So I guess we will get started. Good morning, everyone. Thank you for coming today. I'm Stacy Rasgon. I'm Bernstein Senior Research Analyst. I cover the U.S. Semiconductor and Semicap space, and I can't express what an honor it is to have our guest here today, Haviv Ilan, the President and CEO of Texas Instruments. Now, before I start, I wanna mention, if you have questions you'd like to ask during the presentation, there should be a QR code in your program. It links to our Pigeonhole form. You can submit questions there, and we will have time for Q&A in the end. And so look like, you know, but I go on, I go on and on every year at this session.
TI is usually here, and I always talk about the more remarkable whole year transition story that benefited TI like over the years, and they've made it look easy. They like to say that people see them as boring. I'd say this, the company and the stock and the strategy have been a little less boring lately, and I wanna dig into that. So to help us understand why, it's my great pleasure to welcome Haviv. So thank you. Thank you so much for being with us here today.
Thank you, Stacy, and thanks for having us.
You can all hear?
Can you hear me?
Good. Awesome.
Okay, good. Thank you.
Great. So, Haviv, I wanna start out, I think, with the Elliott in the room. So clearly, like I don't think TI's ever had an activist involved. And maybe just to start out there, can you just a level set where we are, can you just go over what the actual baseline plan is for your capacity and CapEx investments? And then I wanna talk a little bit about, you know, the letter itself and the case that they're kinda laying out and, just get whatever you can... I don't know how much you can say at this point, but, like, whatever you can say, I'd love to hear your thoughts.
T hanks, thanks for the question, Stacy. I definitely expected it. But I can say a lot, so just to give you the facts, we have learned about it on Tuesday morning for the first time. So it's pretty fresh two days in. We are, of course, reviewing the notes and definitely look forward to engage in a constructive dialogue with Elliott. I will say that, as always, we run the company and make decisions to support the company and the entire shareholders of the company, so that will continue to guide us moving forward. Now, I won't go into the details today because I think it's fair we do that with Elliott first, but I would be happy to talk about our plans.
Yes.
First, you know, we can talk about the why and then the what.
Yeah
A nd provide as many details as needed, as you go through your questions.
Got it. So I guess just to level set on the current commitment. Let me know if I've got it wrong. So it's $5 billion in CapEx per year from 2022 to 2026, and that was taken up from 3.5 to 5. My impression there was you're effectively reinvesting the investment tax credit in more CapEx because the depreciation guidance at the time was sort of held constant. So $5 billion through 2026, and then 2027 and beyond was 10%-15% of revenue on CapEx to support, in theory, 7% to 10% revenue growth at that point. Is, do I have that right?
Well, you know, some cases and some, some of the points, yes, some of the points I would like to-
Please
P rovide a little bit more detail. But first, let's think about this investment phase.
Yeah, and why, to your point?
Yes.
Because, I mean, I think by the end of it, by end of 2026, maybe by 2030, the revenue capacity was, I can't even remember, was supposed to support... I can't even remember that number. Was it $30 billion? Maybe it was more.
We again, we'll talk through this, but let's talk about the why first. And the why start with the market opportunity.
Yeah.
I think it's obvious to us that the secular growth in industrial, in automotive, is actually accelerating. Of course, sometimes tough to see when you go through a correction, and there is an inventory correction in the market, but if you put it over a trend line, it's obvious that secular growth is there, and in my opinion, accelerating-
Okay
I nside this decade.
Because this is industrial auto. I mean, you've been pushing this. You were one of the first, I think, to-
Yeah
pivot toward, I mean, it's gotta be 10 to 15 years ago at least, right?
Which leads me to our position. And the position, of, you know, we talk a lot about revenue, but revenue doesn't get created, you know, just by letting the market grow. It's related to our investment and our strategy. And you're right, back in, I remember 2010, 2011, we said, "Hey, we are going to buy us our investment, our R&D, our SG&A towards these markets, with the vision that secular growth is gonna come." And we have done well. I mean, I have seen the R&D machine cranking out more and more parts. We used to have only one business unit building automotive parts, for example, in 2010. Each and every business unit in the company has, and the number of product lines is more than 60 product lines. They all build automotive products.
So the opportunity is very, very fast, and this is how we've built a position. You create a product portfolio, you engage with customers, you invest in your sales and application team, you invest in your website, which led to a strong position of almost 75% of our revenu. In 2023, in industrial and automotive. So to me, to be exposed at such a high level to, I believe, a couple of the fastest growing markets, in the semiconductor industry is a good position to have. The third one, and I think we talked about it even last year, there is also a tailwind on how customers make decisions.
And, you know, we run an average unit price at around $0.40 or so, or less, market. And these decisions would be made at the lowest level of, the engineering force at our customers. But today, especially in the last couple of years, we get, our customers receive guidelines from their leadership teams, from their CEOs, from their CPOs, about the dependability of capacity. TI is in a unique position, to provide this geopolitically dependable capacity at a very affordable cost, competitive, cost competitive, manner, and also, with, a muscle or the size of the capacity. So when you think about these three elements, and you say, "Hey, we have grown in industrial and automotive between 2013 to 2023 at 10% CAGR. Could we accelerate it this decade?
I think we can. Hard to see right now because we are going through an inventory correction, but when I think about the next peak, this is what guides our preparation. I wanna be ready to support this and our potential. And this is why we are invested in the capacity. Okay? And it's, it's done in three phases, and I will... I would like to go through them, if you will.
Yes, please.
Okay.
There's a lot of different pieces here.
Yeah, and it's very complex.
Yeah.
I can tell you that the execution of that plan is complex, and it's actually more than 3 years of an investment period. It's 6 years, starting in 2021. The acquisition of the asset from.
It was Lehi.
Lehi acquisition. Working through 2022 to qualify, the fab and to ramp it to production, which we have done at the end of 2022. I think I was here a year ago, and we talked about-
Yeah
This fab is ramping to production. We have continued to do that. And I love Lehi because simply it doesn't need revenue growth to get utilized. So you think about Lehi as a transfer fab. This is where every week we have, wafers moving away from TSMC, UMC into Lehi. It's done in two waves.
This is Analog, then, or it's Embedded?
Starting with embedded, but analog is following. Let me start, t o give you some more granularity here. 2022 was really F65, we call it, or 65nm
Okay
Embedded non-volatile memory, serving low power NPUs, serving DSPs, serving wireless connectivity. All of these parts are now qualified in Lehi and shipping from Lehi. Not only you get more control shipping out of Utah versus Taiwan, the cost of the pull through is tremendous. Think about what happened with wafer prices at the foundries in the last couple of years. The Lehi investment is predominantly done. The main power was there when you took the team. The pull through is almost a very real cost of a substrate and some gas and some electricity. So very beautiful transfer of revenue from a foundry into inside TI, and allows us also to win more business with this dependability of capacity. Lehi, in essence, as it ramps to full production somewhere in 2025, will be simply revenue replacement. It doesn't need growth t o ramp.
How much revenue does Lehi support when it's at full capacity?
So I think it's close to $4 billion, which is a line. And it depends on the mix. It's gonna start with the embedded, and later on have the what we call high-speed mixed signal of analog. And this is what we do right now, so 45nm this year and next year is analog. So when we complete that, more than $4 billion of revenue support, it's a very nice fall-through. So that's point number one. The point number two that is going right now is RFAB2.
Yes.
RFAB2 is again, a beautiful investment because the fabs are connected. So if you think about the way you add capacity-
RFAB1 was the original one back from 20-
Yeah
10 with, the command assets.
Correct. The cost to build RFAB1 was much lower. Of course, we bought equipment, bought equipment ten cents on the dollar. These days are over, and we can talk about why, but I think we all know why. And the RFAB2, we built the shell. I wish we had it earlier, supporting the previous cycle, but it support us right now. And the way it does it is, as you add equipment into RFAB2, wafers are traveling between the two fabs. So it's actually one big connected fab, which is a very efficient way to add modular modular way of adding equipment. But more importantly, the customers don't see it as a new factory, so I don't need to-
You don't have to re-qualify it.
No re-qualification or change notification. It's the same factory. So that serves us very well, especially now, because, Stacy, I'll be very frank, the part of the headwind we experienced in, 2021, 2022 was simply not having enough capacity. I was in the center of making some very tough decisions of who are the customers that are going to be told: "Hey, we have, we have to find a different supplier. It was very tough. I'm talking about a huge amount of revenue that we had to say, "Not right now.
Yeah.
Mainly outside of industrial and automotive. So think about consumers, think about enterprise, s ome of the comms, and, and tough.... but now as RFAB2 runs, we go back to the socket, and we go and re-win, and it's a very, very effective way to do it, very low cost. And that's one of the ways that RFAB right, RFAB2 right now is highly utilized.
Are you re-winning those sockets now?
We are, yes. And I think customers are smarter. I don't think we will be handed over 100% share, but let's start with 50, and work our fields from there. So I think that is very, very important. And again, RFAB2 is fully utilized these days.
It is full?
Fully utilized.
That's-
It's not at full capacity because equipment is being added.
But for the equipment that's there?
Yeah. I mean, when I say full, above, above 90%. Okay?
That supports another, what? $4 billion or $5 billion when it's full.
When it's fully built, it's gonna be close to $6 billion.
Okay.
Okay? And the other point, which is, again, now it's more tactical, but there is also transfer wafers to RFAB2. This is the first time we do it, it's not simple, but we are transferring wafers from our 150 millimeter wafer fabs t hat we announced, I think you maybe people don't remember that.
The-
$2 billion in Sherman. $2 billion a year kind of fab in Sherman, in Dallas. One 6 inch, 6 inch fab.
I thought you closed all your six-inch fab.
We haven't.
Okay.
We haven't. But we are doing something very unique, is modernizing these parts. So think about parts even from the Burr-Brown days, pre-2000.
Yeah.
Getting redesigned into modern 300mm wafer. So think about from 2,000 chips per wafer into 500,000. These are real cases. Cost becomes, quote, unquote, "close to zero." It stays with really packaged test costs. And the exciting part here is forget about the cost flow-through. It's really the modernization of the part, meaning customers now have the peace of mind that it can run for the next 40 or 50 years. So they know that they are out of old factories and they can design them into future systems, especially the lowest, the lower cost that we are running in them right now. So I'm excited about that. That's another...
Again, when we talk about capacity, yes, there is a $30 billion, but you know, these parts are gonna shut down. So this is also a tailwind that no one takes into, into effect. So what I'm excited about, and we build these factories during a downturn in revenue declining, because revenue is starting to stabilize and maybe starting to, to show momentum. Having these two factories running at full capacity, one on the expense of the foundry, one on the expense of about 200. So underutilization in 200 and 150 may be growing-
Yeah.
but our 300-millimeter wafer fabs that are running it full, and I think the-
Those ones are fully depreciated anyway, so $200.
Yeah, but you know, there's still, there is still labor there, and, you know, there is underutilization associated with them, but that's how it works, to utilize your best capacity or most modern capacity, especially if it prepares you for the future. I'm excited about that, and I think people underestimate that. So y ou will see that fall through. You don't have to wait 15 years for that, okay? That's coming now, as we speak. So this is the phase one, you know, LFAB2.
That's all phase one?
Phase I, yeah. This is LFAB, LFAB1 and RFAB 2. Now, the more complex one is, let's talk about LFAB2 and, talk about, Sherman-
Yeah.
Which is the heavy lifting in Texas. So, LFAB2, in my opinion, more straightforward, it's also on our slide. Large investment, our largest fab.
Eleven billion.
Yeah, and we support probably more than that on revenue.
Oh, wow!
Sophisticated type of product compared to what we do, so running from 65 all the way to 28.
Do you have any products on 28?
We are developing them now. It's our own process and the investment over these tailored to our portfolio, mainly led by embedded, but also high speed, mixed signal for analog. And these will ramp sometime in a... It will ramp out actually in RFAB-
Okay.
but will continue to ramp in LFAB2. In LFAB2, when we think about 2026, why we have 2026 as a milestone, this is when we are ready to receive two.
For LFAB?
Yeah.
Oh, okay.
LFAB2, LFAB2 is a beautiful fab because it's connected. So, we will add in a modular way to, as you know, function of revenue. And we said, I think in the last capital management, we will decide capital beyond 2026 based on where revenue is. First, what's your position, and then growth perspective.
Right.
So you said 10% to 15%, so I disagree. It's—
And that was-
It's those two lines.
On the slide, I think, right?
But it's a funnel.
Yeah.
It's a funnel. Okay, we also... It's thought about should we change the slide? We decided not to touch it, but I think you asked the question during that call. It can go from zero to-
Right, to ten. I mean, if, if revenues are growing 2% in 2027, I'm assuming not spending 50% on revenue.
Absolutely. And then, if you think about the fall through, it's immediate.
Right.
So, that's just to clarify, 2020, 2027 and beyond on that investment. Now, Sherman is a little bit more complex. The reason is, Sherman is what we call a new dot on the map. Luckily, the map is in proximity to North Texas. It's only 30 miles away, but it's a new fab. When you build a new clean room, and we are building two together because it's more efficient.
Yeah.
We said on the call, equip Sherman 1 , but what we say equip is really qualifying, because the capacity is gonna be very small, less than, I think, less than $0.5 billion of capacity in 2026. But you need to build it in order to not only qualify the fab.
Yeah.
You also have to get acceptance of your customers. So it's a long lead time type of approach, and that's how you think. You build the infrastructure, so you can build into that clean room when the time comes. Same example as LFAB2, but need some equipment in 2026. We are running the 65 analog right now, copper and aluminum.
Okay.
130 to 65.
Okay.
It's a very sophisticated Fab. The revenue also there-
That's the plan for Sherman right now?
Yeah, and revenue per fab is gonna be close to $9 billion over there.
There's gonna be 4 fabs eventually.
Correct.
Right?
But again, you build according to revenue.
Yea h.
So the beauty of SM1 and SM2 , and again, they are connected. You can think about it, about them as one large fab, as you add capacity in a modular way versus on revenue. So just to go into more details about our plan and also to clarify the 2027 and beyond. Now, we said we finish it in 2026. We always try to execute at the best level, and you know, it's within 2026. I don't have the exact month yet. But I think we should be done, some of them at the end of 2026, some of them earlier. So that will guide later on our capacity investment.
I guess if I add all of the incremental revenues, we had Lehi, which was $4, although maybe I don't know how much of it is incremental from where we are right now. It's somewhat useful, maybe it's $2.
Lehi, no, don't think about Lehi incremental revenue.
Okay. Okay, and then I guess same thing for them, too.
Lehi 1 .
So Lehi, Lehi 2 is $11 billion, and then it sounds like Sherman, when it's done, is, what? $36 billion, $ 9 billion a piece, is that about this?
That's what we've talked about, 10 to 15 years, you know, but we'll build them as the revenue comes, right?
So the incremental revenue capacity growth between now and 2026 is primarily Lehi 2 then, is what, is what you're suggesting?
No, it's zero from Lehi two. Lehi two, as we said-
The whole $11 billion is all-
By the end of 2022, it's clean room only, okay?
Okay.
The $11 billion is for, of course, clean room and equipment.
Okay.
Mostly equipment.
Okay.
Equipment doesn't start before 2027.
Okay. So there's $5 billion a year in spending beginning on 2026, but there's not a ton of incremental revenue capacity that's ready to go between now and 2026.
No, RFAB 2, RFAB 2 is incremental revenue-
Okay.
Capacity. Of course, we are right now winning back sockets. But it gives you more capacity. So capacity for the company will grow. We mentioned, I think, sustainability of $30 billion.
Yeah.
But that's, that's a sterile number. You think of it, there is like mix, fab never run at 100%, so... And I have a couple of fabs I need to shut down, okay?
Right.
So all of that gives you-
Yeah
the picture of where we can be. Now let's talk about the revenue, Stacy.
Yeah.
And I look at it peak to peak. So 2014, 2018, 2022. TI grew between every peak, every time, 21%, I think, 27%. Then you, you named a year and the growth of maybe, let's take 2026 as an example, because it's exactly that four year span.
Yeah.
I don't think revenue is gonna be flat to 2022. Especially when I was compressed in 2022.
Yeah.
And especially that-
What revenue do you think you gave up or left on the table in, like, 2021 and 2022 because of the constraints you had?
Look, we have done, we have done that discussion internally. It's significant. It was concentrated on the three markets I've mentioned.
'Cause it really is interesting. Like, if you actually look at the data over the last, like, four or five years, I mean, you clearly did lose share. I mean, do you think this is the primary reason? 'Cause it wasn't just NAND, it was analog, and it was embedded. If I just looked like strictly just at China as well, I'm sure we'll get to China, but everybody's worried about the Chinese taking share. And like, broadly, the multinational analog players have been gaining share in China, except for you. Like, you've been losing share in China over the last several years if you just look at the data. Do you think it was just all because of constraints, or you think there was something else going on? And is anything changing there?
So again, first, we like to talk about market share. And you see there is, you know, we have forecasted Q2. I think revenue is building momentum. Let it play out, okay? And we are not done with this asynchronous inventory digestion. So, let's go through this, but I will say that it is a big part of the issue. I will say that the decisions were to how to select the customers in which you will have to take a step backwards. I mentioned the markets. And, you know, in China, we had some big customers on the consumer side, and we had to give up some of the sockets. But they see what we are doing right now, and they're willing to entertain us.
We are going to fight hard and get back on the board. So, that's the China comment. I can talk more about China later, but to me, compression. The second point, and I was just about to say it, how we peaked in 2022, and I never peaked in 2022, but it actually peaked a year after-
Yeah
M iddle of 2023. But if you think about the pricing move of, of Texas Instruments versus competition, now I get into all the numbers, and I, I looked at the, we have to move... We have to distinguish between ASPs and pricing.
Yeah.
It's not always the same. There's just a ton of mix going on-
Yeah
Right now because of the asynchronous nature of the market. But if you look at the, if just to listen to what people did and also hearing it from our customers, I think we were the most customer friendly. We haven't hiked the price to the level that people did, and that's part of the compression of the peak in 2022. That's my belief. This is why I give ourselves a good chance to form a new peak. And also, the probability is very important, Stacy. How bad would it be to have the opportunity again in 2026 and not execute to it?
Yeah.
That would be devastating for the company. We are not going to let that happen.
Yeah. I mean, your inventory strategy flows into this as, as well.
Correct.
I find it interesting that, you know, I wondered about... I know you guys were constrained, but at the same time, you were also building a ton of inventory, and you still are. And again, I'm not gonna... I understand why. I understand the whole idea that the lifetime of the product, as we move on, it doesn't go stale. Did you just build the wrong kind of inventory, like, at that point? Was it just hard to match what you built with where the-
You're talking about the beginning of COVID, maybe?
The beginning. I mean-
But we built very-
A lot of inventory, and yet, like, anecdotally, a lot of the constraints were coming from TI, so.
No, so let's be fair. Maybe we've built a ton of inventory. I wish we built two tons of it.
Yeah. Okay.
Okay, that's it. So it was not the wrong inventory. It was the right one.
It's not enough of it?
Not enough, because this is why we were flying high at the beginning of the upcycle, because we had the inventory.
Yeah.
We were unique there. We had products on the shelf.
You were the only ones that actually decided to keep running when COVID first. Everybody else shut down. It was a problem, right?
Correct. And, I don't think it was well received at the time, but all we wish right now, we had more, right? And that's what guides our current inventory-
Yeah
You know, it's the right part because we have, we have the right data to know it's the right part. And we think it's gonna sell very well, especially if the surge is, you know, it's aggressive. Or at least I want to prepare at least to similar historical surges. And say, if you look at the units. I know you look at units. If you look at the units trend, forget about the AI. Look at the market without memory.
Wow!
We are below 2019 trough.
Right.
Are there more stocks for every end equipment?
Yeah.
There are.
Yeah.
How come we are lower than the trend line by so much?
I hear you, like I-
It's gonna catch up. It always does. Even if it's versus 2019 was the last industry downturn, 2023 was a downturn.
Non-memory revenues in 2023 were down about 2% year-over-year. It wasn't a bad downturn, but revenues were down. Units were down almost 20%. The only other worse year for year-over-year unit growth was 2001.
Correct.
Units in 2023 were below the prior downturn, 2019. ASPs were 30% higher for the industry, and X-memory were up 25%. Yes.
We studied the history as well. By the way, we look at what the WSTS come with soon in a couple of days.
Yeah
For April. But I will tell you that I think we're gonna break a new record of peak to trough on units, even versus the year 2000, you know, kind of model crisis.
Yeah.
So to me, we have to prepare because that thing always catches up. No one believes,
Yeah
I t during the trough, but when you see it, it's too late.
So let's talk about that.
Yeah.
So, broadly, but it's been a weird cycle, right? I mean, it's been very asynchronous, and we've had different end markets that have, like, peaked and troughed and leveled off and everything at different points. It does look like industrial, it has rolled over really, really hard, and I think peak to trough, a lot of these guys are down, I don't know, 40% or whatever, peak to trough on revenue. And auto starting to roll... It's not collapsing, I wouldn't say, but it's starting to roll over. It does look like broadly for even you and most of your peers, people are calling Q2 as the bottom. And you talked about, like, sort of like hopeful signs of recovery. I think there's still some controversy over the shape and trajectory of that recovery.
But, like, what are you seeing in the near term, I guess? I'm looking into the second half and maybe into the end of next year. Are we seeing actual signs of, like, robust recovery yet, or is it... I mean, people are arguing V-shaped or U-shaped or L-shaped, or like, and I know you guys usually don't like the cycle.
Yeah.
In this forum, I don't usually like to, like, drill a ton into short-term questions.
Yeah.
But especially if you're actually looking for a recovery and preparing for it, what are you seeing, in terms of where that recovery might actually be starting now?
Look, the duration of the down cycle was long because of this asynchronous-
Yeah
Nature. I think you're right. And we just see a first in, first out behavior on everything. If you look at markets, call it PE first in, call it,
It's kind of stabilized.
You're done. I think now it gains momentum.
Yeah.
So it's driving growth. There is the overload growth, and it's accelerating. And you'll see that, I believe, I mean, we'll have a good PE year in 2024. You think about the geographies, China first, the up cycle, China first for the down cycle. You know, are we seeing some signs in China that, you know, things can recover? Yeah, we mentioned PE, we mentioned U.S., you know, I think that look better as we go. Industrial, to your point, the last one to maybe to together with automotive.
Yeah.
But in different levels. So industrial, a very strong correction.
Yeah.
Over there, I don't think it's so, it's just inventory, overbuilding, and now just some... I will say that automotive, I'm excited about. First, you know, revenue for TI, I mentioned 10-20 2013-2023, 10% growth for industrial and-
Yeah
Auto, automotive. Auto be 15.
Yeah.
Which I believe-
10% of your revenue is, like, 25 years.
It's actually 34 and growing. Yeah, and growing.
Yeah.
So it's and if you think about last year grew 17%. If you think about Q1, it declined year-over-year for the first time, it's -2%, so very, you know, low single digits. But I think we are seeing a very shallow trough. So and I think it's related to to first secular growth in automotive. Second our position. I think our position is good. So I'm excited about automotive. I think we see momentum building sooner than later. So that's on the automotive side. And yes, you know, we've guided mid-single digits, I think, for Q2. I think in some areas, as I mentioned, like consumer and maybe consumer, maybe enterprise and and communication following.
And I think industrial will be the last to correct, because it declined more or less. It's the tough sectors in industrial declined during the end of last year. Last point I will make on industrial, and you know it, hundreds of end equipment.
Yeah.
Not everyone is the same. You know, power tools and thermostats, call it, appliances and building automation, are already looking t o gain momentum. Factory automation, medical, are still in the correction phase. So every, every end equipment has its own distinction. Now, you asked about the when the market goes back up. When all these... Immediately, when I see a decline, I like to count to four, four quarters. When all these markets go through the, that one year of, inventory correction, we are somewhere at the end of 2024 when it's done, I think. So we can think about, you know, an opportunity moving forward.
I s it fair to say that, like, you guys clearly do a lot more in-house and more control. Everything you're doing is you have more control over your own destiny and... But you don't have as much of a channel, you don't have as much of a buffer. Is it fair to say that what you're seeing is likely closer to what your end customers are actually seeing versus, like, having that in between. Like, people look at you as you were the first to actually start to see the decline, right? Maybe for that reason. People wonder, like, is that a-
No, I think it's a valid cause. And we can see it mainly when we go just in time to customers. So not only we are direct, some of the customers work with consignment. So we build to a buffer, and they consume it real time. So I believe we can see real-time behavior of customers, and Autopilot is a great example. Most of our customer base is on assignment.
Yeah. A nd I think we get real-time signal. On the industrial side, there is a heavier reliance on the channel, on the distribution. Our footprint is not large.
Yeah. What's your distribution revenue mix? It's not that high anymore.
25% in 2023, Mike? Yeah, so 25. And yeah, it's probably gonna go around that 25, maybe a little bit lower over time. But we also are excited about the investment in the other part of the dual channel, as we call it, which is our website, ti.com, for e-commerce. That's something that we are modernizing. We like the investment we are making there.
What are your sales going through there, by the way? You gave some numbers on the cash a couple of years ago on the capital.
I mean, it did very well in during the upturn, because that's where customers went when they needed immediate availability. I think it stabilized to a lower level, but it's still a much more significant versus pre-COVID. So, and it's an established, and we are making investments over there, and this is where data, information-
Yeah
can use bigger wafers are gonna serve us in the future. We think this last mile, as we call it, is kind of messy in our industry, and why not modernizing it and take advantage of the IT developments in the world to serve our customers better? I think our customers expect that, and those who are getting on that modernized channel are excited about that.
I guess it gives you, like, an early view to what customers are doing.
It's all real tim e.
Yeah. So I do wanna talk about China.
China, yeah.
So even beyond TI, like a lot of investors, I mean, they, they see just a ton of capacity coming on in China. There's a lot of question, what's gonna go into those fabs? It's all lagging edge because they can't do it again. So people clearly worry about the potential for, like, local Chinese replacement, particularly analog and other, and other lagging edge technologies. And, I think you guys have, I don't know, 50% of your revenue goes into China, although I, I don't think that a lot of that's multinationals. What's, first, I guess, what's, what's the local consumption in China be around 20%, or?
So I think you quoted a ship-to number.
Yeah.
I think-
That's something about-
With our largest customer, that-
Yeah, yeah
T hat's China, right?
I get that.
So, I think we stopped reporting that because I think it's relevant. So we are now reporting. We are reporting headquarters in China revenue.
Okay.
I mean, the customer headquarters is in China.
Okay.
I think we think you wanted 17%, Mike?
Okay.
17%. It was higher during COVID, but also, as I said, China was first into the down cycle, and I think it's probably gonna be usually it's first in, first out, so we'll see that. And just on the China side, you know, let's put things in perspective about capacity. China is about 17%-20% of all GDP. So for me, the market opportunity is correlated to GDP, I think it is opportunity. But it's still a very important viable market if you wanna compete at, and they have enough capacity to serve them and more, okay? Because their buildup of fab and, and not, I think, for China. They have ambitions to get into our... Think about big parts sitting in our servers-
Yeah
In our medical equipment, in our EVs. I'm not excited about it. I'm sure, I'm not sure the U.S. government is excited about it. I know that my customers are excited about it. So this is where this geopolitically dependable capacity-
Yeah
Is coming in, in big role on, on the future of China. And I think it's very, very important to the parameter of how customers make decisions. So that capacity, again, we compete, we see in China, for China mainly. We see a subset of competitors, all fabless, or predominantly fabless, and we compete. You know, luckily, the set of competitive advantages that we've built over the years is technology and manufacturing. As you said, relying on ourselves, cost inventive. The breadth of the portfolio, which is unique versus the set of competitors in China. The channel advantage, the reach, basic touch to so many customers, some of them very small, that the local players don't even know.
Last but not least, the position we've built in industrial and automotive. It's very similar in China. Also, close to 80% of our revenue in China is on industrial and automotive. Give ourselves a good chance to compete, but it is becoming more competitive, and that's what's going on from a high level over there.
Do you think the pricing trends in China are different from what you've seen elsewhere in the world? Like, do you compete harder in China versus where you have to compete elsewhere?
I do, because I think the local, or the, the emerging competition in China is highly accepted in China. So, and, you know, I think TI can gain premium over there, because we have the breadth, we have the quality, we've been trusted supplier for many years, especially in industrial and automotive. There is a moat around this type of market. But, you know, these guys are capable, and I don't want my team to think they can do only very simple parts, because I've seen these people building more and more complex set of solutions, and we need to compete across the board. Most on the catalog type of of more general purpose parts, but also on the application specific, the ASPs are higher.
Of course, the market price is set by them, but it's not a walkaway market price for us, simply because we have the cost competitiveness to play that game.
I guess to touch on that concept of geopolitical capacity, how important is the CHIPS Act for you? If it wasn't for the CHIPS Act, would you be pursuing the same type of strategy that you're pursuing right now?
Y eah, I think we talked about it last year, and, and look, every time you make an investment, you look at the probabilities and affordability, and it's always a, you know, a set of, parameters. And we were very clear that, we decided to increase our investment or, lean in on capacity because we got some help, or also because we got some help from ITC. So I think we had a plan at the beginning: Hey, let's build for a lower level of revenue opportunity. And we said: Hey, let's take it a step up because we are getting help. Unfortunately, I cannot communicate,
Yeah
W here we are because we are not done yet. So we've submitted the application. Of course, the ITC is helping, but beyond the ITC, there is a-
The ITC is actually bigger, right? I mean-
It's expected to be bigger-
Yeah
Because 25% of your investment in the U.S. But, you know, the grants are also important.
Yeah.
We are waiting to see what, where does that land. I think we've submitted it end of last year, and we should hear, I believe, in the coming months where we landed. So this is definitely helping, but compare that to what China does for-
Yeah
For their industry, you know, it's,
Yeah.
Health is important for us to compete-
Yeah
With some OTC other players.
Got it. I wanna ask about a couple of the growth markets that my outlines include. So how much EV exposure does TI...? Because I was—it's not really, really a joke. I was just like, I would say, like, you're as happy to sell, like, the headlight controller in the car as you are anything else, right? You know, it doesn't seem to be, you're just playing the general content trend. But, like, how do we think about EVs and their impact on technology?
First, let me just, I don't know if it's correction or clarification.
Correct or clarify anything I say.
Automotive, from a very complex ADAS processor-
Yeah
T ens of dollars, all the way to lighting controllers.
Yeah.
So we don't like to brag about radar solutions at 77 gigahertz right now, but our FPD-Link high-speed connectivity that runs at hundreds of millions a year, and OBC controller for onboard charger with our DSP solutions. But we ship them, and we ship them at a high volume. But we also ship, as you said, a bunch of catalog parts for automotive that we are very proud of, and they all add up to a very significant opportunity in automotive. Automotive, I think right now in Q1, was the size of industrial already. They ran at the same level I would say that the breadth of the opportunity is only growing. Now, if you go to EV, more than 2x, Stacy. When you look at our exposure and our opportunity on.
Like, how much of your-
I search as EV.
How much of your auto revenue today is EV? I don't know if it's.. There's going to be overlap. I'm assuming-
I don't have that, I don't have that, like, in front of me. Let me check that. But it's significant, and the beauty of it, EVs are not fully, run their S-curve reduction, right? So there is room to grow over there. But as I talked about the last decade, it was mainly on ICE. And I think about the future ten years, it's mainly on-
Yeah
A mix of ICE and EVs.
Do you think your content and opportunity in EV is too excessive then?
Our opportunity on EV, and we've done that math, and we kind of look at the dollar opportunity is more than $1,000, and it's, it's-
Let's see apple. How much TI... What's the average TI content in the average car today?
Oh, today, we are shipping $hundreds, you know, in some cases, close to $1,000, in some cases, $4.
Wow!
But very high content. It depends, also on the-
Yeah
How many cameras, how many screens, you know, it depends on the, on the vehicle tier. But you look at the high tiers, well, about $500 per vehicle.
Got it. And then I'd, I'd be remiss. I don't know if you want to hear, but I have to ask about AI. Like, so is there-
Yeah.
Is there an AI play for TI? Like, where do you guys benefit from that?
Yeah, a growing opportunity.
Okay.
So, let's say the biggest one I'm excited about is, you know, think about cloud and servers and, the compute guys. They are starting powers in levels that I couldn't imagine, you know, above 1,000 amps going into a processor. And, you know, you see, start to see, you know, close to a kilowatt and higher. And if you look at our roadmap, it just wants to be higher, right? So the way you serve that is what we call a multi-phase, control solution. Many, many parts going and serving this, this process. So I'm excited about this, what Sherman 1 comes into play, because this is our newest technology.
How low does Sherman go down to?
At 65 nanometer.
Okay.
We think we are building the best BCD process on Earth. And, in terms of, the opportunity to serve a very fast-growing market from the U.S. again, because you want these, these TI PMIC parts to come from dependable source, I see a growing opportunity. The product portfolio is there. We are talking with all the big ones, and I'm excited about the progress. We're also gaining share platform to platform, but I think we are still kind of only scratching the surface. So I think at the peak it ran... I actually looked at it lately, close to a billion-dollar business. If you look at datacom in the server and also power, but power is the biggest one.
Okay.
So that's on AI. The second part, and, you know, maybe I should have mentioned it even last year when you asked me the question, think about the edge. We have a good low power processing business. Think about our DSP days. You know, think about neural networks and learning. These are, these are all metric manipulation. We have a very good accelerators that we've developed over the years. They sit in our processors, and when you think about inference, TI is there. DMS or Driver Monitoring Systems
We have a leading solution winning with many, many OEMs, and that is using that machine learning accelerator. Think about, even, you know, arc detection in energy infrastructure, TI DSPs are solving that problem as well. So AI at the edge, for instance, is a growing opportunity for the company, and we have early wins. The last one early, but robotics, or content per robot for the company is significant. Tens of microcontrollers, some of them with machine vision capabilities. So that is still... Think about, you know, the humanoids or human robotics. They have a lot of content for AI.
Do you think that's gonna be a thing, humanoid robots?
I know that there is a lot of content per robot. I just know how to serve the end equipment. Calling out how many units will be sold in the future, I was never good enough to guess that. But, winning the socket is where we are busy.
Yea Yeah. I wanna ask a little bit more about, like, industry structures, but particularly on M&A. So, like, TI has not really participated in, like, large scale M&A for quite a while. Like, you've had some competitors who have, and I'd argue have been fairly successful at it. What are your thoughts on that? Do you... Is there anything, I guess, like, have your priorities on what you would be looking for if you were looking for deals, have those changed at all over the years, or is that what kept you out of it?
Look, I think this is... Nothing has changed there. We are always looking. It's, we always said it's gonna be analog-centric. We always look at, hey, is there a technology that can complement our portfolio? We don't have a, you know, a lot there when you look at it. Even in a, take an example of wide band, wide bandgap, technology. Like, again, we are investing this in internally. So I think we many times prefer the make versus the buy opportunity. But, you know, depending on the right time, and also need, we need to think about affordability. Right now we are so busy in getting ahead on our capacity and getting back to that, frequency of share trend line, as we mentioned.
That's a priority right now, but as we go back to the trend, you know, maybe opportunities will pop up and, you know, TI will make a move. But at least what I like about our position right now, there is nothing that we really need to build the scale and, t o set the future-
Yeah
For the company.
I guess talk about that free cash flow per trend.
Yeah.
Free cash flow per share trend.
Yeah.
Do you have a number in mind? I know you've given, like, sort of revenue capacity, like, kind of, kind of numbers. And, and they go back to the Elliott letter, I know they were throwing out $9, I think it was $11.27. I presume you guys have done the math on, "Hey, if we cut our CapEx in half, the free cash flow per share goes up." But do you have some idea of, like, what that, that free cash flow per, per share trend ought to get you in 2026 or 2027 or 2030, or, like, whatever it is? Like, what, what are we shooting for here?
Yeah, I thought we were very. I think we had a slide in capital management, right?
That's why I'm asking. I can't remember what the number was.
Okay. It's 12 in 2026. It's 13.3 in 2027. Not that I remember, okay? But that's the trend line. That's the trend line.
That was not a target, that was the-
No, but-
Okay
T hat's good. The trend line is there, and I think Mike said or Dave said it, this is what guides us. That trend line, we haven't shown a trend line of 15 years, okay? Not that I'm picking on you, about 15 years. But, you know, we show the 2004, I think, through 2022, and they extrapolated that trend line through 2030. So when I think about when you get back, it's within that timeframe. And I don't have 2030 in my brain, okay? I, I see the modular capacity plan coming in play as soon as, you know, back end of 2026 to 2027.
That's the expectation. I mean, that's expectation from the team. Of course, revenue has to do something-
Sure
D uring that time. But even if revenue wants to be very poor, then what do you do with CapEx? You just, you're done. Okay? You let it fall through, and then it will fall through very nicely. So we have run all kind of scenarios over there. I'm excited about it. And the beauty of where we are, I don't think you have to wait two or three cycles to get that, to come into play. So we are, I think, marching into the next cycle. It's the first time we have predicted or guided for sequential growth after two years, I think. Two years, I think.
We've actually been down three quarters in a row, six months.
Seven, maybe. Yeah, something like that. But guided the first time out for maybe two years. And it's easier to invest when you have revenue momentum on your side. So I think that's what my... When I compare QTA a year ago, I see markets joining the inventory digestion, and I think a big part of it is behind us. There is a momentum building on revenue, the factory that we are operating, producing chips at a wonderful cost, a wonderful alternative cost. And as revenue grows, I think that investment will come to fruition very quickly. So I'm super excited about that space.
Got it. That makes sense. Look, we've got about a minute and a half left.
Yeah.
You've kind of been doing it, but I'll give you your soapbox. Why should investors buy TI stock today?
Okay. First, I think, and again, the investment, yes, we make investment, thinking like long-term owners, and I think about those as 10 or 15 years. But you don't need to wait 10 or 15 years for this investment to come to fruition. We are a very unique supplier in terms of have a combination of enough capacity, the cost competitive or affordable position, and very dependable. And as customers hear more and more about it, and as revenue starts to build momentum again after a very tough cycle, I think the company could not be in better shape than it is today. I'm excited about the future, and I'm excited for us and for all of our shareholders. So that would be, that would be my short pitch.
Got it. I think that's a great place to leave it, Haviv. Thank you so much.
Thank you, Stacy. Thank you.