Broadcom Inc. (AVGO) Earnings Call Transcript & Summary

All right. Thank you. Hello, everyone. Welcome to this keynote session at our Bank of America Global Technology Conference. I'm Vivek Arya. I head the semiconductors and semi cap equipment research here at BofA Securities. And it's my absolute pleasure and honor to have Hock Tan, the President and CEO of Broadcom, join us for this keynote session. I think it's -- suffice to say, Hock has kind of been a visionary in this industry. And we are really delighted that he could join us to share his thoughts and insights about the industry and about Broadcom, which right now with the profitability, I think, is arguably the most successful in terms of profitability, the industry and the company -- in this semiconductor industry. So with that, a very warm welcome to you, Hock. Thank you for joining us today.

Thank you for inviting me, Vivek. Happy to be here.

So maybe Hock, just to kind of set the stage, give us a state of the union as you see it in the industry. It's going through a lot of changes that is unprecedented, kind of broad-based demand. We are also hearing of capacity shortages left and right. So give us the state of the union of the industry as you see it. What are the few things that interest and excite you the most? But then also what are the few things that you are keeping an eye on from a headwind perspective?

Sure. Sure, Vivek. And that's a very, very interesting answer -- very interesting questions. So let me try to address that because there are multiple factors moving this around. But the most obvious one is, obviously, to state the obvious, we are in a semiconductor up cycle, and it's not something generically that unusual. This industry, over the last 40, 50 years, has been going through cycles. Over the last 10 years, perhaps those cycles have mitigated because the industry is a maturing industry even as technology in different functionality applications in semiconductors continue to evolve. And it has -- matured sort of cycles have kind of been more mitigated but still cycles. And we are right in the middle of one of those cycles. And what may have been different in this cycle is that it is matching the emergence out of the pandemic. As you know, pandemic hit in 2020. And we're kind of the economy that is kind of a stage of broad economy -- or recovery. But even during the pandemic, in the midst of the pandemic last year, especially the second half of last year, we have seen certain key trends, one of which is for certain technologies. And I'll touch on it in a second. There has been accelerated adoption to adapt to a work-from-home, learn-from-home, play-from-home environment and we've seen that. And one of the exciting things out of it that we see Broadcom side and more than a few other items more than that, obviously, we see the strength in 5G smartphones. Everybody wants the best connectivity out there given that they are working away from offices where there's infrastructure and also the need for PCs. What has been a slow-moving market has all of a sudden become very hard PCs. But what particularly, from our side on the infrastructure, we saw a lot is the strength in connection to the home broadband. Now use of -- [indiscernible] for the last 10 years, a very boring pipe, kind of pipeline business, connecting to the homes. What we have seen now is that connection, especially on the last mile to homes, has taken on a lot of interest, both political but also out of necessity. And we've seen that popping up last year, shown very strong. And this past quarter, when we announced it, we are up 28% year-on-year demand. Broadband being -- for us, wireline broadband is PON and DSL and WiFi connectivity as well as for the cable operators, cable modem, the latest generation. What's also driving broadband even more now that we're seeing, and I almost call it, it's a surge in interest in broadband is really driven by 5G in my view. So a lot of telcos out there, Europe, U.S., in particular, have spent a lot of money investing in spectrum and also in some of the front haul, a lot of the front-haul initial investment, initial cycle on 5G networks, on 5G spectrum. And the challenge now is you do that, you attract subscribers. Just 8 years ago, they did it to attract more subscribers into their networks. Same thing for 5G. It's a fight, marketing fight for subscribers. The question is, so find a use case for 5G that will get them going early because use cases will take time. It's essential, but this is not going to happen anytime soon. So what is happening and we are seeing out there in quite a few telcos is to converge it, as they call it, I call it to bundle it with broadband. And suddenly, there is a strong set of initiatives by these telcos to really create bundles where broadband -- more than bundle, gateways to your house where broadband combines with 5G cell. That's a very exciting opportunity for someone like Broadcom, where we do the silicon for the broadband wireline. Of course, we don't do it for the cellular, but we just do it for -- whether it's copper, fiber and definitely WiFi. So there's one strong trend that pops out of this whole thing. Of course, the other area is working from home or living at home. You see we go online a lot. So we see a lot of hyper cloud, and we have seen it since middle of last year, late last year, continuing to today. A lot of strength in investment by the hyperscale guys in scaling out their data centers. We see it particularly in the networks, the network silicon, the switching, the routing silicon as well as some of the compute offload that we provide to the hyper cloud very strongly. And -- but most of all, most recently, as we should all expect, but we'd like to see that happen, is over the last -- since our last earnings call just over 3 months ago, we now have seen strong recovery in orders. Now these are not revenues just yet because lead times for delivery of semiconductors, as you all know, has stretched up quite a bit. But these are orders from enterprise, what I consider traditional enterprise, which has been a market over the last year down on its back. We've started to see revenue recovery in our fiscal Q1, beginning of this calendar year. We still see traditional enterprise being down. We see it getting better last quarter. We will see it continue to improve by the end of this year to a point where you might reach breakeven, but we will see. I believe based on the bookings, we are getting orders that are accelerating strong recovery on enterprise demand late this year, definitely '22.

Got it. So we will talk about the different growth drivers in a second, but under your leadership, Avago and then Broadcom have gone through a very interesting M&A journey in the space, right, starting from LSI and then the legacy Broadcom. But in 2016, you kind of changed strategy and the path somewhat, right, with Brocade and then CA and then Symantec. So talk to us what is kind of that unifying strategic intent behind all this M&A? And how should that inform investors about how you are thinking about M&A going forward?

Okay. That's a hell of a question, Vivek. And it makes me think back to how this whole journey, this whole strategy and business model we have -- that we have today was put in place. And truly, we haven't really changed that basic underlying business model and thesis. And it's -- I'll put it this way. We look at semiconductors, that's where our roots of Broadcom was and Avago. But we look at semiconductors as an enormous profit pool, multiple verticals under -- over -- in an industry of supplying key components or better yet, supplying technology, [ deeps ] of technology embedded in whether it's silicon. And by the way, the roots of Broadcom started with Avago. And Avago was the spin-off of HP's semiconductor group. It was the HP semiconductor group many, many years ago. And it came out [ about ] 15 years ago from HP Agilent and Avago. Its roots, its strengths were not silicon, even though it's semiconductor. [ You want ] compound semiconductor, we're the guys. We do all this, gallium arsenide, indium phosphide, all these lasers, all those RF stuff, that's the roots. From there, we broadened into silicon. And today, you correctly said, we added on software, infrastructure software elements. So unifying thinking behind each of those product lines within the technologies we have expanded into is simply this, that within technology, there are verticals, there are products driven by specific technology applications, which are mission critical, which are very, very, very, I guess, critical and will sustain. And the whole over -- unifying strategy of Broadcom is to identify those product lines and either grow them or buy them. Now with Avago, mostly compound semiconductors, we had inherited from HP some 8 -- 7, 8 strategic product lines that meet those criteria. Those criteria, I've articulated many times, are those that have -- there are verticals -- we could just call them niche markets that are -- and some niche markets get enormous because the mass market hits toward -- most weighted like RF but -- as an example. But those are markets which are sustainable. Best way to look at it, it has been around for the last 20 years. As far as we can see out there, they'll be around for the next 20 years. It's all about sustainability. And two, sustainable markets is one where we are the market leader and secondly, the technology leader. And the underlying investment model or thesis is that owning those assets and continuing to reinvest to ensure you remain the leader in technology and market will ensure a growing and sustainable source of profits. So it's all about getting those. We started with, I think it was 8, and we decided buying was -- in this market was definitely more predictable, more likely to be successful and trying to organically grow in the market where incumbency, #1 players are pretty much in place. So we embark on this series of acquisitions simply because it's less risky, more predictable, more likely, playing the odds to acquire winners then to try to create winners in verticals where we have limited presence. And that's how we ended up in today's environment today where -- by the way, we started with largely compound semiconductor with acquisition of LSI, subsequently Broadcom Inc. We broadened into silicon semiconductors. And since 2016 with Brocade and thereafter CA and Symantec, we broadened further into infrastructure software. Each of those that we acquire, each of the product lines we have today as a franchise have those common characteristics. And that's one -- it is -- we're creating a platform of technology into the ecosystem. And underlying basis is create value through sustainable and growing revenue streams.

Got it. So Hock, I kind of get that in that, look, you have a lot of very important franchises that are stand-alone, right, have a lot of merit, right, on their own basis. But what is the proof point? Where is the evidence to show that bringing together semiconductors and software has created any synergies, right? So how would you kind of grade that part of the business? Because I get that stand-alone, they make sense, but where has been the synergy in bringing them together?

Oh, tremendous. Actually, from the viewpoint of -- if you look back, just look back the last 8 years, let's just pick a long enough term, 6 years, 8 years, even 5 years. You look back at this company, and we've done a lot of acquisitions already since then. You see that we have increased our revenue top line, I would say, basically very nicely. But what you're seeing is while we're doing that, unlike -- contrary to rumors, I would say, we have actually increased R&D spend over that period of time, faster than we increased revenue spend. And over that -- say, take that 8-year spend. And the end result through all this is over that same 8-year spend, our operating profit increased multiple of our revenue increase. And my view, as a proof of [indiscernible] track back the last 8 years, just that 10 years, 8 years, even 6 years, you will see that we have spent more on R&D growth-wise than we have increased revenue, but [ with the ] outcome of it all because of synergies on the platform and our ability to focus resources much more effectively, we have increased operating profitability over a period of time much more. You want more -- some statistics, I may be off a bit. I would say over the last 8 years, also our revenue probably increased about sixfold, sevenfold. Our R&D increased 15-fold, and our operating profit increased something like 30x over that 8-year period. That's the interesting part. It tells a couple of things other than the fact the statistic show by itself, which is -- I'll be blunt. The reason we are able to pick those assets and grow and make -- and basically grow their revenues and profit while spending R&D in the statistics I mentioned is simply that -- I'll be direct. I think we know how to run businesses much better than most of the guys out there. I'll be as direct as I can. That's what gives us the opportunity. That, in a way, is an arbitrage of sorts. And the model we have of focusing verticals, focusing key businesses to do better in the area they're doing and to basically optimize the return out of those investments is phenomenal. It's really good. The other way to look at it is, people talk a lot about software. What have you done on software, this new enterprise software. So let me give you some more statistics on software. If you put Brocade, which we consider really a software appliances -- with appliance, and we did that 2016. And then we did CA and we did Symantec. Today, the revenue from these 3 groups, we call it, we classify and report it as a segment as infrastructure software, is about -- and that's 3.5 years now that we own them. And so we have some cycles under us. We have some experience. Revenue is just over $7 billion a year on this acquisition. And we generally report today in this $5 billion of operating profit. EBITDA, if you want to call it that, $5 billion. And to get that $5 billion, we have invested in acquiring those assets some $33 billion since 2016. But we generate over $5 billion of EBITDA a year. As I said, I've always laid out, we go for at least 10% cash-on-cash return. And in most of our acquisitions, in fact, in all acquisitions over the last 10 years or so, we have exceeded that hurdle.

Got it. So Hock, when I look at the current makeup of the business, so kind of 75-ish percent semiconductors, right, 25-ish percent on the infrastructure software side. If you polish your crystal ball and you look out 3, 4, 5 years from now, what do you think that ratio will become? What do you want that ratio to become?

I -- so I -- to be honest, I don't know. There's no specific driver for -- to make it become something. There is -- as I said before, the overlying -- the overarching investment strategy and the unifying model is that we look for technology products with sustainability and which we could acquire at a fair price that would give us, under our model, under our operating model, a very good return on investment. And the criteria we have always used is at least a 10% cash on cash, that is EBITDA, annual EBITDA to enterprise value that require return on those assets. And we've always used that. And we go when those assets are available. And we are very picky on what those assets are. And even if they don't meet that criteria as we buy companies and we have done that with multiple of those assets, we only keep those -- that had those criteria I articulated. And we spin off, sell off, close those assets, which do not meet those criteria. So Broadcom today, 23 different product lines, assets, all have that quality. And we keep -- keeping our eye -- we keep our eye on that [ ball ]. But in terms of where do we find them, I don't preclude any area. It could still be in semiconductors, though we see less and less of those opportunities, and it could still be in infrastructure software. And in semiconductors, it could be in, as I say, compound semiconductors, where we have legacy strengths or it could be more advanced semiconductor solutions or it could be even infrastructure software. We continue to be on the lookout for those opportunities. So we're not trying to drive to a certain mix. We're not. If what is available, what we're comfortable doing and discipline we have is that -- under our operating model, we have to have that cash-on-cash return.

Got it. And maybe just the last one on this M&A aspect because I think it's very interesting to look at the way you think about the industry, right? And M&A is just kind of one of the aspects of that. But if you look at the semiconductor industry today, do you see at least some interesting targets out there? And if there are, then what is the lens with which you look at them? Is it the profitability? Is it a specific end market or growth? Or is it just the fact that, look, if I do this X, Y and Z, will the regulators even agree to it? Because the last big acquisition you're trying to do in semiconductors might make -- made a lot of sense in hindsight. But obviously, the regulators were against it. So is it that regulatory aspect that keeps you kind of a little bit at arm's length when it comes to looking at semiconductor M&A?

Not really because by definition, I mean, in semiconductors alone, we have about 18 as -- which is to be precise, yes, 18 separate product divisions. And we believe these are in mission-critical areas of application in the semiconductor -- in the component industry. And so regulators, it only comes from overlap -- direct overlapping of those areas. And without trying to sound arrogant about it, I'm saying there's no point acquiring a #2 in any one of those space or #3. It's -- we don't think so. We -- once you're #1 and you invest, right, you stay way ahead. So there isn't -- regulator is not really an issue. And in areas that do not overlap, and we know in semiconductor, there are areas we do not yet go into specifically, we don't want to -- compute, per se, general purpose computing or areas related to that directly or for that matter, memory. We don't do memory. Obviously, there's no overlap at all. In connectivity, I'll be direct. There are very few franchises, I call them, left that would interest us. And we don't look at whether it's growing faster or not. Generally, yes, they do grow faster or not, but the overall industry and growth of that particular vertical doesn't drive us. As I say, it's more sustainability. It can be growing at single digits. We're fine. Just we want to be sure that it's an evolving space, technology-wise. And it's an application -- it's a product that is needed to sustain a particular vertical out there over the long haul. And that's really how we look at it. So it doesn't have to grow. It just -- be there for the next 10 years. And then we really look at, is it actionable? For us, that -- we're not trying to pick up what is a high-growth area because, by the way, high-growth area tends to be still in the state of changing. And that's hard to say -- that makes us pause as -- do we want to buy some -- into a sector when you're not very sure you're going to come out on top. Remember, all our product franchises have to have technology and market leadership.

Got it. Maybe, Hock, moving on to this old topic of supply shortages, right, I think Broadcom was one of the first companies to kind of flag that last year. And you recently reported earnings, so you gave us kind of good color on where you stand, right? You mentioned your lead times have started to kind of stabilize. They're still quite extended, but at least they are not growing any further. Where is Broadcom in that journey? Do you think that you get back to trend line sometimes middle of next year, late next year? And then also importantly, the fact your customers are living through this experience, do you think it changes their behavior going forward once we are getting back to a more normal state? Like would we have learned something from this experience, good or bad, that can impact Broadcom?

Well, that's -- as I said, we are in the midst of one of those imbalance between demand and supply. As I said, semiconductor industry has gone through a multiple of those through the years. This seems to be rather extreme, and I think it's accelerated by a couple of trends, one of which was clearly the pandemic, which drove certain technology areas very strongly because of -- accelerated the adoption of those technologies. And other areas, less so as it relates particularly to enterprise and some other areas. But as it relates to our behavior of working from home, learning from home, that drove those. So there is an accelerated trend. And overlaying on it was this push that started in 2019 and started pushing even more '20, 5G and created some unique demand, especially in content increases in 5G handsets, the fastest conversion from older technology to new technology. And that's just the small handset side. So that was '20. And normally, these cycles do not last that long because -- last maybe a couple of years because there's time for capacity to come in to some extent. But also demand may start to, after initial [ peak ], slow down. The difference here, as we say, is the supply is coming in, for sure. But demand, as we see it right now, has not slowed down. Supply -- my lead times have stretched out and remain stretched out. But part of it was the way we chose to also do it also to truly reflect what's manufacturing cycles, also give us the opportunity to scrub the demand we are seeing in terms of orders. We have orders that is huge backlog right now. But one has to be very -- one, we have to be mindful and be practical that a lot of the demand is -- could be, as you all are worried about, double booking. And so -- and one of the benefits we have is the largest 100 customers in the semiconductor space of Broadcom, which is some $20 billion of revenues or $18 billion of revenue, let's say, or $20 billion, no, sorry, represents 75% of it. So -- and we sell to these guys directly. So it's easier to scrub that through. The others go through distributors, tends to be small players and that, we manage differently. But we have an opportunity to really scrub that and to really try to ship what we feel as true end-user demand and needs. And as far as we can, what we report is as closely as we can make it, what is truly needed. Of course, the existence of buffer inventory, whether it's on our books or on our distributor books or on our end customer books are largely gone. It's almost running a whole supply chain on just in time. And that's what we're seeing right now. And it's -- we're managing that, I think, fairly decently.

Got it. So if, Hock, let's say, a large OEM asks you, when do you think the semiconductor industry will be able to supply me what I need when I need it? Is that in middle of [ 2020 ]?

Well, right now, '22 is -- has filled up pretty much in terms of our order book. And so I would say this cycle is going to last '23. 6 months ago, I would pooh-pooh that idea. Today, seeing -- as I said, some of the trends, some of the tailwinds you talked about earlier that we see in the industry, which is strong renaissance on broadband, the fact that even enterprise is recovering very aggressively, yet hyperscale is still maintaining at a high level of bookings. It's not revenue. Bookings, which goes way as far out as enterprise bookings now, gives me some sense that, no, this may actually run longer than '22. It might actually run to '23 at this point.

Got it. And then, Hock, the last question, as the cycle kind of extends versus what we have seen before, do you anticipate a shift in the kind of markets that perhaps -- whether it's enterprise or maybe even parts of 5G that did not do as well that maybe could start to do better in this extended cycle? Like should investors be prepared for different kind of a regime change or different kind of markets maybe taking over versus what we saw in the last 12 months?

Could you give me more example of what you mean by regime change?

Yes. For example, you mentioned enterprise, right? You have started to see some signs of recovery in the enterprise like as part of the cycle extension, just that there are parts of your end markets that have not yet been very strong, and they are still to kind of come in the second part of the recovery. Do you see any change?

What you're asking is a very interesting question, too, which also is at the back of, I'm sure, a lot of people's mind, which is -- our economy, I'd like to believe, is rebalancing. And you've seen -- I've been asked this question a lot other than when you see is the peak. But the second question, what you're basically trying to say is the -- I've always said that for our model is that the semiconductor industry, long-term compound CAGR growth is about mid-single digit, GDP plus. So this recent set of events, the current set of events we're all sitting on, pandemic, change in it, fundamentally changed this growth rate, long-term growth rate of the semiconductor industry into something different. I think, Vivek, that's probably the -- a way for me to interpret that question. And I'll say this, it's an interesting thing before I jump quickly to answer because it depends on horizon, right? For instance, I'm now kind of perhaps speculating that this industry upturn may last until maybe '23. And you've seen the growth and maybe not as high as what we're seeing today, what I reported last quarter, which is a 20% year-on-year growth on semiconductors alone or at least mid-teens for this year, but still better than mid-single digits for 3 years. So if you take the horizon as 5 years, because of the 3-year strength beginning, you're right. You probably adjust on the mid-single-digit compounded growth rate over 5 years to probably high single digits, mid to high single digits. I believe that. But then -- and I think that's likely going to happen because this cycle may take longer than we think. And so it's not a 1-year thing. It may not be a 2-year thing, if it's a 3-year thing. So a 5-year compounded growth rate might be high single digits. But if you take 10 years, I believe it will probably be mid single -- down to mid-single digits. So it depends on the time horizon you want to take. But also to say -- one of our regime change, well, the first thing that will slow down quickly, my guess, if it starts to hit -- almost starting to hit a wall is always consumer. Consumer just -- it flips up very quickly. It probably be the first to flip down in -- when the cycle rolls over. And I think you'll be -- consumer products, whether it's PCs or it's smartphones. But infrastructure, fortunately for us, Broadcom, which is where most of our businesses, will take longer. It goes up, tends to go up slower though the last few months has shown us otherwise, but tends to. And equally, these are programs that roll out multiple years. Especially, they're pushed out by service providers, less hyper cloud or service providers. They plan, and they roll it out over 2 years, 3 years, takes a while. So we tend to think that -- that's why I tend to think the cycle take longer. And infrastructure will be able to maintain longer. And that's why enterprise popping up in '22 and sustaining perhaps beyond '22. We'll keep that rate of growth year-on-year and above what I consider very long-term trend.

Excellent. Hock, I think with that, we are at the end of our allotted time for this keynote. Really want to thank you for taking the time to share your insights about the industry and about Broadcom. Thanks to everyone who joined. If you have any follow-ups, please feel free to write to me. But with that, let's close this session. And thank you so much, again, Hock, for joining us this morning.

Thank you, Vivek, for inviting me, and thanks to everyone for joining in. Thank you.

Appreciate it. Have a good day.