Here we go. Okay, good morning, everybody. Welcome to the Goldman Sachs Communication and Technology Conference. My name is Jim Schneider. I'm the Semiconductor Analyst here at Goldman Sachs, and it's my pleasure to welcome Texas Instruments and CEO Haviv Ilan to the stage today. Welcome, Haviv.
Thank you, Jim. Good morning, and thanks for having us here.
Thank you for being here. Maybe let's start on the macro side for a moment. Texas Instruments has a relatively broad end-market exposure. You see a lot across the globe, across a wide range of end-markets. Maybe give investors some color on where you're seeing incremental strength and other verticals where you're seeing incremental weakness or sluggishness.
Yeah, let's start with a really high level. If you think about what we are going through this year in 2025, I think two things are playing in parallel. First, we see a semiconductor recovery. I'll comment about the markets in a minute, but you know we throttled somewhere in the first half of 2024, I would argue Q1 2024, if you look at unit strength. We've seen a recovery since then. Different markets at different phases, right? I think this is continuing. We are now, Texas Instruments at least has commented about four out of the five markets we operate in are in recovery, with one of them, which is the automotive market, still not there, but I think it will come as well.
In parallel, trade tensions, geopolitical tensions are putting a lot of, there is a lot of disruption noise, and I think it also influences the way the market wants to go back to trendline. If you think about where we see strength, obviously, the markets that have recovered, the first were the personal electronics market. Some people would call it the consumer market. Comms is recovering, but data center is probably the fastest recovering market. This is where we see very strong growth at about 50%. It's also establishing almost like new peaks, so it's going back to the 2022 level. I do believe that in 2022, we were also shipping for inventory builds. Right now, when I see that market doing well, I get encouraged.
I would say that on the industrial side, we are seeing now two or three quarters of recovery, but at a, if you will, a slower pace. Other than aerospace and defense that established a new peak, we see most of the sectors maybe 20% to 40%, sometimes 35%, sometimes 30%, below their previous peak. I think that's related to the macro environment of, and I see it with my customers. They're cautious. They're making up their mind on capital investment. They want clarity, and I think that makes the recovery a little bit slower.
That's a good point. Just to clarify, on your last conference call, I think most investors seemed to perceive a downtick relative to your perceived level of bullishness coming into that call and the strength of the uptrend that we're likely to see. I think you may have taken issue with that, but maybe perhaps recalibrate for us how you see this in the level of tailwinds and just your overall level of optimism we see on recovery right now.
Yeah, I think I'll just be very open. You know, let's talk about the second quarter. The second quarter was a good quarter for the company. The company grew at mid-teens, I think 15 or 16% year over year. Also, sequentially, it was above average quarter. I think it grew at 9% sequentially, and it was a strong quarter. I do think that we have discovered along the quarter that the two elements that I've described are almost superpositioned. The recovery of the market, but also the trade tensions and maybe tariffs and pullings, you know, we kind of deciphered it a little bit better. Specifically, with our business in China, when we ended the quarter, we looked at our growth over there. It grew sequentially close to 20%. This is a little bit above natural, okay?
We made, you know, customers don't tell us why they pull parts and why they want more of them. We have to assume there was some sort of pull in there, and we've made that assumption. When we came into July, we just said, look, that was an element of Q2. This is why you see a Q3 a little bit, you know, a little bit lower than our traditional 6 or 7% sequential growth. I like to take it to the bigger picture, Jim. I think this is why we're here. When I look at the markets we operate in, and we are an analog and embedded company, I call it, we make foundational chips, but we operate in three markets that are amazingly promising. I'll start with the industrial market.
Although it's still not ripping back to the trendline, underneath, we are seeing investments related to automation, to electrification, to digitization that are going to continue secular growth, even if it's going to take a little bit more time. I think it's clear in automotive, there is content growth. This is why the automotive market, we talk about recovery, but even if you look at where we are versus the peak, we are single digits before or lower than the previous peak, and it will recover as well because of content growth. This is why we saw a very shallow down cycle, if you will, or recovery. I go back to the fastest growing market for us, which is data center. In today, in analog and embedded, still not a big part of the TAM, probably about 5 or so mid-single digit percentage of TAM, but growing very quickly.
If I look at TI, I'll give a little bit more color, and we'll provide a little bit more clarity as we arrange our markets differently into 2026. We'll call out data center, but if you think about the way I look at data center, it's running back to its 2022 levels, as I've said before. It's growing more than 50%. It's our fastest growing market this year, and I see more and more opportunity to play a bigger game for the company. The run rate is not, it's not low. It's like probably a $1 billion or $1.2 billion in between. That's how I expect us to finish 2025. I see a tremendous opportunity for growth in the future.
Today, maybe a single digit portion of the market, but growing to be, in my opinion, 20% of our market, not so far away from now because of the clip it runs in terms of growth. I'm excited about that. 75% to 80% of our revenue is centered around the three markets. This is why I'm excited, you know, and this is why we prepared the company for that in terms of investments, in terms of, of course, inventory, in terms of our product portfolio. We have to build them, to build our parts, and to compete at more sockets per system as we move forward.
Yeah, makes sense. Maybe on a very high level of strategic note, I believe you've been CEO for about two and a half years at this point. When you first started as CEO, what were the core issues of the company you wanted to address? Have you made progress over that period of time? What are the areas of business you're still working to improve?
Yeah, by the way, it feels more than two and a half years to me, but yeah, you're right. It has been two and a half years. I will say that when I took the company, the biggest thing was to learn or to apply the learnings from the previous upcycle. We felt that we could have done better in the upcycle. We were almost not prepared for success in terms of how we support our customers. That meant investment, and you've seen us starting that or announcing that back in 2021, get capacity in place, get capacity ahead of demand. If you see an opportunity, you can respond.
The problem with the fabs, and by the time you understand you have a surge of demand and to put that capacity in place, it's three to four years from moving dirt all the way to a clean room qualified, and it's just too late. Get capacity ahead of demand is one thing. The second thing is inventory. This is part of the reason we have built a strong inventory position right now to sell the future because that's how you react short-term to a surge of demand. I mentioned before, there is a bumpiness, there is noise. It can happen still in this cycle. We don't know. It's very noisy out there. Having a good inventory position to support whatever scenario the market will throw at you is very, very important. We modeled the previous surge of demand, and we want to support at least that in the future.
We've done that, and I am very pleased about these two elements because you can argue that we fully control our destiny, but it's in this period to build factories on budget, on time, with sophisticated technology, qualifying it on time. I'm very proud of the team. We have done it very well. Also, the inventory was built in a very thoughtful way, part by part, material by material, using data to inform us the level of inventory we need to have. We are really approaching, and Rafael touched it last week, we are approaching that level. We are getting to that comfort level that now we can support any scenario that the market will throw at us. Of course, you have to prepare the best case or the toughest surge that the market will throw at you.
If you think about the other element that I'm working on right now, and we'll continue in the second half of this decade, it is to work on our product portfolio. You know, I grew up making products in TI, and it's dear to my heart. I do feel that there are some obvious places where we had opportunity. The easiest example is embedded. You know, embedded, 20% of our business could be higher. We have not grown revenue peak to peak from 2018 to 2022. I put it on our portfolio. We are investing there. We can talk more about it later if you have questions, but I'm encouraged with that progress. There are also elements, once you get capacity and inventory in place for analog, you also can expand the portfolio over there.
We are expanding there as well into areas like high voltage, areas that are going to be important for data centers, more parts that can serve the automotive market, also areas in industrial that I feel we can further expand our portfolio and have this unique position. I think we are very unique here as a supplier to solve almost any problem on the board. Whether it's a $0.05 part and a general purpose amplifier or a voltage regulator, all the way up to a very sophisticated integrated application-specific part, and TI can play in both. Our competitive advantages are serving us across the portfolio. We want to be the house of solution for our customers, and I think our customers value that capability.
Yeah, great. Why don't we just kind of touch on some of your targets and goals you've laid out? I mean, at this kind of point in time when investors see cyclical recovery happen, many people immediately gravitate to gross margins and earnings power. However, you've always used free cash flow as sort of your North Star metric. Maybe you want to kind of touch on both those elements, the P&L and the free cash flow for a second. Gross margin has been complicated by a few factors at Texas Instruments because of the capacity additions you referenced earlier. So DNA is kind of stepping up to the $2 billion to $3 billion range. Inventories are fairly elevated at this point.
In light of the cyclical view you provided there, how should investors think about the path and maybe the timeline to get back to sort of mid-60% product gross margins?
Yeah, I mean, of course, when you make investments in capacity and inventory, the premise is that revenue will come, and that's how you expand margins over the long term. We even gave formulas of how it behaves for the company with a fourth row to the GPM line once we account for depreciation. Let me start to go back to our religion almost. We think we run the company on free cash flow per share. That's how we think. That's also how I talk to my team. I talk to the board. This is how we presented the targets for the company. We were very open about it in our capital management call exactly a year ago where we showed, look, this is a, let's give you a plan through the decade that gave you a span of revenue.
I think there is a funnel almost there that you can see at 2030. It represents about, you know, this is not exact math, but about $30 billion to $40 billion of revenue and how the company can behave in terms of free cash flow per share. It also gives you some intermediate points in 2026. Investors can be informed of our plan. This is a plan we put together from the get-go, and we are very committed to make it. Now, margins are a factor of how you get there, right? You know, you mentioned the biggest element is CapEx. CapEx is dependent on demand, especially now where we've built our clean room. We've got capacity ahead of demand. We are, as we call, on the phase three. You know, let's see what revenue wants to be and let's equip the factory accordingly. That will inform depreciation and margins.
What I can tell you, and you can run the math, when you fall through at 75% once you account for depreciation, you will see that in a case of a lower revenue plan, you just have to spend less because you've done it. Remember, part of the investment was to bring, this is a one-time thing, to bring manufacturing inside the company. You actually, manufacturing foundries in Taiwan, for example, are embedded business that 80% was done outside of the company. All of that goes inside. That's a one-time investment that, you know, once it's behind you, it's a very lucrative fall through, right? Because the variable cost to run a wafer externally versus your own fab is tremendous, tremendous fall through.
As you run your math and the fall throughs we provided, you will see, you know, that in any event, and that's how we model it, it's going to be in the high 60s easily. You can actually see that if you grow faster, you have to inject more CapEx. Depreciation still lingers longer, and it actually weighs itself down on margins. I want to say one more time, we don't run the company thinking that way. We run the company thinking about free cash flow per share. I think this is what owners at the end care about, and that's what also brings value long term for a shareholder.
Yeah, that's fair. On vis-a-vis the capital allocation that you talked about, you did lay out some intermediate scenarios, especially for 2026. I think that even at a lower revenue target and a lower CapEx target, your sort of like minimum free cash flow per share target was about $8.
Yeah, I think it was for the lower revenue cases, I think it was $8 to $9 of free cash flow per share. Yeah. This is not like we don't build the company. Part of our ambition is, and this is how we live our culture, is we think like owners that will own the company for decades. That's how we think. We think 10, 15, 20 years. I think we are very unique in that sense, and I'm very proud of it. I think we had a capital management call back in February, and we gave the same framework. Yes, that's our plan. There is no lie. You shall be at $8 in 2026 because then you are not optimal. If we had something to comment, we will tell you. We are always being very upfront with our owners.
I think our next call is going to be in February. We'll probably report back there, but I'm not expecting surprises there.
Okay, very good. Maybe pivoting back to your manufacturing strategy, your U.S.-centric manufacturing strategy for a second. I think that you've been very clear about your desire to build in the U.S. and have that be kind of the core of your manufacturing base, if not entirely. Maybe just give us a sense about how that approach got started, and what did you decide to do or how do you decide to take that approach versus a very different approach taken by many of your competitors, whether it's global or China for China or these other kind of ideas that your competitors have?
No, I think it's a great question. To us, as you know, we've embarked on this plan in 2021. I think at the time, there was not a lot of discussion about trade tensions and geopolitics. At least you've seen us steady hands since then. I think that's very important to remember how we think and what drives us. You talked about the U.S. capacity. I think about it as, you know, we call it dependable because I think when you build capacity, you want to serve the globe. I think our capacity is dependable for the globe. This is not based on, this is not U.S. or China or whatever. If I think about the Western part of the world, I'll put the U.S., I'll put Europe. You can also add areas like Japan, areas like Korea, and beyond.
To me, building a factory in our case in the U.S., and I'll talk about why in the U.S., is dependable. I think these are partners. You're even seeing early indicators of the trade discussions that they are partners. I think our factories in Utah and in Texas are not only supporting in an excellent way our U.S. customers, but also our Europeans and our other partners that I've mentioned. In that sense, it's very, very dependable. The reason of U.S. is because that's where we were. In manufacturing, we always talk about efficiency of manufacturing and cost. By the way, that you see some of our foundry, you want to concentrate your supply base. You get more efficiency when you have clusters of fabs co-located. You have access to talent. Your suppliers come closer to you and support you in a better way.
You can really utilize the factories better. Texas or North Texas is the area for us. We operate in three cities, but in Dallas, in Richardson, in Sherman, they're all kind of 40 miles apart. That's our cluster in North Texas. It's a great state to build a plant in, very business-friendly, good supply of power. Water is there for us. Most important, talent and know-how, rather than a new green point on the map that is a from zero exercise. The other part that we invest is in Utah. You can say Utah is a new place for you, but not for our team. Over there, we not only formed a new fab, we acquired an existing team, an existing factory. I think great investment for us. We like what the team is doing. They are ramping their new products. That's why we build there.
This is also what informed our decision. You know, we could have looked at an alternative to Lehi too, but we decided Utah is the best place, and I'm very happy about it. That investment is ongoing. It's also why we have still some minimum level of CapEx in 2026, but we'll be done with that towards the end of 2026, beginning of 2027. We are really complete in phase three. That's what we call geopolitical dependable capacity. The last point I will make, and I'm sure you'll have a follow-up on that, Jim, is that China is an important market for us. China is less than 20% of world GDP. To me, it's about the same, call it 20% of the semiconductor TAM that we want to support. We need to support China.
Today, we can support China from the U.S., but I have to be ready for a case we cannot. This is where we have the rest of the footprint of our capacity. We have a fab in Europe. We have two fabs in Japan. We have a fab in China. We have learned a little bit, we have tuned our machine in April because we got a little bit of a dry run. What happens if there are U.S.-based tariffs on whatever? We had a good answer to our customers. We are now making it even better. Some of the cases for technologies that were only built in one fab, we are now requalifying. We own our technology. We own our destiny. We are moving them into places like Japan, like Germany. We have our assembly and test houses in China and in Malaysia, in the Philippines.
When I talk with our, and I just was in China a month ago, the answer we have right now is excellent to our customers in China. Of course, we have to be prepared to be dynamic. I can't call what the trade tensions will develop into, but I think TI is a great answer for both our China-based customers, but also the outside of China-based customers. That also informs our strategy moving forward.
It sounds like the customer response to that strategy, including the potential kind of geographic diversification, has been positive.
Yes, again, look, I'm not trying to call what's going to happen in two, five, or seven years, and the rules can change every week. We saw a lot of dynamics in that, but you can't prepare yourself to any scenario. I like our position. I like our position both for our China customers. I like our position for the rest of the world. I think as a TI owner, you can feel good about TI is prepared to a set of scenarios. I think we actually can be one of the most resilient suppliers in our industry.
Great. Maybe moving on to embedded processing, that's an area you touched on a few minutes back. I think you've been quite positive on rejuvenating your embedded portfolio. Talked a little bit about how your team is looking to regain share. Can you kind of elaborate on what the medium to long-term goal is for embedded and their specific sort of applications or geographies you're looking to target to kind of like drive that share retake?
Yeah, again, embedded, I commented before, we will be the first one to say, hey, this is not something we are happy with the results that we have generated for our owners during the last run. If you look at the bigger picture, since 2017, we've been losing market share in embedded. I think we have stabilized that share loss. I think we have to establish a trend, but I will dare and say that it's stabilized. Amik, I will argue that we are starting to show, I want to make it a trend. Quarter in, quarter out, let's show a trend of market share gains in embedded. Embedded has to be an equal contributor to free cash flow per share growth in analog.
By the way, when you build your portfolio in embedded differently, and this is what we are doing right now, higher breadth of products, serving customers like industrial and automotive, having this diversity and longevity, and bringing most important manufacturing inside, inside Lehi, there is no reason that that will be very different than analog, especially when you go to the more catalog-ish part type of embedded. That's in the center of our strategy, with the low-power MCUs, with the wireless connectivity that is getting revived now in our portfolio, with real-time control. Think about all these data centers. Someone has to convert all the 800 volts coming into the data center in the future, all the way down to VCore.
Our DSPs or our C2000 MCUs are in the center of that platform, and we are winning share over there, and we plan to play a bigger role over there. Of course, analog will also help with all the power management parts that we are building. I am encouraged with where we have progress in embedded, but I think there is so much more opportunity. If you talk with the embedded processing team, their ambition is higher. They say, okay, our compare is easier. Meaning we are coming from a lower point. We need to outshine the analog folks. That's an internal competition in the company. I will tell you that I will be very pleased if embedded is at least an equal contributor to the long-term growth of free cash flow per share. I think they are set to do it.
You wouldn't see TI investing $15 billion in Lehi if we were not convinced that the team can do it.
Yeah, great. Industrial and market, if we could talk about that for a second. I think it's obviously you have a very, very broad range of applications here. Any applications that stand out to you in terms of their growth potential within industrial? Maybe you think, where do you think you're most differentiated in your product offerings relative to peers?
Yeah, look, on the industrial side, as I mentioned before, this is a huge market for us. I think at the peak, it was about 40% of our revenue or so, and so much room to grow. We are far away from trendline. I mentioned before, 30% lower than the previous peak, and the next peak should be higher if you think about secular growth. A lot of room. You do see, and we have nine sectors in industrial now. We kind of did some consolidation there. We look at the nine sectors. Some of them are already back to where they used to be, or back to the previous peak. Not like all of our peers. Aerospace and defense established a new peak for us, unfortunately, with what's going on globally.
Areas that are related to electrification, specifically around the data center, all this power delivery into the rack, if you will, is a great opportunity for Texas Instruments, both on analog and embedded. We are seeing again, we established again the same level of revenue on a quarterly basis versus the previous peak, but it's growing very, very quickly. I'm encouraged about that. The other market, and the biggest one for us, is automation robotics. I know this is an opportunity waiting to happen. I know that because I see what we're doing in our factories, inside Texas Instruments, inside our assembly and test. There is a clear efficiency gain on cost per pin that we make when we make a package using robotics and AGVs and even humanoids in our factories. It is just early phases, and we are investing. We are adding automation.
I think a lot of our customers are still in a wait and see mode related to the geopolitical tensions, related to the trade. Underneath that, there is a lot of work on automation, on digitization, on electrification, on the grid. You talk about energy infrastructure to serve all these data centers. I'm very, very excited about the long term. Can't wait for it to happen. I know right now we're getting impatient. The secret is to stay persistent here because the secular growth is real. Of course, anecdotally, I can give you a gazillion amount of examples, but we also have to show it in the numbers. Again and again, it showed in the past that it does it. I'm very convinced that we are going to see it, maybe later than sooner because of what we discussed before.
Once we have stability on all the noise we are feeling with trade and tariffs, I think it's going to be a great opportunity.
Yeah, a little more tactically, last quarter you talked about, you put up a very good growth in industrial. You talked about the pull-in effect earlier. Maybe just kind of walk us through exactly what you saw in Q2. Was that pull-in activity just limited to China? Do you believe it was broader than that? I mean, I guess, if you could share anything that's happened this quarter, that would be great.
Yeah, first on this quarter, let the quarter play out. That's what informed our Q3 discussion or July discussion on the earnings call. The biggest tale was China. China grew sequentially in that quarter 20% or 19%, I don't remember, 19%. This is above natural. We also saw signs that companies took some inventory and then started to deplete back some of it because they got comfortable again. To me, it's again noise. What I've learned also as an engineer on noise, the more you sample it, the less information you get. You just make more noise. Let it play out. To me, that doesn't change the big picture, Jim. I am convinced that the industrial secular trend continues and actually accelerates in the next decade.
Yeah. Automotive, you know, we've talked about it sort of lagging the other markets, but I think, you know, historically and consistently, you've always talked about that being a big opportunity for TI. Maybe just kind of talk about what is going on in that market more from a design perspective. Where are you seeing traction? Are you seeing most traction in EVs or ICE or ADAS systems, and then geographically, Europe, U.S., China, etc.?
Yeah, automotive is a great market. I think that's part of the reason you saw a correction. It is very minor, you know, peak to trough, call it mid to high single digits. We call it hovering at a trough, but if I look at our quarterly revenue in automotive over the span of three or four years, it's kind of, it's ready to grow again. It's been a very shallow correction. The reason is the secular growth. The reason is more content per vehicle. Our opportunity is across the sectors in automotive. It's ADAS, of course, and electrification on EVs that everybody talks about, but it's also on body and on new braking systems and on infotainment, more speakers and more displays and more connectivity and more, more, and more.
I don't think it runs forever because one day you do have enough speakers and massage seats and fans and displays in a car, but we are not there yet, right? I'm very encouraged about that. Geographically, look, the EV adoption in China is just, it's almost, I would say, done. Every new platform is planned on EVs. Customers prefer EVs. They're the most cool cars, the cars that people want to buy. I think 55% of new cars sold right now are EVs in China. I think that S-curve adoption is very quick to the point that even the government said, "Hey, slow down. Okay, let's tone down the craziness of that growth." I think worldwide it takes a little longer on EVs, and that's good. I think the market will be of higher quality.
If I look at our customers long term, they are shifting their R&D into EV platforms or plug-in hybrid EV. I think that trend continues, maybe at a slower pace versus China, but I'm very encouraged about that. Our opportunity in EVs, and especially on the embedded side, is to add or to attract or to attack more sockets. This goes back to our investment, you know, expand your portfolio, analog and embedded, have tailwinds or create tailwinds across the sectors inside the cars, across the sockets. I think we are well positioned to do that in automotive.
One thing we didn't talk about relative to your manufacturing strategy is just sort of like the, I think the industry's always focused on the issue of not just units, because we look at the same trend line you do, but also pricing. How do you think about pricing structurally going forward for the next several years? Is pricing a lever that you think TI can use to gain market share?
Look, in terms of what happened, we know what happened. Pricing, you have to go back to the late 1990s to see a price hike as the industry saw during the 2021, 2022, maybe even 2023, but mainly 2021, 2022 timeframe. I think TI was very conscious about, yeah, this is probably not going to last forever. If the market price changes, respond. We responded. I give the team credit of staying on sockets, not walking away and calling, "Hey, this is how to compete." We know how to compete. Sometimes we have to up our game on cost, and we've done that to kind of shrink dice sizes, get test programs to be more efficient. We stayed the course. I think we are gaining share with, you know, we are seeing this. I think pricing always gets rationalized over cycles.
This cycle right now is very, you know, is not very rapidly going back into trendline. The opportunity is not there yet, I would say. Prices we see go up and down all the time. For this year, I would say it behaves as we expected. It's about kind of a low single-digit decline versus last year. It all depends on supply and demand for the future, as you know. The beautiful thing about our position is we can play either game because of our cost structure, because of our competitive advantages, because of our appeal to multiple sockets. TI has an opportunity to play across sockets. I see it as a competitive advantage, meaning I don't have to sub-select. I don't have to walk away from sockets.
We can make good money and good profit and good growth of free cash flow per share across the portfolio, which I think is harder to do in a different manufacturing strategy.
Great. I think with that, we're out of time. Haviv, thank you for being with us. It was a great overview.
Thank you.
Thank you.