, if you are looking at historical data, a lot of that was driven by the rise of smartphones, the rise of China, industrialization, et cetera, that maybe it was on a slope that we may not see that same slope? To compare it to that past history may not be as useful.
Yeah. I think when we look at the market, that longer-term trend line kind of sits between 5% and 6%. It's a unit trend line. When you look at the secular trends in industrial, the secular trends in automotive, the secular trends in enterprise or servers and AI, there's just more content going in systems. Our belief longer-term is that those are the markets that will drive. There were other markets that drove it in the past. We're positioned well to take advantage of that secular growth.
Industrial, is it broad-based recovery? Is it an all-industrial market, especially factory automation? Is it China and non-China? Just give us a sense for how broad-based industrial is.
Yeah. As we saw through the first quarter, that strength was very broad-based. We've got about 10 sectors that make up industrial. They were all contributing growth to the total. There was not one region that was contributing. They are all in the same zip code from that. That is why after eight-10 quarters of sequential declines, we think the first signal kind of coming out is that cyclical recovery.
Got it. And then the last, actually, before I go to that last one, automotive. What are you seeing in that market? Is there a similar sense, X% below parity? I know there's a lot more direct customers rather than distribution in automotive. But give us a flavor for what are the trends you're seeing in automotive, China versus non-China.
Yeah. Just to give it a reference point and go back, automotive was the last market to begin to weaken almost two years after we saw weakening in personal electronics. From peak to trough, it was only down mid-teens, so fairly shallow. It has troughed and kind of run more sideways since then. Inside of that, if you go back a year, we had a couple of quarters of 20% + sequential growth in China, followed by a third quarter of high single digits. This last quarter, that has moderated. I think we are up low single digits both sequentially and year- on- year. That low single digits kind of feels like it is still running sideways. It does not feel like an inflection. That is kind of where it is at this point.
You think it's low single digit because the inventory restocking is done, or is there still a lot left? Is that the content argument because of the slowdown in EVs is not playing out the way one would think it should?
Yeah. I think it's hard to tell with just so few data points. EV expectations, of course, have come down. They're still expected to grow, though. That is content increases. If you look at cars being delivered today versus three or five or 10 years ago, there's just more content in them. I think one of the things that we saw is there used to be a 3 or 4x delta between a combustion engine car and an EV. If you look at a combustion engine car, everybody wants a good screen inside of it. They want a good cockpit experience. They want LED lights, LED turn signals, ADAS, safety systems. That has narrowed some inside of those cars. We're positioned well in China, in markets in the West.
We service about 1,000 different automotive OEMs out of our 65 product lines, 60-ish ship products into automotive. Very intentionally trying to be very broad-based so we're not dependent on a region or a technology or one sector inside of that.
I see. Rafael, this is more kind of a broad subjective question, but how are you feeling about the second half of the year? We are kind of middle of the year qualitatively versus what you might have thought, right, three or six . I get the cyclical recovery aspect, but there's also been a lot of tariffs and other macro factors. How are you feeling about the second half versus the first half of the year?
Yeah. As we said at the earnings call in April, and it's reflected in our guidance that we published at the time, we think we're in the early stages of a cyclical recovery.
So it means no seasonality? Seasonality will not play a role for the first.
I didn't say that.
Okay. Okay.
Unlike some of your competitors, and just for clarification for others, we do not update our guidance mid-quarter. We do not give a new guidance.
No body language.
We don't talk about further quarters other than the quarter. So that's all I can give you.
You are smiling, so I take that as a good sign.
It's a burrito he had.
Good to see you in an optimistic mood. Let me ask this last question on pricing. We have heard we had some of your peers present earlier, and they suggested that for some new designs, they are starting to see they are kind of proactively pricing a little more aggressively. My broad question is that, first, are you seeing any of that phenomena? In general, are customers becoming more aggressive? Because just like we see all this inventory on their balance sheets, they have to see it. They see all this capacity that you are putting on. Are they getting to be more aggressive in asking for price concessions?
Let me answer that last question first, and then the first question last. No. We have the leverage in that situation. It's not, "Oh, we have capacity, we have inventory, we're going to ask for a price concession. Not at all. We don't have any pressure to drain that inventory. That inventory is long-lasting. It's not a fire sale, nothing anywhere close to that. No, that's not an issue. On the first part of your question, the way we think about prices is we're a price taker. We're not a price setter. Whatever the market price is, we go and meet it. That's an oversimplification potentially because, of course, parts are so complicated. Price is but one of 10 different factors that they take into account.
At a high level, a voltage regulator for this many volts and that much current with this specs will sell for X amount. And then we go and price for X amount in order to win the business.
Got it.
What we're seeing, I'm sorry, what we're seeing from a trend standpoint is low single-digit price declines on an annual basis. That's what we've been saying for some time. It's the end of the.
A couple of years now.
Yeah, for a couple of years now. That is what we're seeing, and that's what we're expecting.
Got it. Final question. I know we are out of time. Capital returns, at what point do you think you will get back to kind of the prior trend line of buyback? As the CapEx comes down next year, does that start to help that?
Yeah. Big picture, over the last three, four years, our focus on capital allocation has been on the CapEx line to make those investments, on cash to protect those investments, and the dividend to make sure we maintain the dividend and continue to grow the dividend. What took a second place was the buybacks, M&A, and anything else on capital allocation. Another place that we protected was, of course, R&D. And R&D, we've actually continued to invest and even increase investment. To your point, once we get to even late this year, but definitely by next year, when the $5 billion goes to something lower, the tables start turning. We will do less CapEx. If we're in a cyclical upturn, revenue will be up. That should translate to we won't need as much cash on the balance sheet.
All that translates into more cash to return to owners. My expectation would be that you see buybacks to be higher than what they've been in the last few years.
With that, thank you. Thank you so much, Rafael.
All right. Thank you very much.