Broadcom Inc.

Welcome to Broadcom's Inc. Third Quarter Fiscal Year 2024 Financial Results Conference Call. At this time, for opening remarks and introductions, I would like to turn the call over to GU, Head of Investor Relations of Broadcom, Inc.

Thank you, Operator, and good afternoon, everyone. Joining me on today's call are Hop Tan, President and CEO, Kirsten Spears, Chief Financial Officer, and Charlie Kowals, President, Semiconductor Solutions Group. Broadcom distributed a press release and financial tables after the market closed describing our financial performance for the third quarter of fiscal year 2024. If you did not receive a copy, you may obtain the information from the investor section of Broadcom's website at broadcom.com. This conference call is being webcast live and an audio replay of the call can be accessed for one year through the investor section of Broadcom's website. During the prepared comments, Hawk and Kirsten will be providing details of our third quarter fiscal year 2024 results, guidance for our fourth quarter fiscal year 2024, as well as commentary regarding the business environment. We'll take questions after the end of our prepared comments. Please refer to our press release today and our recent filings with the SEC for information on the specific risk factors that could cause our actual results to differ materially from the board-looking statements made on this call. In addition to U.S. GAAP reporting, Broadcom reports certain financial measures on a non-GAAP basis. A reconciliation between GAAP and non-GAAP measures is included in the tables attached to today's press release. Comments made during today's call will primarily refer to our non-GAAP financial results. I will now turn the call over to Hawk.

Thank you, Gee, and thank you, everyone, for joining us today. In our fiscal Q3 2024, consolidated net revenue of $13.1 billion was up 47% year-on-year, and operating profit was up 44% year-on-year. These strong results reflected three key factors. One, AI revenue continues to grow and grow strongly. Two, VMware bookings continue to accelerate. And three, non-AI semiconductor revenue has stabilized. Before I give you more color on our two reporting segments, Let me give you a quick update on guidance. Now we started the year providing annual guidance with quarterly updates as we run the process of integrating VMware. Things are now much more stable and we're in the first, sorry, and we're in the final quarter of 2024. So instead of giving you annual guidance, we now revert to providing the quarterly guidance for Q4. Starting with software, in Q3, infrastructure software segment revenue of $5.8 billion was up 200% year-on-year, driven by $3.8 billion in revenue contribution from VMware. the transformation of the business model of VMware continues to progress very well. In fact, last week, we held a well-attended VMware Explore conference in Las Vegas, our first as a combined company. This event was all about promoting VMware Cloud Foundation, or VCF, which is the full software stack that virtualizes an entire data center and create a private cloud environment on-prem for enterprises. The success of this strategy is reflected in our performance in fiscal Q3. We booked more than 15 million CPU calls of VCF. representing over 80% of the total VMware products we booked during the quarter. And this translates into an annualized booking value, or ABV, as I had described before, of $2.5 billion during Q3, up 32% from the preceding quarter. Meanwhile, we continue to drive down costs in VMware, We brought VMware spending down to $1.3 million in Q3 from $1.6 in Q2. And when we acquired VMware, our target was to deliver adjusted EBITDA of $8.5 billion within three years of the acquisition. We are well on the path to achieving or even exceeding this EBITDA goal in the next fiscal 25. Now turning to semiconductors. In networking, Q3 revenue of $4 billion grew 43% year on year, representing 55% of semiconductor revenue. This was again driven by strong demand from hyperscalers for both AI networking and on custom AI accelerators. As you know, our hyperscale customers continue to scale up and scale out their AI clusters. Custom AI accelerators grow three and a half times year on year. In the fabric, Ethernet switching driven by Tomahawk 5 and Jericho 3 AI grew over four times year on year. while our optical lasers and pin diodes used in optical interconnect grew threefold. Meanwhile, PCI Express switches more than doubled, and we are shipping in volume our industry-leading 5 nanometers, 400 gigabits per second NICs, and 800 gigabits per second DSPs. So now let me give you more color on our networking products, which are not used in AI. As we had indicated last quarter, we believe we hit bottom in Q2. And in Q3, non-AI networking was up actually 17% sequentially, even as it was down 41% year on year. We expect this level of revenue to sustain in Q4 and the year-on-year decline to moderate to 30%. So in adding the strength we continue to see in AI, we expect total networking revenue to grow over 40% year-on-year in Q4. Across enterprise infrastructure, we see the same trend of recovery in server storage. Our Q3 server storage connectivity revenue was $861 million, up 5% sequentially, and down 25% year-on-year. In Q4, we expect server storage revenue to grow mid to high single-digit percent sequentially, even as revenue is expected to be down high single-digit percent year-on-year. Moving on to wireless, Q3 revenue, wireless revenue of $1.7 billion grew 1% year-on-year, representing 23% of semiconductor revenue. And in Q4, reflecting the launch of next-generation devices and our North American customer, We expect wireless revenue to actually grow over 20% sequentially, even as it will be relatively flat year-on-year. On to broadband, Q3 revenue declined 49% year-on-year to $557 million, and represented 8% of semiconductor revenue. Broadband remains weak on a continued pause in telco and service provider spending. And in Q4, we expect broadband to continue to be down over 40% year-on-year, but we do expect that recovery to begin in 2025. Finally, Q3 industrial resales of $164 million declined 31% year-on-year. We believe we are approaching bottom in Q3 as Q4 resales are expected to recover sequentially. Year on year, Q4 industrial resales will still be down approximately 20%. In summary, here are the trends we are seeing in semiconductors. In aggregate, we have reached bottom in our non-AI markets. and we're expecting recovery in Q4. AI demand remains strong and we expect in Q4 AI revenue to grow sequentially 10% to over $3.5 billion. This will translate to AI revenue of $12 billion for fiscal 24 up from our prior guidance of over $11 billion. Putting it all together with software, here's our forecast for Q4. We expect Q4 semiconductor revenue of approximately $8 billion, up 9% year on year. For infrastructure software, we expect revenue to be about $6 billion. So we are guiding Q4 consolidated revenue to be approximately $14 billion, which is up 51% year-on-year. We also expect this will drive Q4 consolidated adjusted EBITDA to achieve approximately 64% of revenue. This Q4 guidance would imply we're raising the outlook for fiscal 2024 revenue to $51.5 billion and adjusted EBITDA for the year to 61.5%. And with that, let me turn the call over to Curtis.

Thank you, Hock. Let me now provide additional detail on our Q3 financial performance. Consolidated revenue was $13.1 billion for the quarter, up 47% from a year ago. Excluding the contribution from VMware, Q3 revenue increased 4% year on year. Gross margins were 77.4% of revenue in the quarter. R&D was $1.5 billion, and consolidated operating expenses were $2.2 billion, up year on year primarily due to the consolidation of VMware. Q3 operating income was $7.9 billion and was up 44% from a year ago with operating margin at 61% of revenue. Excluding transition costs, operating profit of $8 billion was up 45% from a year ago with operating margin of 62% of revenue. Adjusted EBITDA was $8.2 billion, or 63% of revenue. This figure excludes $149 million of depreciation. Now a review of the P&L for our two segments, starting with semis. Revenue for our semiconductor solution segment was $7.3 billion and represented 56% of total revenue in the quarter. This was up 5% year on year. Gross margins for a semiconductor solution segment were approximately 68%, down 270 basis points year-on-year, driven primarily by a higher mix of custom AI accelerators. Operating expenses increased 11% year-on-year to $881 million on increased investment in R&D, resulting in semiconductor operating margins of 56%. Now moving on to infrastructure software. Revenue for infrastructure software was $5.8 billion, up 200% year on year, primarily due to the contribution of VMware, and represented 44% of revenue. Growth margins for infrastructure software were 90% in the quarter, and operating expenses were $1.3 billion in the quarter, resulting in infrastructure software operating margin of 67%. Excluding transition costs, operating margin was 69%. Moving on to cash flow. Free cash flow in the quarter was 4.8 billion and represented 37% of revenues. Excluding cash used for restructuring and integration of 529 million, free cash flows of 5.3 billion were up 14% year on year and represented 41% of revenue. Precash flow as a percentage of revenue has declined from the same quarter a year ago due to higher cash interest expense from debt related to the VMware acquisition and higher cash taxes due to a higher mix of U.S. income and the continued delay in the reenactment of Section 174. We spent $172 million on capital expenditures. Day sales outstanding were 32 days in the quarter in line with a year ago We ended the third quarter with inventory of $1.9 billion, up 3% sequentially. Note that we continue to remain disciplined on how we manage inventory across the ecosystem. We ended the third quarter with $10 billion of cash and $72.3 billion of gross principal debt. During the quarter, we replaced $5 billion of floating rate notes with new fixed senior notes. We use the proceeds from the completed sale of the NWIRS end user computing business to KKR and cash on hand to reduce floating rate debt by an additional $4.2 billion. Following these actions, the weighted average coupon rate and years to maturity of our $53 billion in fixed rate debt is 3.6% and 7.7 years respectively. The weighted average coupon rate and years to maturity of our $19 billion in floating rate debt is 6.7% and 3.1 years respectively. We expect to repay approximately $1.9 billion of fixed rate senior notes due in Q4. Turning to capital allocation, in Q3 we paid stockholders $2.5 billion of cash dividends, which based on a split adjusted quarterly common stock count represented a cash dividend of 52.5 cents per share. For Q4, we are rounding up the quarterly cash dividend to 53 cents per share. In Q3, the split adjusted non-GAAP diluted share count was 4.92 billion in line with expectations. We paid 1.4 billion in withholding taxes due on vesting of employee equity resulting in the elimination of 8.4 million AVGO shares. In Q4, we expect a non-GAAP diluted share count to be approximately 4.91 billion shares. Now on to guidance. Our guidance for Q4 is for consolidated revenue of 14 billion and adjusted EBITDA of approximately 64%. For modeling purposes, we expect consolidated growth margins to be down approximately 100 basis points sequentially on the higher revenue mix of semiconductors and product mix within semiconductors. GAAP net income and cash flows in Q4 are impacted by higher taxes, restructuring, and integration-related cash costs due to the VMware acquisition. As Hawk just discussed, we are resuming quarterly revenue and adjusted EBITDA guidance for fiscal 2025, as fiscal year 24 has been a transition and integration year following the VMware deal close. That concludes my prepared remarks. Operator, please open up the call for questions.

Thank you. To ask a question, please press star 1-1 on your telephone and wait for your name to be announced. To withdraw your question, press star 1-1 again. Due to time restraints, we ask that you please limit yourself to one question. Please stand by while we compile the Q&A roster. And our first question will come from the line of Vivek Arya with Bank of America. Your line is open.

Thanks for taking my question. Just a clarification, Hawk, and then the question. So I think AI revenue roughly 3.1-ish billion in Q3, flat-ish sequentially. What was the mix in terms of compute versus networking? And in the 3.5 billion for Q4, what do you see as that mix? And then as we get into fiscal 25, I realize you're not guiding overall AI, but just how's your general kind of confidence and visibility? Do you think that Broadcom can kind of grow in line or better than the overall AI silicon industry in fiscal 25?

Yeah. Well, as we indicated in the last earnings call, for this past quarter, I think we're talking about two-thirds in... compute and wanted in networking. And we kind of expect Q4 to run the similar trend. And as far to answer your second part, no, we don't guide for yet for fiscal 25, but we do expect fiscal 25 to continue to be strong, to show strong growth on our AI revenue.

Thank you. Thank you. One moment for our next question. And that will come from the line of William Stein with Truist Securities. Your line is open.

Great. Thanks for taking my question. Hock, one of the things that we've picked up from both suppliers and the broader ecosystem in AI, I think we heard this from NVIDIA as well, that there was a shift in their revenue in the quarter somewhat away from cloud service providers towards enterprise. And I wondered if that might potentially have a slowing effect on your revenue outlook in this end market, because your participation is really pretty focused on the cloud customers. I wonder if you're seeing that, if you view it as a challenge, or maybe you have Um, a contrary view. Thank you.

Okay. Well, that's an interesting question in terms of the shift, but see, we do not, uh, focus very much on enterprise AI market, as you know, uh, will is our. products in the ai are largely very much largely focused especially on the ai accelerator or xpu side but even so i'm also in just as much on networking side on hyperscalers on cloud those with large platform and some additional natives what you call big guys we don't deal very much on ai with enterprise so we obviously don't see that trend

Thank you. One moment for our next question. And that will come from the line of Ross Seymour with Deutsche Bank. Your line is open.

Hi, thanks for letting me ask you a question. I wanted to pivot over to the software side of things. Hawk, it seems like, you know, obviously the VMware business had a great fiscal third quarter. It seems like the classic Broadcom software fell off. So I guess the two-part question is, What happened in the classic Broadcom side of things to create that volatility? And are we now kind of reaching that $4 billion base in the fourth quarter that you talked about with VMware? And kind of, if so, what are the puts and takes in the growth rate as we look into the future on that business?

Well, as we indicated, the VMware business continues to book very well. as we convert our customers very much in two ways, one from perpetual to subscription license, but also those subscription licenses are for the full stack of DCF. And that has been very successful, as I indicated, given the high ratio of DCF subscribers, new subscribers that we have achieved. And we see this trend continuing in Q4, very much so, and very likely through into 25. So in terms of directional trend, other than the guidance I'm giving you in Q4, 24, directionally, We continue to see accelerated bookings and, by extension, accelerated growth.

Thank you. One moment for our next question. And that will come from the line of Stacy Rascon with Bernstein Research. Your line is open.

Hi, guys. Thanks for taking my questions. I have two short ones, one on each segment. On semis, you know, the non-AI networking is like more than 50% below where it was running before it rolled off. And clearly the other businesses are also way below their peaks. Is there any reason why those shouldn't event? Is this just cyclical or something else going on? Is there any reason why those shouldn't get back to prior levels once recovery happens? And then on the software side, you know, so the non-VMWare pieces looks like it's back to that 2 billion-ish a quarter level or so that it was at before. Is that just brocade falling off? And is this sort of 2 billion-ish a quarter, is that bottomed as well? Is that the right level we ought to be thinking about for growth for the non-VMWare software business as we go forward from here?

Yeah, on the same side, the answer is very simple. We have, as you all know, we've gone through your typical down cycle of semiconductors, and I'm referring particularly to non-AI, and we have talked about that before many times. We've gone through a down cycle, and as the ecosystem, as many of our customers, but the broader ecosystems work on and adjustment in the inventory levels at all stages in the supply chain. And we are not immune from it, obviously, as we try to insulate ourselves from it as much as possible. We've gone through it, and the indications we have seen very clearly is we have, in fact, passed through the bottom. The best indicator is the bookings we are receiving. In non-AAI, our bookings in Q3 of non-AAI semiconductor demand is up 20%. And so that tells us we are well on the way to a recovery. Now, by end markets, as I indicated, the level of the amount of recovery, the timing of recovery somewhat varies. But we're seeing largely on enterprise, enterprise data center, enterprise IT spending, we're past the bottom. And Q3 was, in fact, sequentially a recovery from the bottom of, we believe, Q2 or Q1 this fiscal year. And we'll see Q4 continuing that recovery, and obviously, in our view, into 25. in terms of the cycle. Broadband, we are not seeing it yet in terms of the bottom, but we see that as close to bottom in the sense that here, again, bookings are up from where it used to be. And so we are very, very clear in our thinking that broadly we have, as a whole, non-AI semiconductors. We've gone through the down cycle. It's on an uptick. And like all previous cycles, my sense to you, Stacey, is we will get up back to the level we used to be. There's no reason at all why it doesn't. And given the rate of bookings, it will go. I dare say even put a thought in your mind that as AI permits enterprise, enterprises all across and digital natives. You need to upgrade service. You need to upgrade storage. You need to upgrade networking, connectivity across the entire ecosystem. And if anything else, we are headed, we could be headed for upcycle timing precisely when, we're not sure, but an upcycle that could even meet or even surpass what our previous upcycles would be. Simply because the amount of bandwidth you need, the amount to manage a store, manage all those workloads that come out of AI, would just put a need to refresh and upgrade hardware. So that's my two cents worth on where we're headed from this down cycle. So my belief is 24 was the lowest point where they are up there. That's part of the reasons we are stating it very clearly here. On the software side, your question, no, I think we have reached a level of stability that puts and takes. Brocade is one of those that goes up and down very volatile, and that's largely it. But on the non-VMware revenue, on the software revenue, I think we've reached a level of very clear stability. And what we are looking towards more is how VMware picks up over the next year, year and a half.

Got it, that's helpful. Thank you, Hawk.

Thank you. One moment for our next question.

Hello.

And that will come from the line of Ben Rietzes with Mellius Research. Your line is open.

Hey, thanks a lot for the question. Hawk, I wanted to ask you about semiconductors, your AI revenue. If you could just clarify some of your comments. Was the third quarter 3.1-ish in line with your expectations, and was anything weaker than expected? And then with the sequential growth to 3.5, where are you expecting that to come from? And then if you don't mind, you said next year AI revenue should grow quite a bit. I was just wondering if that was due to any additional customers within your hyperscaler and consumer internet portfolio. Thanks.

Well, our number in the third quarter is pretty much in line with what we expect AI revenue to be. And our revenue in Q4 was what forecast for Q4 is what's giving us the basis to a large extent to step up our guidance for AI revenue for the full year to over $12 billion. So if nothing else, that continues to indicate, I hope, to us that next year the trend will continue to be strong. And again, it's all largely hyperscalers, cloud, and digital natives. And it's again a mix of AI accelerators and networking. And it's also largely based on backlog we have in place for that. Beyond that, and it shows a growth. Beyond that, no, we're not guiding you beyond the backlog we have. So I kind of answer your question indirectly on do I have any more customers? We shall see.

Okay, thank you, Hock.

Thank you. One moment for our next question. And that will come from the line of Carl Ackerman with BNP Paribas. Your line is open.

Yes, thank you. Kirsten, I was hoping you could speak to the relocation of IP back to the U.S. that is causing a $4.5 billion tax liability. Historically, Broadcom has redomiciled ahead of a pending transaction, and I'm getting questions from investors saying, if this action may relate to any asset sales as the company seeks to pay down debt. So if you could clarify that, that would be helpful. Thank you.

Yeah, no, it was just the timing of when we chose to do it this time. And, no, it doesn't have anything to do with that. It's just we relocated the IP, and that caused the $4 billion charge. The offset to that is a deferred tax liability. So think of that as non-cash, very little cash impact to that.

Thank you. One moment for our next question. And that will come from the line of Timothy Arcuri with UBS. Your line is open.

Thanks a lot. Hawk, I wanted to ask about the growth rate in your AI revenue versus what we're seeing on the GPU side. Your AI revenue grew in the same zip code this year as what the GPU computes growing. And you did say that it would be up next year, but your main customer is ramping a new version of their custom ASIC next year, and there's some thought that they might shift some of their purchasing back to GPUs next year. So do you think that the growth of your AI revenue should still approximately track how much GPU compute is going to grow next year? If you can give us any qualitative or quantitative thoughts there, that'd be great. Thank you.

That's been me. Hey, Dean. I think we had some communication gaps here. Could you repeat the question?

Yeah. So the question, Hawk, really is around the growth rate of your AI revenue versus what we're seeing on the GPU side, because this year you grew about the same as what GPU computes growing. And the question is, is there anything happening next year that would change that equation so that your growth rate of your AI revenue would be materially different than what GPU computes growing next year?

You know, that's a very difficult question for me to answer because it comes in two parts, right? In terms of GPU growth, you should ask the guys who does merchant GPU or GPU, which is obviously NVIDIA and AFD. And I don't see, I don't play in the enterprise market at all. See, that's part of their market I don't see. Having said that, they do both play a somewhat in the hyperscalers where I'm totally focused on doing. So that's really very, and that's really no connection one with the other. It is indirect, but it not suffice for me to say long term, I'm saying that thoughtfully, long term, the large hyperscalers, few and large hyperscalers with very large platforms, huge consumer platforms, subscriber base, have the entire model predicated on running a lot of large language models, a lot of AI requirements, workloads out there. And it will drive, in a matter of time, towards creating as much as possible their own compute silicon, their own custom accelerators as a matter of time. And we are in the midst of seeing that transition. It may take a few years for that to happen. So that is on a different trajectory, a different path. And I'm in that path of doing that, enabling custom accelerators. I'm in that path. I'm not in the path of, in the meantime, a different trajectory of enabling enterprises to do AI on their own workloads. That's more the merchant guys. Some of the merchant guys are obviously also in the AI, in the hyperscalers today, but that's a process, obviously, of a transition going on. So one doesn't really connect to the other team in that regard. But I would likely say, obviously, as the transition occurs, we have a good tailwind in the business model we have of providing accelerators and networking to the AI data centers of those large hyperscales.

Right, Hock. Okay. Thank you so much.

One moment for our next question. And that will come from the line of Harsh Kumar with Piper Sandler. Your line is open.

Yeah, I was curious about the profitability of VMware. Historically, your software businesses have had operating margins greater than 70%. VMware, I know, is newer and you're doing things a little different. You're keeping more customers than you historically have kept. But I was curious if you see a similar profile as the rest of your software businesses for VMware after you're done with all the cuts and everything?

Well, I'll let you draw your own conclusion, Ash, but I was a pain to lay out, as you probably heard. In Q3, our revenue from VMware was $3.8 billion, and our operating expenses is $1.3 billion. And you can pretty quickly figure out where we're headed in terms of operating margin and as I indicated EBITDA margin. And Q4 will continue the trajectory of revenue continuing to grow and expenses still dropping even as it starts to stabilize but continue to reduce.

Thank you, Hawk.

One moment for our next question. And that will come from the line of CJ Muse with Cantor Fitzgerald. Your line is open.

Yeah, good afternoon. Thank you for taking the question. I wanted to focus on software gross margin. So when you close the acquisition of VMware, we ticked lower from low 90s to kind of high 80s. And we're now pushing a bit higher into July. And curious, as we kind of get to that $4 billion threshold and you've kind of indicated higher in fiscal 25, how should we think about the gross margin trajectory overall for software?

Well, you know, for us, software gross margin is actually redirect. It's not that relevant. You know that, right? Software is, unless I'm running SaaS big time. Now, a lot of our products are subscription, but they're not SaaS. We have some of our products on CES cloud-based, but most of them are not. And our gross margin will be around 90% at least.

And one moment for our next question.

And that will come from the line of Chris Caso with Wolf Research. Your line is open.

Yes, thank you. Good evening. I wonder if you could speak to the custom AI revenue and perhaps the contribution from some of the other customers aside from that largest customer. You know, how meaningful are the other customers in that segment? And, you know, what do you expect into next year as some of those newer projects start to ramp?

I know, we're dancing around the thing. As I indicated, we have three customers now going on, and all three of them are meaningful. Otherwise, we won't call them customers as the criteria we have used. Until we get meaningful shipments out to them on AI accelerators, we do not really consider that as a customer. Simply because it's a new, this is an emerging trend. It's not an easy product to deploy for any customer. And so we do not consider proof of concepts as production volume. These are all production accelerators deployed in data centers, AI data centers of those three customers.

One moment for our next question. And that will come from the line of Christopher Rowland with Susquehanna.

Your line is open.

Hi. Thanks for the question. My question is actually on storage. And, Hawk, you bought Seagate's hard disk drive SOC assets earlier in the year. Can you talk about what you actually bought there? what it means in terms of economics for your company and whether this accelerates your storage business over the next few years. Thanks. This is more of a partnership than anything else.

Basically, what we essentially created in that transaction was To begin with, we actually believe long-term in the sustainability of hard disk drive media as a great long-term sustainable storage alternative or storage medium for those hyperscales. It makes sense. One would like to think eventually everything goes to flash. Don't think so. hard disk drive storage will still be meaningful. And the technology, which is most interesting for us, has a lot of ways to go. As hard disk drives goes on to from where it is today, which is, you know, 22, 23, 24, you know, terabytes, to going to 30, 40, and even 50 terabytes. A lot of technology along the way, and a lot of that resides in silicon. So what we're doing in effect is a collaboration more than anything else, though it's structured obviously as a purchase of intellectual property, but we're also taking engineers, designers, combining it with the designers we have, and basically enabling Seagate and eventually the entire industry to continue a roadmap that goes towards 50 terabytes. That's our ambition, that's our vision, and to be able to do that within five years or less. So that's pretty much what it is. It's a statement of our belief that nearline hard disk drive storage will sustained very well over the next five years, if not longer.

Thank you.

One moment for our next question. And that will come from the line of Aaron Rakers with Wells Fargo. Your line is open.

Yeah, thanks for taking the question. Kind of thinking strategically as we look forward ahead to NVIDIA's Blackwell product cycle, there's been some indications that possibly Broadcom has an opportunity to participate more deeply in the optical side of that product platform for NVIDIA. I'm curious, do you see that as an opportunity relative to prior generations of NVIDIA just to deepen a participation or just to participate in general in kind of the areas of VSPs and maybe other things related to the Blackwell cycle from NVIDIA? Thank you.

that's an interesting question and got a simple answer i'm not really participating in it that's invidious roadmap uh and i'm really not directly uh in that kind of market in that kind of uh product roadmap that's invidious product roadmap in terms of blackwell impressive product on the way to coming out uh now in terms of base technology we developed of course it could be used, it could be applied, and we are very happy to share that as it may be useful to enable Blackwell to be part of that, whether it's on the optical component side, which is what you're referring to, or even the DSP side in terms of providing the interconnects to enable clusters of Blackwell to be built. That's as far as our engagement in that. We're happy to be part of that ecosystem, as I said, but directly we're not in that market, as you know.

Yep.

One moment for our next question. And that will come from the line of Joe Moore with Morgan Stanley. Your line is open.

Great. Thank you. I wonder, Hawk, if you could talk about your thoughts on further M&A You know, is that still on your radar down the road? And, you know, if you did, would it be still software-focused or any possibility of semiconductors becoming interesting to you again?

Joe, that's a beautiful question. I'll tell you this bluntly so that you're not disappointed. Right now, I'm having my hands really full and enjoying myself doing it on really turning, transforming the business model of EMOS. It's a great experience, and you're feeling great about it when you're doing it. And when you're doing it, it's pretty much running way beyond expectation, as we indicated in that slide. So, no, I'm very focused on getting VMware as it continues to accelerate in getting private cloud deployed in the largest enterprises in the world. And you know what? I've got another year, two years to go. to make that transformation totally complete.

Very clear. Thank you.

One moment for our next question. And that will come from the line of Harlan Sir with J.P. Morgan. Your line is open.

Good afternoon. Thanks for taking my question. You know, Hoff, last quarter you talked about an acceleration in R&D investments by your AI customers, and you talked about your follow-on wins for their next generation XPU ASIC programs. It also looks like they're trying to accelerate their deployment of their TPUs, XPUs, and networking into their data centers here in the second half of the year. We know that on AI accelerators specifically, supply is quite tight given the co-op packaging and HBM memory constraints. So has the team seen upside orders in demand for XPUs and networking? here in the second half, have you been able to meet that upside demand or is the team somewhat supply constrained? I guess in other words, is AI demand greater than your supply here in the second half of the year?

Yeah, no, we continue to see all this. We continue to see upsides and you're right in the pattern of that behavior that's going It's, you know, as our customers, these are hyperscalers, trying to deploy more and more capacity of AI data centers, in AI data centers, and you start to hear them talk in terms of power. They don't even talk in terms of how many XPU or GPU classes they talk in. you know, 500 megawatt, one gigawatt. Not yet, but people are dreaming that. So we are, as they get this enabled, we're getting pull-ins, we're getting upsides. And I expect that to happen a lot more in 2025. We're not putting that in any guidance or indication we're giving you, but I'm pulling what you said exactly right on. We do expect to see upsides, as we've been seeing, uh recently we continue to see that probably going forward over the next 12 months especially in related to xpus getting deployed and create getting infrastructure available and rushing to deploy them we see on quite a bit of that have you been able to meet that upside or are you somewhat limited by supply constraints we can meet those upsides perfect

Thank you, Hawk.

Thanks.

One moment for our next question. And that will come from the line of Edward Snyder with Tartar Equity Research. Your line is open.

Thank you very much. Hawk, that was a perfect segue into my question. You've said in the past calls that you thought that AI compute would move away from ASICs and go to merchant market, but it looks like the trend kind of heading the other way are you still the opinion that that's going to be the long-term trend of this and secondly as you just pointed out um power is becoming the defining factor for deployment um with all the big guys at this point given the performance of uh performance per watt of the asics over gpus which is superior to gpus why shouldn't we see more of these guys moving to to uh custom asics i know it takes a long time it takes a lot of funding etc but especially as the enterprise starts getting more involved with this, there are going to be some applications that are kind of standard across some of the enterprises. Wouldn't we even see some of the bigger, like AWS moved to a custom silicon for a specific workload? So basically the overall trend in ASICs in AI. Thanks.

Okay. Ed, did I hear you right to say at the beginning, maybe, that you meant that there's a trend towards asic or xpu from general purpose gp right yep you're right and you correct in pointing out to me that hey i used to think that general purpose merchant silicon will win at the end of the day well based on history of semiconductors mostly so far general purpose merchant silicon tends to win but like you i flipped in my view and i did that by the way six last quarter maybe even six months ago but then nonetheless catching up is good and i i actually think so because i actually think there are two markets here on ai accelerators there's one market for enterprises of the world And none of these enterprises are incapable nor have the financial resources or interest to create the silicon, the custom silicon, nor the large language models and the software going maybe to be able to run those AI workloads on custom silicon. It's too much. And there's no return for them to do it because it's just too expensive to do it. But there are those few cloud guys, hyperscalers, with the scale of the platform and the financial wherewithal for them to make it totally rational, economically rational, to create their own platform. and custom accelerators because it's all, right now, I'm not going to, I'm not trying to over-emphasize it. It's all about compute engines. It's all about, especially training those large language models and enabling it on your platform. It's all about constraint to a large part about GPUs. Seriously, it's getting to a point where GPUs are more important than engineers. It is some of these hyperscalers in terms of how they think. Those GPUs are much more, or XPUs, are much more important. And if that's the case, what better thing to do than bring it under the control of your own destiny by creating your own custom silicon accelerators. And that's one. I'm seeing all of them do. It's just doing it at different rates, and they're starting at different times. But they all have started, and obviously it takes time to get there, but a lot of them, there are a lot of learning in the process versus once. The biggest guy of them who has done it longest have been doing it for seven years. Others are trying to catch up, and it takes time. I'm not saying it will take seven years. I think it will be accelerated, but it will still take some time, step by step, to get there. But those few hyperscalers, platform guys, will create their own if they haven't already done it and start to train them on their large language models. And that's Yeah, you're right. They will all go in that direction totally into ASIC or we call it XPUs, custom silicates. Meanwhile, there's still a market in enterprise for merchant silicates.

Right, but that basically suggests that you're on the early part of your curve where I'm not trying to call GPUs or whatever, but you could be getting to something closer to the peak of the GPU market just because everything, right, besides the cost expense and as you're spending all this money and you're paying all this money for power, the ASICs become more and more attractive. So the curves are going to look different, right?

It's an accelerating curve. It may take longer than we all want it to happen, but definitely accelerating. Because the size of the demand from those hyperscalers will totally rival that in the enterprise.

Thank you. And that is all the time we have for our question and answer session. I would now like to turn the call over to GU for any closing remarks.

Thank you, operator. This quarter, Broadcom will be presenting at the Goldman Sachs Communicopia and Technology Conference on Wednesday, September 11th in San Francisco. Broadcom currently plans to report its earnings for the fourth quarter and fiscal year 2024 after the close of market on Thursday, December 12th, 2024. A public webcast of Broadcom's earnings conference call will follow at 2 p.m. Pacific. That will conclude our earnings call today. Thank you all for joining. Operator, you may end the call.

This concludes today's program. Thank you all for participating.

You may now disconnect. you Thank you. I'm

Thank you. you you

Welcome to Broadcom's Inc. Third Quarter Fiscal Year 2024 Financial Results Conference Call. At this time, for opening remarks and introductions, I would like to turn the call over to GU, Head of Investor Relations of Broadcom, Inc.

Thank you, Operator, and good afternoon, everyone. Joining me on today's call are Hop Tan, President and CEO, Kirsten Spears, Chief Financial Officer, and Charlie Kawaz, President, Semiconductor Solutions Group. Broadcom distributed a press release and financial tables after the market closed describing our financial performance for the third quarter of fiscal year 2024. If you did not receive a copy, you may obtain the information from the investor section of Broadcom's website at broadcom.com. This conference call is being webcast live and an audio replay of the call can be accessed for one year through the investor section of Broadcom's website. During the prepared comments, Hawk and Kirsten will be providing details of our third quarter fiscal year 2024 results, guidance for our fourth quarter fiscal year 2024, as well as commentary regarding the business environment. We'll take questions after the end of our prepared comments. Please refer to our press release today and our recent filings with the SEC for information on the specific risk factors that could cause our actual results to differ materially from the board-looking statements made on this call. In addition to U.S. GAAP reporting, Broadcom reports certain financial measures on a non-GAAP basis. A reconciliation between GAAP and non-GAAP measures is included in the tables attached to today's press release. Comments made during today's call will primarily refer to our non-GAAP financial results. I will now turn the call over to Hawk.

Thank you, Gee, and thank you everyone for joining us today. In our fiscal Q3 2024, consolidated net revenue of $13.1 billion was up 47% year-on-year, and operating profit was up 44% year-on-year. These strong results reflected three key factors. One, AI revenue continues to grow and grow strongly. Two, VMware bookings continue to accelerate. And three, non-AI semiconductor revenue has stabilized. Before I give you more color on our two reporting segments, Let me give you a quick update on guidance. Now we started the year providing annual guidance with quarterly updates as we run the process of integrating VMware. Things are now much more stable and we're in the first, sorry, and we're in the final quarter of 2024. So instead of giving you annual guidance, we now revert to providing the quarterly guidance for Q4. Starting with software, in Q3, infrastructure software segment revenue of $5.8 billion was up 200% year-on-year, driven by $3.8 billion in revenue contribution from VMware. the transformation of the business model of VMware continues to progress very well. In fact, last week, we held a well-attended VMware Explore conference in Las Vegas, our first as a combined company. This event was all about promoting VMware Cloud Foundation, or VCF, which is the full software stack that virtualizes an entire data center and create a private cloud environment on-prem for enterprises. The success of this strategy is reflected in our performance in fiscal Q3. We booked more than 15 million CPU costs of VCF. representing over 80% of the total VMware products we booked during the quarter. And this translates into an annualized booking value, or ABV, as I had described before, of $2.5 billion during Q3, up 32% from the preceding quarter. Meanwhile, we continue to drive down costs in VMware, We brought VMware spending down to $1.3 million in Q3 from $1.6 in Q2. And when we acquired VMware, our target was to deliver adjusted EBITDA of $8.5 billion within three years of the acquisition. We are well on the path to achieving or even exceeding this EBITDA goal in the next fiscal 25. Now turning to semiconductors. In networking, Q3 revenue of $4 billion grew 43% year on year, representing 55% of semiconductor revenue. This was again driven by strong demand from hyperscalers for both AI networking and on custom AI accelerators. As you know, our hyperscale customers continue to scale up and scale out their AI clusters. Custom AI accelerators grow three and a half times year on year. In the fabric, Ethernet switching driven by Tomahawk 5 and Jericho 3 AI grew over four times year on year. while our optical lasers and pin diodes used in optical interconnect grew threefold. Meanwhile, PCI Express switches more than doubled, and we're shipping in volume our industry-leading 5 nanometers, 400 gigabits per second NICs, and 800 gigabits per second DSPs. So now let me give you more color on our networking products, which are not used in AI. As we had indicated last quarter, we believe we hit bottom in Q2. And in Q3, non-AI networking was up actually 17% sequentially, even as it was down 41% year on year. We expect this level of revenue to sustain in Q4 and the year-on-year decline to moderate to 30%. So in adding the strength we continue to see in AI, we expect total networking revenue to grow over 40% year-on-year in Q4. Across enterprise infrastructure, we see the same trend of recovery in server storage. Our Q3 server storage connectivity revenue was $861 million, up 5% sequentially, and down 25% year-on-year. In Q4, we expect server storage revenue to grow mid to high single-digit percent sequentially, even as revenue is expected to be down high single-digit percent year-on-year. Moving on to wireless, Q3 revenue, wireless revenue of $1.7 billion grew 1% year-on-year, representing 23% of semiconductor revenue. And in Q4, reflecting the launch of next-generation devices and our North American customer, We expect wireless revenue to actually grow over 20% sequentially, even as it will be relatively flat year-on-year. Onto broadband, Q3 revenue declined 49% year-on-year to $557 million, and represented 8% of semiconductor revenue. Broadband remains weak on a continued pause in telco and service provider spending. And in Q4, we expect broadband to continue to be down over 40% year-on-year, but we do expect that recovery to begin in 2025. Finally, Q3 industrial resales of $164 million declined 31% year-on-year. We believe we are approaching bottom in Q3 as Q4 resales are expected to recover sequentially. Year on year, Q4 industrial resales will still be down approximately 20%. In summary, here are the trends we are seeing in semiconductors. In aggregate, we have reached bottom in our non-AI markets. and we're expecting a recovery in Q4. AI demand remains strong, and we expect in Q4 AI revenue to grow sequentially 10% to over $3.5 billion. This will translate to AI revenue of $12 billion for fiscal 24, up from our prior guidance of over $11 billion. Putting it all together with software is our focus for Q4. We expect Q4 semiconductor revenue of approximately $8 billion, up 9% year-on-year. For infrastructure software, we expect revenue to be about $6 billion. So we are guiding Q4 consolidated revenue to be approximately $14 billion, which is up 51% year on year. We also expect this will drive Q4 consolidated adjusted EBITDA to achieve approximately 64% of revenue. This Q4 guidance would imply we're raising the outlook for fiscal 2024 revenue to $51.5 billion and adjusted EBITDA for the year to 61.5%. And with that, let me turn the call over to Kurt.

Thank you, Hock. Let me now provide additional detail on our Q3 financial performance. Consolidated revenue was 13.1 billion for the quarter, up 47% from a year ago. Excluding the contribution from VMware, Q3 revenue increased 4% year on year. Gross margins were 77.4% of revenue in the quarter. R&D was 1.5 billion, and consolidated operating expenses were 2.2 billion, up year-on-year primarily due to the consolidation of VMware. Q3 operating income was $7.9 billion and was up 44% from a year ago with operating margin at 61% of revenue. Excluding transition costs, operating profit of $8 billion was up 45% from a year ago with operating margin of 62% of revenue. Adjusted EBITDA was $8.2 billion, or 63% of revenue. This figure excludes $149 million of depreciation. Now a review of the P&L for our two segments, starting with semis. Revenue for our semiconductor solution segment was $7.3 billion and represented 56% of total revenue in the quarter. This was up 5% year on year. Gross margins for a semiconductor solution segment were approximately 68%, down 270 basis points year-on-year, driven primarily by a higher mix of custom AI accelerators. Operating expenses increased 11% year-on-year to $881 million on increased investment in R&D, resulting in semiconductor operating margins of 56%. Now moving on to infrastructure software. Revenue for infrastructure software was $5.8 billion of 200% year-on-year, primarily due to the contribution of VMware, and represented 44% of revenue. Growth margins for infrastructure software were 90% in the quarter, and operating expenses were $1.3 billion in the quarter, resulting in infrastructure software operating margin of 67%. Excluding transition costs, operating margin was 69%. Moving on to cash flow. Free cash flow in the quarter was $4.8 billion and represented 37% of revenues. Excluding cash used for restructuring and integration of $529 million, free cash flows of $5.3 billion were up 14% year-on-year and represented 41% of revenue. Precash flow as a percentage of revenue has declined from the same quarter a year ago due to higher cash interest expense from debt related to the VMware acquisition and higher cash taxes due to a higher mix of U.S. income and the continued delay in the reenactment of Section 174. We spent $172 million on capital expenditures. Day sales outstanding were 32 days in the quarter, in line with a year ago, We ended the third quarter with inventory of $1.9 billion, up 3% sequentially. Note that we continue to remain disciplined on how we manage inventory across the ecosystem. We ended the third quarter with $10 billion of cash and $72.3 billion of gross principal debt. During the quarter, we replaced $5 billion of floating rate notes with new fixed senior notes. We use the proceeds from the completed sale of the NWERS end user computing business to KKR and cash on hand to reduce floating rate debt by an additional $4.2 billion. Following these actions, the weighted average coupon rate and years to maturity of our $53 billion in fixed rate debt is 3.6% and 7.7 years respectively. The weighted average coupon rate and years to maturity of our $19 billion in floating rate debt is 6.7% and 3.1 years respectively. We expect to repay approximately $1.9 billion of fixed rate senior notes due in Q4. Turning to capital allocation, in Q3 we paid stockholders $2.5 billion of cash dividends which based on a split adjusted quarterly common stock count represented a cash dividend of 52.5 cents per share. For Q4, we are rounding up the quarterly cash dividend to 53 cents per share. In Q3, the split adjusted non-GAAP diluted share count was 4.92 billion in line with expectations. We paid 1.4 billion in withholding taxes due on vesting of employee equity resulting in the elimination of 8.4 million AVGO shares. In Q4, we expect a non-GAAP diluted share count to be approximately 4.91 billion shares. Now on to guidance. Our guidance for Q4 is for consolidated revenue of 14 billion and adjusted EBITDA of approximately 64%. For modeling purposes, we expect consolidated gross margins to be down approximately 100 basis points sequentially on the higher revenue mix of semiconductors and product mix within semiconductors. Gap net income and cash flows in Q4 are impacted by higher taxes, restructuring, and integration-related cash costs due to the VMware acquisition. As Hawk just discussed, we are resuming quarterly revenue and adjusting EBITDA guidance for fiscal 2025, as fiscal year 24 has been a transition and integration year following the VMware deal close. That concludes my prepared remarks. Operator, please open up the call for questions.

Thank you. To ask a question, please press star 11 on your telephone and wait for your name to be announced. To withdraw your question, press star 11 again. Due to time restraints, we ask that you please limit yourself to one question. Please stand by while we compile the Q&A roster. And our first question will come from the line of Vivek Arya with Bank of America. Your line is open.

Thanks for taking my question. Just a clarification, Hawk, and then the question. So I think AI revenue roughly 3.1-ish billion in Q3, flat-ish sequentially. What was the mix in terms of compute versus networking? And then the 3.5 billion for Q4, what do you see as that mix? And then as we get into fiscal 25, I realize you're not guiding overall AI, but just how's your general kind of confidence and visibility? Do you think that Broadcom can kind of grow in line or better than the overall AI silicon industry in fiscal 25? Yes.

Well, as we indicated in the last earnings call, for this past quarter, I think we're talking about two-thirds in... compute and wanted in networking. And we kind of expect Q4 to run the similar trend. And as far to answer your second part, no, we don't guide for yet for fiscal 25, but we do expect fiscal 25 to continue to be strong, to show strong growth on our AI revenue.

Thank you. Thank you. One moment for our next question. And that will come from the line of William Stein with Truist Securities. Your line is open.

Great. Thanks for taking my question. Hock, one of the things that we've picked up from both suppliers and the broader ecosystem in AI, I think we heard this from NVIDIA as well, that there was a shift in their revenue in the quarter somewhat away from cloud service providers towards enterprise. And I wondered if that might potentially have a slowing effect on your revenue outlook in this end market, because your participation is really pretty focused on the cloud customers. I wonder if you're seeing that, if you view it as a challenge, or maybe you have Um, a contrary view. Thank you.

Okay. Well, that's an interesting question in terms of the shift, but see, we do not, uh, focus very much on enterprise AI market, as you know, uh, will is our products in AI are largely, very much largely focused, especially on the AI accelerator or XPU side. But even, so more so, just as much on networking side. On hyperscalers, on cloud, those with large platform, and some digital natives, what you call, big guys. We don't deal very much on AI with enterprise. So we obviously don't see that trend.

Thank you. One moment for our next question. And that will come from the line of Ross Seymour with Deutsche Bank. Your line is open.

Hi, thanks for letting me ask you a question. I wanted to pivot over to the software side of things. Hawk, it seems like, you know, obviously the VMware business had a great fiscal third quarter. It seems like the classic Broadcom software fell off. So I guess the two-part question is, What happened in the classic Broadcom side of things to create that volatility? And are we now kind of reaching that $4 billion base in the fourth quarter that you talked about with VMware? And kind of, if so, what are the puts and takes and the growth rate as we look into the future on that business?

Well, as we indicated, the VMware business continues to book very well. as we convert our customers very much in two ways, one from perpetual to subscription license, but also those subscription licenses are for the full stack of VCF. And that has been very successful, as I indicated, given the high ratio of VCF subscribers and new subscribers that we have achieved and we see this trend continuing in q4 and very much so and possibly very lightly through into 25 so in terms of directional trend other than the indication i'm giving you then the guidance i'm giving you 24 in q4 24 directionally We continue to see accelerated bookings and, by extension, accelerated growth.

Thank you. One moment for our next question. And that will come from the line of Stacy Rescon with Bernstein Research. Your line is open.

Hi, guys. Thanks for taking my questions. I have two short ones, one on each segment on semis. You know, the not AI networking is like more than 50% below where it was running before it rolled off. And clearly the other businesses are also way below their peaks. Is there any reason why those shouldn't event? And this is this just cyclical or something else going on? Is there any reason why those shouldn't get back to prior levels once recovery happens? And then on the software side, you know, so the non-VMWare pieces looks like it's back to that 2 billion-ish a quarter level or so that it was at before. Is that just brocade falling off? And is this sort of 2 billion-ish a quarter, is that bottomed as well? Is that the right level we ought to be thinking about for growth for the non-VMWare software business as we go forward from here?

Yeah, on the same side, the answer is very simple. We have, as you all know, we've gone through your typical down cycle of semiconductors, and I'm referring particularly to non-AI, and we have talked about that before many times. We've gone through a down cycle, and as the ecosystem, as many of our customers, but the broader ecosystems work on an adjustment in the inventory levels at all stages in the supply chain. And we are not immune from it, obviously, as we try to insulate ourselves from it as much as possible. We've gone through it, and the indications we have seen very clearly is we have, in fact, passed through the bottom. The best indicator is the bookings we are receiving. In non-AI, our bookings in Q3 of non-AI semiconductor demand is up 20%. And so that tells us we are well on the way to a recovery. Now, by end markets, as I indicated, the level of the amount of recovery, the timing of recovery somewhat varies. But we're seeing largely on enterprise, enterprise data center, enterprise IT spending, we're past the bottom. And Q3 was, in fact, sequentially a recovery from the bottom of, we believe, Q2 or Q1 this fiscal year. And we'll see Q4 continuing that recovery, and obviously, in our view, into 25. in terms of the cycle. Broadband, we are not seeing it yet in terms of the bottom, but we see that as close to bottom in the sense that here again bookings are up from where it used to be. And so we are very, very clear in our thinking that broadly we have, as a whole, non-AI semiconductors, we've gone through the down cycle. It's on an uptake. And like all previous cycles, my sense to you, Stacey, is we will get up back to the level we used to be. There's no reason at all why it doesn't. And given the rate of bookings, it will go. I dare say even put a thought in your mind that as AI permits enterprise, enterprises all across and digital natives. You need to upgrade service. You need to upgrade storage. You need to upgrade networking, connectivity across the entire ecosystem. And if anything else, we are headed, we could be headed for upcycle timing precisely when, we're not sure, but an upcycle that could even meet or even surpass what our previous upcycles would be. Simply because the amount of bandwidth you need, the amount to manage a store, manage all those workloads that come out of AI, would just put a need to refresh and upgrade hardware. So that's my two cents worth on where we're headed from this down cycle. So my belief is 24 was the lowest point where they are uptake. That's part of the reasons we are stating it very clearly here. On the software side, your question, no, I think we have reached a level of stability that puts and takes. Brocade is one of those that goes up and down very volatile, and that's largely it. But on the non-VMware revenue, on the software revenue, I think we've reached a level of very clear stability. And what we are looking towards more is how VMware picks up over the next year, year and a half.

Got it, that's helpful. Thank you, Hawk.

Thank you. One moment for our next question.

Hello.

And that will come from the line of Ben Rietzes with Mellius Research. Your line is open.

Hey, thanks a lot for the question. Hawk, I wanted to ask you about semiconductors, your AI revenue. If you could just clarify some of your comments. Was the third quarter 3.1-ish in line with your expectations, and was anything weaker than expected? And then with the sequential growth to 3.5, where are you expecting that to come from? And then if you don't mind, you said next year AI revenue should grow quite a bit. I was just wondering if that was due to any additional customers within your hyperscaler and consumer internet portfolio. Thanks.

Well, our number in the third quarter is pretty much in line with what we expect AI revenue to be. And our revenue in Q4 was what forecast for Q4 is what's giving us the basis to a large extent to step up our guidance for AI revenue for the full year to over $12 billion. So if nothing else, that continues to indicate, I hope to us, that next year the trend will continue to be strong. And again, it's all largely hyperscalers, cloud, and digital natives. And it's again a mix of AI accelerators and networking. And it's also largely based on backlog we have. in place for that. Beyond that, and it shows the growth. Beyond that, no, we're not guiding you beyond the backlog we have. So I kind of answer your question indirectly on, do I have any more customers? We shall see.

Okay, thank you, Hock.

Thank you. One moment for our next question. And that will come from the line of Carl Ackerman with BNP Paribas. Your line is open.

Yes, thank you. Kirsten, I was hoping you could speak to the relocation of IP back to the U.S. that is causing a $4.5 billion tax liability. Historically, Broadcom has redomiciled ahead of a pending transaction, and I'm getting questions from investors saying, if this action may relate to any asset sales as the company seeks to pay down debt. So if you could clarify that, that would be helpful. Thank you.

Yeah, no, it was just the timing of when we chose to do it this time. And, no, it doesn't have anything to do with that. It's just we relocated the IP, and that caused the $4 billion charge. The offset to that is a deferred tax liability. So think of that as non-cash, very little cash impact to that.

Thank you. One moment for our next question. And that will come from the line of Timothy Arcuri with UBS. Your line is open.

Thanks a lot. Hawk, I wanted to ask about the growth rate in your AI revenue versus what we're seeing on the GPU side. Your AI revenue grew in the same zip code this year as what the GPU computes growing. And you did say that it would be up next year, but your main customer is ramping a new version of their custom ASIC next year, and there's some thought that they might shift some of their purchasing back to GPUs next year. So do you think that the growth of your AI revenue should still approximately track how much GPU compute is going to grow next year? If you can give us any qualitative or quantitative thoughts there, that'd be great. Thank you.

That's been me. Hey, Dean. I think we had some communication gaps here. Could you repeat the question?

Yeah. So the question, Hawk, really is around the growth rate of your AI revenue versus what we're seeing on the GPU side, because this year you grew about the same as what GPU computes growing. And the question is, is there anything happening next year that would change that equation so that your growth rate of your AI revenue would be materially different than what GPU computes growing next year?

You know, that's a very difficult question for me to answer because it comes in two parts, right? In terms of GPU growth, you should ask the guys who does merchant GPU or GPU, which is obviously NVIDIA and AFD. And I don't see, I don't play in the enterprise market at all. See, that's part of their market I don't see. Having said that, they do both play a somewhat in the hyperscalers where I'm totally focused on doing. So that's really very, and that's really no connection one with the other. If there is, it's indirect. But it now suffice for me to say long term, I'm saying that thoughtfully, long term, the large hyperscalers, few and large hyperscalers with very large platforms, huge consumer platforms, subscriber base, have the entire model predicated on running a lot of large language models, a lot of AI requirements, workloads out there. And it will drive, in a matter of time, towards creating as much as possible their own compute silicon, their own custom accelerators as a matter of time. And we are in the midst of seeing that transition. It may take a few years for that to happen. So that is on a different trajectory, a different path. And I'm in that path of doing that, enabling custom accelerators. I'm in that path. I'm not in the path of, in the meantime, a different trajectory of enabling enterprises to do AI on their own workloads. That's more the merchant guys. Some of the merchant guys are obviously also in the AI, in the hyperscalers today, but that's a process, obviously, of a transition going on. So one doesn't really connect to the other team in that regard. But I would likely say, obviously, as the transition occurs, we have a good tailwind. In the business model, we are providing accelerators and networking to the AI data centers of those large hyperscales.

Right, Hock. Okay. Thank you so much.

One moment for our next question. And that will come from the line of Harsh Kumar with Piper Sandler. Your line is open.

Yeah, I was curious about the profitability of VMware. Historically, your software businesses have had operating margins greater than 70%. VMware, I know, is newer and you're doing things a little different. You're keeping more customers than you historically have kept. But I was curious if you see a similar profile as the rest of your software businesses for VMware after you're done with all the cuts and everything?

Well, I'll let you draw your own conclusion, Ash, but I was a pain to lay out, as you probably heard. In Q3, our revenue from VMware was $3.8 billion, and our operating expenses is $1.3 billion. And you can pretty quickly figure out where we're headed in terms of operating margins and, as I indicated, EBITDA margins. And Q4 will continue the trajectory of revenue continuing to grow and expenses still dropping, even as it starts to stabilize, but continue to reduce.

Thank you, Hawk.

One moment for our next question. And that will come from the line of CJ Muse with Cantor Fitzgerald. Your line is open.

Yeah, good afternoon. Thank you for taking the question. I wanted to focus on software gross margin. So when you close the acquisition of VMware, you know, we tick lower from low 90s to, you know, kind of high 80s. And we're now, you know, pushing a bit higher into July. And curious, as we kind of get to that, you know, $4 billion threshold and you've kind of indicated higher, in fiscal 25, how should we think about the gross margin trajectory overall for software?

Well, you know, for us, software gross margin is actually redirect. It's not that relevant. You know that, right? Software is, unless I'm running SaaS big time. Now, a lot of our products are on subscription, but they're not SaaS. We have some of our products on SaaS cloud-based, but most of them are not. And our gross margin will be around 90% at least.

And one moment for our next question.

And that will come from the line of Chris Caso with Wolf Research. Your line is open.

Yes, thank you. Good evening. I wonder if you could speak to the custom AI revenue and perhaps the contribution from some of the other customers aside from that largest customer. You know, how meaningful are the other customers in that segment? And, you know, what do you expect into next year as some of those newer projects start to ramp up?

Well, I know, we're dancing around the thing. As I indicated, we have three customers now going on, and all three of them are meaningful. Otherwise, we won't call them customers as the criteria we have used. Until we get meaningful shipments out to them on AI accelerators, we do not really consider that as a customer. Simply because it's a new, this is an emerging trend. It's not an easy product to deploy for any customer. And so we do not consider proof of concepts as production volume. These are all production accelerators deployed in data centers, AI data centers of those three customers.

One moment for our next question.

And that will come from the line of Christopher Rowland with Susquehanna. Your line is open.

Hi, thanks for the question. My question is actually on storage. And Hawk, you bought Seagate's hard disk drive SOC assets earlier in the year. Can you talk about what you actually bought there uh what it means in terms of economics for your company um and and whether this accelerates your storage business over the next few years thanks oh this is more of a partnership than anything else basically you know it's what we what we essentially uh created in that transaction was

To begin with, we actually believe long-term in the sustainability of hard disk drive media as a great long-term sustainable storage alternative or storage medium for those hyperscales. It makes sense. One might think eventually everything goes to flash. Don't think so. hard disk drive storage will still be meaningful. And the technology, which is most interesting for us, has a lot of ways to go as hard disk drives goes on to from where it is today, which is, you know, 22, 23, 24, you know, terabytes, to going to 30, 40, and even 50 terabytes. A lot of technology along the way, and a lot of that resides in silicon. So what we're doing, in effect, is a collaboration more than anything else, though it's structured, obviously, as a purchase of intellectual property, but we're also taking engineers, designers, combining it with the designers we have, and basically enabling Seagate and eventually the entire industry to continue a roadmap that goes towards 50 terabytes. That's our ambition, that's our vision, and to be able to do that within five years or less. So that's pretty much what it is. It's a statement of our belief that nearline hard disk drive storage will sustained very well over the next five years, if not longer.

Thank you.

One moment for our next question. And that will come from the line of Aaron Rakers with Wells Fargo. Your line is open.

Yeah, thanks for taking the question. Kind of thinking strategically as we look forward you know, ahead to NVIDIA's Blackwell product cycle, there's been some indications that, you know, possibly Broadcom has an opportunity to participate more deeply in the optical side of that product platform for NVIDIA. I'm curious, do you see that as an opportunity relative to prior generations of NVIDIA just to deepen a participation or just to participate in general in kind of the areas of VSPs and maybe other things related to the Blackwell cycle from NVIDIA? Thank you.

That's an interesting question. I've got a simple answer. I'm not really participating in it. That's NVIDIA's roadmap. I'm really not directly in that kind of market, in that kind of product roadmap. That's NVIDIA's product roadmap in terms of Blackwell. Impressive product on the way to coming out. Now, in terms of base technology we developed, of course, it could be used, it could be applied, and we are very happy to share that as it may be useful to enable Blackwell to be part of that, whether it's on the optical component side, which is what you're referring to, or even the DSP side in terms of providing the interconnects to enable clusters of Blackwell to be built. That's as far as our our engagement in that. We're happy to be part of that ecosystem, as I said, but directly, we're not in that market, as you know.

Yep.

One moment for our next question. And that will come from the line of Joe Moore with Morgan Stanley. Your line is open.

Great, thank you. I wonder, Huck, if you could talk about your thoughts on further M&As You know, is that still on your radar down the road? And, you know, if you did, would it be still software-focused or any possibility of semiconductors becoming interesting to you again?

Joe, that's a beautiful question. I'll tell you this bluntly so that you're not disappointed. Right now, I'm having my hands really full and enjoying myself doing it on really turning, transforming the business model of EMOS. It's a great experience, and you're feeling great about it when you're doing it. And when you're doing it, it's pretty much running way beyond expectation, as we indicated in that slide. So, no, I'm very focused on getting VMware as it continues to accelerate in getting private cloud deployed in the largest enterprises in the world. And you know what? I've got another year, two years to go. to make that transformation totally complete.

Very clear. Thank you.

One moment for our next question. And that will come from the line of Harlan Sir with JP Morgan. Your line is open.

Good afternoon. Thanks for taking my question. You know, Hawk, last quarter you talked about an acceleration in R&D investments by your AI customers, and you talked about your follow-on wins for their next generation XPU ASIC programs. It also looks like they're trying to accelerate their deployment of their TPUs, XPUs, and networking into their data centers here in the second half of the year. We know that on AI accelerators specifically, supply is quite tight given the co-op packaging and HBM memory constraints. So has the team seen upside orders in demand for XPUs and networking? here in the second half, have you been able to meet that upside demand or is the team somewhat supply constrained? I guess in other words, is AI demand greater than your supply here in the second half of the year?

Yeah, no, we continue to see all this.

We continue to see upsides and you're right in the pattern of that behavior that's going It's, you know, as our customers, these are hyperscalers, trying to deploy more and more capacity of AI data centers, in AI data centers, and you start to hear them talk in terms of power. They don't even talk in terms of how many XPU or GPU classes they talk in. 500 megawatt, one gigawatt. Not yet, but people are dreaming that. So we are, as they get this enabled, we're getting pull-ins, we're getting upsides. And I expect that to happen a lot more in 2025. We're not putting that in any guidance or indication we're giving you, but I'm pulling what you said exactly right on. We do expect to see upsides, as we've been seeing uh recently we continue to see that probably going forward over the next 12 months especially related to xpus getting deployed and create getting infrastructure available and rushing to deploy them we see on quite a bit of that have you been able to meet that upside or are you somewhat limited by supply constraints we can meet those upsides perfect

Thank you, Hawk.

Thanks.

One moment for our next question. And that will come from the line of Edward Snyder with Tartar Equity Research. Your line is open.

Thank you very much. Hawk, that was a perfect segue into my question. You've said in the past calls that you thought that AI compute would move away from ASICs and go to merchant market, but it looks like the trend kind of heading the other way are you still the opinion that that's going to be the long-term trend of this and secondly as you just pointed out um power is becoming the defining factor for deployment um with all the big guys at this point given the performance of uh performance per watt of the asics over gpus which is superior to gpus why shouldn't we see more of these guys moving to to uh custom asics i know it takes a long time it takes a lot of funding etc but especially as the enterprise starts getting more involved with this, there are going to be some applications that are kind of standard across some of the enterprises. Wouldn't we even see some of the bigger, like AWS moved to a custom silicon for a specific workload? So basically the overall trend in ASICs in AI. Thanks.

Okay. Ed, did I hear you right to say at the beginning, maybe, that you meant that there's a trend towards asic or xpu from general purpose gp right yep you're right and you correct in pointing out to me that hey i used to think that general purpose merchant silicon will win at the end of the day well based on history of semiconductors mostly so far general purpose merchant silicon tends to win but Like you, I flipped in my view. And I did that, by the way, last quarter, maybe even six months ago. But nonetheless, catching up is good. And I actually think so because I actually think there are two markets here on AI accelerators. There's one market for enterprises of the world. And none of these enterprises are incapable nor have the financial resources or interest to create the silicon, the custom silicon, nor the large language models and the software going with it to be able to run those AI workloads on custom silicon. It's too much. And there's no reason for them to do it because it's just too expensive to do it. But there are those few cloud guys, hyperscalers, with the scale of the platform and the financial wherewithal for them to make it totally rational, economically rational, to create their own and custom accelerators because it's all, right now, I'm not going to, I'm not trying to over-emphasize it. It's all about compute engines. It's all about, especially training those large language models and enabling it on your platform. It's all about constraint to a large part about GPUs. Seriously, it's getting to a point where GPUs are more important than engineers. It is some of these hyperscalers in terms of how they think. Those GPUs are much more, or XPUs, are much more important. And if that's the case, what better thing to do than bring it under the control of your own destiny by creating your own custom silicon accelerators. And that's one I'm seeing all of them do. It's just doing it at different rates, and they're starting at different times. But they all have started, and obviously it takes time to get there, but a lot of them, there are a lot of learning in the process versus what the biggest guy of them who has done it longest have been doing for seven years. Others are trying to catch up, and it takes time. I'm not saying it will take seven years. I think it will be accelerated. But it will still take some time, step by step, to get there. But those few hyperscalers, platform guys, will create their own if they haven't already done it and start to train them on their large language models. And that's, yeah, you're right. They will all go in that direction, totally. into ASIC, or as we call it, XPUs, custom silicon. Meanwhile, there's still a market in enterprise for merchant silicon.

Right, but that basically suggests that you're on the early part of your curve where, I'm not trying to call the GPUs or whatever, you could be getting to something closer to the peak of the GPU market just because everything, right, besides the cost expense and as you're spending all this money and you're paying all this money for power, the ASICs become more and more attractive. So the curves are going to look different, right?

It's an accelerating curve. It may take longer than we all want it to happen, but it's definitely accelerating because the size of those and the size of the Demand from those hyperscalers will totally rival that in the enterprise.

Thank you. And that is all the time we have for our question and answer session. I would now like to turn the call over to GU for any closing remarks.

Thank you, operator. This quarter, Broadcom will be presenting at the Goldman Sachs Communicopia and Technology Conference on Wednesday, September 11th in San Francisco. Broadcom currently plans to report its earnings for the fourth quarter and fiscal year 2024 after the close of market on Thursday, December 12, 2024. A public webcast of Broadcom's earnings conference call will follow at 2 p.m. Pacific. That will conclude our earnings call today. Thank you all for joining. Operator, you may end the call.

This concludes today's program. Thank you all for participating. You may now disconnect.