Marvell Technology Group Ltd
NASDAQ:MRVL
If you don’t study any companies, you have the same success buying stocks as you do in a poker game if you bet without looking at your cards.
Utilize notes to systematically review your investment decisions. By reflecting on past outcomes, you can discern effective strategies and identify those that underperformed. This continuous feedback loop enables you to adapt and refine your approach, optimizing for future success.
Each note serves as a learning point, offering insights into your decision-making processes. Over time, you'll accumulate a personalized database of knowledge, enhancing your ability to make informed decisions quickly and effectively.
With a comprehensive record of your investment history at your fingertips, you can compare current opportunities against past experiences. This not only bolsters your confidence but also ensures that each decision is grounded in a well-documented rationale.
Do you really want to delete this note?
This action cannot be undone.
| 52 Week Range |
50.43
93.94
|
| Price Target |
|
We'll email you a reminder when the closing price reaches USD.
Choose the stock you wish to monitor with a price alert.
This alert will be permanently deleted.
Marvell Technology Group Ltd
You don't have any saved screeners yet
Ladies and gentlemen, thank you for standing by and welcome to the Q3 2020 Marvell Technology Group Earnings Conference Call. At this time, all participants are in a listen-only mode. After the speaker presentation, there will be a question-and-answer session. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions]
I would now like to hand the conference over to your speaker, Ashish Saran, Vice President, Investor Relations. Sir, please begin.
Thank you, and good afternoon, everyone. Welcome to Marvell's third quarter fiscal year 2020 earnings call. Joining me today are Matt Murphy, Marvell's President and CEO; and Jean Hu, our CFO.
I would like to remind everyone that certain comments today may include forward-looking statements, which are subject to significant risks and uncertainties and which could cause our actual results to differ materially from management's current expectations. Please review the cautionary statements and risk factors contained in our earnings press release, which we filed with the SEC today and posted on our website as well as our most recent 10-K and 10-Q filings. We do not intend to update our forward-looking statements.
During our call today, we will refer to certain non-GAAP financial measures. A reconciliation between our GAAP and non-GAAP financial measures is available on our website in the Investor Relations section.
With that, I'll turn the call over to Matt for his comments on our performance.
Great. Thanks, Ashish, and good afternoon to everyone on the call. As a reminder, we closed the acquisition of Aquantia in our third fiscal quarter, so our consolidated results include a partial quarter of results from their operations as well.
During the third quarter of fiscal 2020, we delivered solid results and achieved $662 million in revenue. Our GAAP loss per share was $0.12, our non-GAAP earnings per share was $0.17 at the midpoint of guidance. Our storage business grew sequentially and we had better-than-expected results from our networking business. While we cannot influence the macroeconomic environment, I am very happy with our execution on the things we do control. Earlier this year, we discussed a number of significant design wins for our products in 5G base stations, enterprise switches, automotive Ethernet, and datacenter storage.
In the third quarter, we continue to add to this growing list with more wins which I will discuss later in the call. Our engagement with customers continues to strengthen, which is becoming evident in our success in winning high value sockets in all our target markets, which we expect to drive sustainable growth over many years.
In fact, one of our important growth drivers has already started to benefit us and I am very pleased that we shipped a significant amount of our 5G products in the third quarter, helping our key customer rapidly rollout base stations for the initial wave of 5G deployments in Korea. Of course, this is just a start of 5G deployments and we are looking forward to a lot more growth from 5G in the longer-term.
More recently, we closed the acquisition of Avera on the first day of our fourth quarter. The acquisitions of Aquantia and Avera are very strategic to broadening our semiconductor solutions portfolio for our infrastructure customers. We are well on our way to integrating both teams within our networking business, and I would like to extend a warm welcome to all the new employees who have joined the Marvell team.
We also expect to complete the divestiture of the Wi-Fi Connectivity Business to NXP in the first half of this month. However, as we currently still own the Wi-Fi business, my comments today on revenue expectations for the fourth quarter, include a full quarter of expected results from that business.
I'd like to take this opportunity to personally thank the 600 dedicated employees in that business for their years of service to Marvell and for staying focused through this transition. We wish them all the best.
Now moving on to the performance of our two core businesses. First, in our networking business, revenue during the quarter was $330 million, flat sequentially as compared to our expectations of a low single-digit decline. This improvement was split between partial quarter contributions from the Aquantia acquisition and organic Marvell businesses.
During the quarter, macroeconomic uncertainty continued to impact demand from the enterprise end market. Additionally, as expected, demand for our Wi-Fi products declined seasonally. In contrast, our embedded processors delivered double-digit sequential revenue growth, driven by strong 5G shipments.
As you may recall in the early part of this year, we had anticipated the start of initial production for our 5G products in the fourth quarter of this fiscal year. However, last quarter, we announced that we were able to start production in the third quarter. Our key customer requested very aggressive delivery timelines to meet their commitments supporting 5G rollouts in Korea.
Outstanding execution by our engineering and operations team, allowed us to rapidly ramp production and successfully deliver 5G products in volume within the third quarter itself. This was also very impressive demonstration of the maturity of our platform and quality of our NPI process that enabled us to transition from sampling our new 5G processors in the first quarter to driving a production ramp shortly thereafter.
The third quarter was just the start of the 5G ramp for us as it was driven primarily by a single region. We expect these initial deployments in Korea to continue to be the main driver for our 5G revenue for another couple of quarters and we should start to see the benefit from other geographies later in fiscal 2021.
By the time we approach the second half of next fiscal year, we expect a strong increase in our 5G-related revenue driven by continued Korea deployment and the start of 5G adoption in Japan and other countries such as the U.S. as they start to install 5G base stations in higher volumes.
In addition, we expect to start production of baseband processors for our second Tier 1 base station customer in the fourth quarter of fiscal 2021 further adding to our overall 5G revenue ramp.
Our progress to date has also enabled us to engage with them in longer-term technology roadmap discussions. The broadening adoption of our 5G platform is a testament to the advanced capabilities of our multi-core embedded processor portfolio further complemented by our Ethernet switch and PHY products. And we now have added Avera's full custom ASIC capabilities, enabling us to offer an industry-leading 5G suite of technology.
But before I move on to the Avera acquisition, let me spend a couple of minutes on the unique capabilities of our OCTEON-embedded processor platform. As you may recall, OCTEON was the first multi-core processor architecture with a complete set of on-chip hardware accelerators for networking applications. Since then the OCTEON platform has been widely adopted for control and data plane processing across a broad range of networking applications, including base stations, switches, routers, broadband access and firewall appliances.
This platform now scales up to 48 optimized ARM cores and includes hardware optimized packet processors, crypto engines, virtualization support and a wide variety of network interfaces. Our Fusion baseband processors are built on the same proven OCTEON architecture, adding baseband specific hardware accelerators and DSP cores.
An important benefit for OCTEON and Fusion customers is the ability to reuse the entirety of their sizable control and data plane software investment as they design in our products across their portfolio. The software reuse not only speeds our customers development cycles, but also creates very sticky designs for Marvell.
Our OCTEON and Fusion family include fully programmable ARM & DSP cores, enabling an extremely flexible and scalable architecture further optimized by implementing certain fixed functions and hardware acceleration blocks.
The speed, simplicity and scalability of the OCTEON programming model has been a key differentiator and an important driver for the adoption of our processors in multiple applications. In addition, as processing spreads into new domains and base stations, such as the radio head, customers are also turning to our processors for new functions.
As an example, as we discussed earlier this year, one of our customers adopted our existing Fusion baseband solution from massive MIMO processing. This represented their fastest route to market as their software was already up and running on the OCTEON architecture and the Fusion implementation was flexible enough to extend to their massive MIMO application. Longer-term, customers will often require a customized OCTEON Fusion processor optimized for specific applications, such as the complex Beamforming needed for massive MIMO.
I am pleased to announce that we now have been awarded such a design win by a key customer for a Fusion-based solution customized for massive MIMO processing in the radio head. The additional complexity in the solution is also enabling a higher ASP than our existing baseband products.
In addition, we are driving a fast time-to-market by leveraging the flexibility in our processor platform and expect to complete development in under one-year allowing us to start ramping this product into production by the beginning of calendar 2021. We are also in advanced discussions with the second customer for processing solutions for massive MIMO.
Moving onto Avera. The combination of Avera and Marvell creates an infrastructure ASIC powerhouse with the scale and design flexibility leading wireline and wireless OEMs are looking for from their semiconductor partners. Kevin O'Buckley, who is Avera’s General Manager has joined Raghib’s team and will be leading the ASIC business for Marvell.
In addition to Kevin, we are fortunate to be adding multiple key leaders from Avera who along with the talented global Avera team bring a tremendous amount of industry experience and technical capability. This team originally part of IBM's Microelectronics business brings a proven track record of delivering a large number of highly-complex ASICs to very demanding customers.
In terms of revenue, we continue to expect that Avera will contribute approximately $300 million in annual revenue next fiscal year. A significant amount of Avera’s revenue is from the wireless infrastructure market primarily for digital front-end ASICs and radio heads which aligns extremely well with Marvell’s expanding presence in base stations. The base station revenue from Avera will be an addition to the organic Marvell opportunities I discussed earlier.
We estimate that Avera brings an incremental $4 billion to Marvell’s addressable market across the datacenter, carrier, enterprise and automotive end markets. Avera’s design capabilities coupled with our broad technology platform will allow us to leverage our combined IP portfolio across the full spectrum of custom, standard and semi-custom solutions. We also expect to benefit from a broader combined customer base.
As I mentioned earlier, we have now added Aquantia’s leading multi-gig Ethernet products to Marvell’s broad switch and PHY portfolio, extending our reach in both the network, infrastructure and the rapidly growing automotive Ethernet market.
Faraj Aalaei, who is Aquantia’s CEO has joined Marvell and taken on a broader role has the General Manager for all our Ethernet products, including automotive. Faraj brings a tremendous amount of industry experience and we're looking forward to working with him and his team. We have already integrated roadmaps to create a very compelling end-to-end Ethernet connectivity portfolio and in fact the combined Marvell and Aquantia team have already secured their first major design win together.
This win was for a multi-gig Ethernet PHY in a Tier 1 networking OEMs campus switch platform, and we expect to start ramping production in the second half of next fiscal year. In addition, the response from multiple automotive customers to the combined company roadmap has been very strong and we are jointly working on additional opportunities to add to the 16 design wins Marvell has already won across a broad range of OEMs.
From a revenue perspective, macroeconomic uncertainty impacted Aquantia’s revenue in the first half of Aquantia’s fiscal year as their key customers reduced purchases to better master inventory with demand. With the acquisition now complete, we expect to grow this business to approximately $100 million in annualized revenue in fiscal 2021. In fact, Aquantia’s revenue was already started to recover from their June quarter on a run rate basis.
Turning now to the growing ecosystem for ARM-based compute in the datacenter. At the Supercomputing Conference in Denver, we announced the availability of NVIDIA GPU support for ThunderX2-based server platforms to provide HPC and cloud customers with a powerful solution for exascale computing.
In addition, we recently announced that Microsoft has started to deploy servers based on Marvell's ThunderX2 server processors for internal Azure development workloads. We believe that this is an important milestone on the path to the broader adoption of ARM-based servers across the Azure platform. The progress of ThunderX adoption at our cloud and HPC customers gives us confidence and expecting revenue growth next fiscal year from our server processors starting from a base of low initial volumes and then continuing to ramp in the second half.
As we look beyond the datacenter server market, we are also seeing multiple opportunities to leverage our ThunderX processor cores in SoC or ASIC solutions for applications requiring embedded high-performance compute engines.
Moving onto our outlook for the fourth quarter for our networking business, driven primarily by full quarter contributions from Avera and Aquantia, we expect an approximate 25% sequential increase in networking revenue. We are expecting revenue from Aquantia products to continue to recover from the low point they reported earlier this year.
For the rest of Marvell's networking business, we expect 5G shipments to remain strong, while demand from the enterprise end market remains weak, including a decline from a number of Chinese customers.
As you may recall in prior quarters, we had indicated that these customers appear to be building inventory to guard against any future potential supply chain disruptions. As a result in the fourth quarter of fiscal 2020, we are anticipating a sequential decline from that elevated run rate, in addition to general end market softness. We are also projecting the Wi-Fi business to decline seasonally in the fourth quarter.
Turning to our storage business. Storage revenue for the third quarter was $288 million, growing 5% sequentially. As expected, this growth was driven primarily from the enterprise and datacenter market. However, we did experience weaker than expected revenue from the edge market, with a softening in demand from gaming and video surveillance.
We also continue to see a rapid conversion from HDDs to SSDs in the PC market, which presents a headwind for our client business. Despite challenges from current trends in the client and edge markets, our strategy to focus on the enterprise and datacenter market, including a strong position in high-capacity cloud drives, continues to payoff to drive growth for the overall storage business.
Last quarter, we introduced the industry's first PCIe Gen 4 SSD controller based on a 12-nanometer process. This product delivers higher performance in a lower power envelope in a space optimized package to enable our customers to develop smaller form factor solutions. In addition, our solution also allows customers to eliminate DRAM from their SSDs, which also reduced the size and power while lowering their bill of material.
Multiple customers, including several NAND OEMS are deeply engaged in evaluating our product as it uniquely addresses a number of their emerging mainstream SKUs where they need to balance performance and power in a small form factor solution.
This PCIe Gen 4 architecture is also the basis for our customized SoC flash controllers for embedded and do-it-yourself applications. Our first major DIY design at a Tier 1 system level OEM is in the final stages of development. We expect a sharp ramp in revenue from this product in the third quarter of next fiscal year when our customer begins mass production of their system.
Turning to the fourth quarter for our storage business. We project revenue to be flat to up slightly on a sequential basis. We expect our revenue from the datacenter and enterprise end markets to continue to grow from a strong third quarter and offset weaker trends in the edge and client markets.
In closing, we continue to win new designs and position ourselves for solid growth with strong execution on new product introductions. Our 5G products have started shipping in volume, helping us offset macro weakness in other end market such as enterprise. We expect 5G base station deployments to expand into multiple geographies later next fiscal year, which we anticipate will drive a substantial increase in revenue.
In addition, we continue to expand our engagement with a broad set of Tier 1 5G customers. Our storage business has stabilized, benefiting from our growing exposure to enterprise and datacenter applications and grew appreciably in the third quarter. The acquisitions of Aquantia and Avera meaningfully increased our addressable market opportunity and add the Marvell's infrastructure capabilities to create a very unique technology platform. I am looking forward to an exciting fiscal 2021.
With that, I'll turn the call over to Jean for more detail on our results and outlook.
Thanks, Matt, and good afternoon, everyone. I will start with a review of our financial results for the third quarter, which includes a partial quarter results from the Aquantia acquisition, which we closed on the 19 of September. As a reminder, the guidance provided on the 29 of August for the third quarter did not include any contributions from the pending Aquantia acquisition. Revenue in the third quarter was $662 million versus our guidance of $660 million at midpoint.
Networking represented 50% of our revenue in the third quarter, with storage contributing 43%. Revenue from our other business was $45 million in the third quarter below expectations and declined 14% sequentially. Other products accounted for 7% of revenue.
GAAP gross margin was 51.3%. Non-GAAP gross profit was $421 million or 63.5% of revenue at midpoint of our guidance range. GAAP operating expenses were $402 million.
Non-GAAP operating expenses were $283 million, slightly above the midpoint of guidance due to the closure of Aquantia acquisition within the quarter. Without the impact from the Aquantia acquisition, non-GAAP operating expenses would have been below the low end of guidance.
Our current level of OpEx reflects a lower level of bonus accrual. GAAP operating loss was $62 million. Non-GAAP operating profit was $138 million or 21% of revenue. GAAP loss per diluted share was $0.12. Non-GAAP earnings per diluted share was $0.17 at a midpoint of our guidance.
Now turning to the balance sheet. In the third quarter, we returned $40 million to shareholders through dividend and then we exited the quarter with $438 million in cash and cash equivalents. Long-term debt was $2 billion, an increase of approximately $350 million from the prior quarter. The increase was from drawing down our revolving credit facility for funding acquisition of Aquantia.
Inventory increased to $68 million from the prior quarter, $31 million of the increase with from the Aquantia acquisition including the accounting fair value step up. The balance of the increase was to support the expected growth of our 5G business.
Let me now move on to our current outlook for the fourth quarter of fiscal 2020, which include the expected results from the Aquantia and Avera acquisitions for the full quarter. This guidance also includes expected results for the full quarter from the Wi-Fi business.
Once the sale of the Wi-Fi business is completed, we will update the revenue portion for our guidance for the fourth quarter. We continue to expect our revenue in the current quarter to remain impacted by the U.S. government’s export restrictions on certain Chinese customers.
Specifically, we expect revenue to be in the range of $750 million plus or minus 3%. We anticipate our GAAP gross margin will be approximately 45.5% and our non-GAAP gross margin will be approximately 62%. Avera is a cover out from GlobalFoundries and that we have a certain one-time transition cost that are expected to negatively impact our gross margin in the fourth quarter.
We expect our gross margin to return to more than 63% once we complete the divestiture of Wi-Fi business and a fully integrate Aquantia and Avera into Marvell supply chain. We project our GAAP operating expenses to be in the range of $445 million to $455 million. We anticipate our non-GAAP operating expenses to be in the range of $315 million to $320 million.
We expect net interest expense to be approximately $26 million. As a reminder, at the beginning of our fourth quarter, we financed the acquisition of Avera with a $600 million bridge loan, and this increase in debt is reflected in our guidance for interest expense. We anticipate GAAP loss per diluted share in the range of $0.17 to $0.23 and the non-GAAP income per diluted share in the range of $0.15 to $0.19.
Before we open the line for questions, let me make a couple of longer-term comments on operating expenses expectations given we have a few moving parts from three transactions. First, due to the typical seasonality in payroll taxes, merit increases and the reset of bonus accruals from the current low level, we expect our non-GAAP OpEx in the first quarter of next fiscal year to grow by approximately 5% on a sequential basis.
Second, we have quickly started to integrate both Aquantia and Avera and the ERP integration have been largely completed, enabling a faster start toward synergy achievement. We will continue to drive the remaining synergy achievement and the operating efficiency improvement through fiscal 2021, while investing in a growth initiative to position us to deliver long-term profitable growth.
Once we complete the divestment of Wi-Fi business and the fully integrated Aquantia and Avera, we expect OpEx to decline through the rest of fiscal 2021 from the Q1 high point I just outlined. As a result, we project our non-GAAP operating expenses in Q4 next fiscal year will be approximately $300 million.
Operator, please open the line for Q&A. Thanks.
[Operator Instructions] Our first question comes from the line of Vivek Arya of Bank of America. Your question, please.
Thanks for taking my question. Matt, a few quarters ago, you gave $600 million long-term opportunity in 5G, what is the confidence in achieving that? And just what does that number look like now that Avera in the mix, and I think you highlighted some additional opportunities also in massive MIMO.
Yes. Great. Thanks, Vivek. So I think to answer your first question, you're right. A couple of quarters ago, we articulated the $600 million number which was really the business that we had already secured design wins for. We still feel very good about that opportunity.
And in fact, as we indicated in the prepared remarks, there's even been some incremental change to that on the positive side as we transition at least one of our customers to a higher ASP product in the massive MIMO area. So that tends to be tracking well in terms of the design opportunities and our customer positioning. So we're quite confident there.
On Avera, what we said was about half their revenue was from the wireless infrastructure market and base stations of which a portion of that is tied to 5G. And certainly, we see that contributing as well and that would be on top of the $600 million.
And I look forward to completing full quarter with Avera under our belt and then articulating the combined opportunity in the future, but it's definitely above $600 million just given the fact that Avera is bringing in pretty significant revenue in the base station market.
All right. And for my follow-up, Jean, thanks for giving us the gross margin and the OpEx puts and takes. I was hoping, you could just for clarification give us just the contributions of the different M&A moving pieces in your Q3 and Q4, so that we can align the models accordingly.
Yes. Vivek, it's actually quite straightforward. If you think about our operating expense guidance for Q4, which include both Aquantia and Avera for the full quarter, which also include the Wi-Fi business. So moving forward, if we end up closing the Wi-Fi sale in our Q4, the best way to think about it is, I also guided you Q4 next year, the OpEx is going to be $300 million without Wi-Fi business.
So a simple way to think about this, that's how we're going to migrate. We're going to have a merit and bonus increase reset in Q1, but we're going to absorb those increase by improving efficiency and achieving additional synergies. So you can largely view the difference between Q4 this year, and the Q4 next year. It's the difference, it's the Wi-Fi business getting out.
I meant on the sales side, Jean, if you could give us the puts and takes in Q3, Q4 from –you completed Aquantia in Q3, which added probably some small amount. And then in Q4, right you have the two acquisitions. But then Wi-Fi goes out, but is down seasonally. I'm just trying to contrast the outlook that you gave versus what people's expectations were going into the quarter. Thank you.
Okay, great. So Q3, Q4 is actually when you think about that is Q3 our OpEx is about $280 million without any acquisitions. We actually achieved the significant amount of the synergy right at the beginning of the two acquisitions. So the two acquisitions probably add about $37 million of OpEx. That's why the Q4 midpoint of OpEx guidance is the current level.
So the way to think about it is $280 million step up because of the two acquisitions. And then when you get into Q1 next year, of course, we're going to have the Wi-Fi business coming out if we close the transaction. But at the same time the OpEx will stay around the same level because that 5% sequential increase from the merit and the bonus reset.
Hey, Vivek, this is Ashish. So I think from a – I think your question is also on the topline side. So I think for Q3, again, Aquantia was therefore, a fairly small period and it's a fairly small number. So the Q3 revenue is pretty much, you should assume it's really Marvell with a very small help from the Aquantia side.
I think on Q4, again, I think what we can do to remind you of kind of what we expect on a full-year basis. So as Jean said, Avera is a $300 million business, right for next year. Aquantia, next year, we believe, is about $100 million. But as Jean said, it's still recovering toward that number, right as you remember from where they were originally.
Wi-Fi, again, if you remember when we said the divestment, it's a $300 million business, right. So you can do the quarterly number and it's on the low end of that number given we are at the seasonally low point. I think that's how you can think about revenue for Q4.
Thank you.
Thank you. Our next question comes from Blayne Curtis of Barclays. Your line is open.
Hey, guys. Thanks for taking my question. I was just kind of curious, I mean, you've obviously had, you've talked about more opportunities on the wireless side and you can deal with. You haven't had ASIC capabilities. Just kind of curious the timelines that kind of combine these roadmaps and scale of Avera and kind of if you can just talk about the pipeline, as you look forward, opportunities that you've looked at with your customers that you could potentially move over with those capabilities.
Sure, great question. So a couple of things, so one is, I talk about in two dimensions. The first is, we've already had hit the ground running on the customer side. So there has been a pretty extensive global roadshow with the combined Marvell and Avera team to articulate our strategy for both 7 and 5-nanometer in our technology platform.
We're doing this in a very targeted manner and really think about it as we're going to address. We're going to line up our IP roadmap that we've had inside the company with the end customer orientation. So we're going to go more narrow and go much bigger with a handful of large customers across those end markets I mentioned which is cloud, enterprise, automotive and also base stations and that's gone extremely well.
The second is we've done already a very good job with the team, integrating the IP team from Avera and also the backend teams into one consolidated Central Engineering effort. So when we hit the ground running on our next-generation process technology platform, we're going to have one platform that we can leverage for our ASIC business for our semi-custom business and for our standard product business.
And that's resonating very well with the customer base. I think they wanted to see us actually bring the team together quickly and have one integrated offering. And I would say the feedback has been phenomenal from the roadshow in terms of interest from all those end markets that I mentioned.
Thanks. And then maybe if you just talk about the ramp of your lead base station customer, you said you shipped it early, you've been hitting these aggressive timelines. Obviously, they needed to deploy those, get them quality carriers and ramp. I think you’re looking for contribution in January. Can you speak to whether that's still on track in your guidance? And then if there is any visibility, you mentioned some regions, but in terms of the traction they're getting and your visibility into getting those targets for that customer? Thanks.
Sure. Yes, I know it ramped up and it ramped up fairly steeply, which was positive, it's going to probably remain certainly from a Korea standpoint, it will keep deploying at that level. And then we see continued momentum as our customer deploys in other regions and within those other regions, other carriers within those regions. And certainly even in the Korea market, I think the best is still in front of us, if you look at subscriber penetration, which has been very fast in Korea. I think to-date, it's around 15% of total subscribers around 5G and certainly that's going to trend significantly higher next year.
So I think of it as sort of the first step is Korea's ramps and it's going to stay strong for the next few quarters and then we anticipate additional regions kicking in, in the outer quarters with what looks to be a very strong second half as 5G really picks up momentum and our design position should enable us to participate in that.
Great. Thank you.
Yes.
Thank you. Our next question comes from Ross Seymore of Deutsche Bank. Your question please.
Thanks for letting me ask a question. Matt, I want to focus more on the enterprise side of things. In your preamble, you talked a little bit about the China inventory digestion. But last quarter you also talked about some of the big customers that hit you with some order cancellations relatively close to your call. So overall, I just wanted to see, how long do you think that inventory burn in China is going to last and did the weakness that you saw from that large customer continue into the fourth quarter guidance, or is that kind of plateaued?
Sure. Thanks, Ross. Let me talk about them in two distinct pieces because they are a little bit different. I think on the China side what we signaled was we did see inventory being built. I think, primarily due to concerns about restrictions from an entity list point of view, and we had signaled that over the last couple of quarters, we saw it happen and we called out that we now see the – we see that reflected in our Q4 outlook, which is that inventory had been built in Q3 and that was not going to continue into Q4.
So that's sort of a separate geo political issue on the enterprise market relative to the broader enterprise market. We certainly were impacted. Last quarter, we did signal that we had a slowdown in the outlook fairly close to the call. That's continued. And I think if you just look at the commentary in general, out of the market. I think most of the large companies selling into the enterprise are continuing to have a weaker outlook primarily due to the macro.
And I think the way we look at it is, certainly there is some – if you look at United States that while the consumer is strong, I think the business environment tends to be fairly cautious and until there is maybe more clarity on some of the trade related issues or there is some conclusion, it's going to continue to remain choppy and so that's what we're seeing currently.
I'd say – I’d end this whole thing by saying at the same time, because of our – the investments we made in some of these new emerging areas and candidly the teams very strong execution on our new product development, we've been able to offset some of this weakness and 5G was one example. This past quarter, which we expect to continue, but we also as we said in our prepared remarks, we see multiple growth drivers heading in the next year, which even with a weaker macro we think our product cycles will carry us through.
That's great. Thanks for that. And my follow-up one for you Jean, I guess maybe a clarification, then the question. The clarification on the OpEx, does it got 5% sequentially in the first quarter for all the reasons you described before, but then comes back down when the Wi-Fi comes out, so that nets out kind of flat?
And then the question behind that would be on the gross margin side of things, you gave great color about what it would do once you get the integration done. Would you care to give us any idea on exit rates like you said on OpEx being $300 million next fiscal year. What the gross margin might be? Is that about when you get over the 63% again or is there some different timetable we're talking about?
So if you think about gross margin, right the two major drivers are revenue level and the product mix. We are quite confident once we complete the Wi-Fi divestiture and the fully integrated two acquisitions that we should get to about 63% very quickly, but then beyond that really is going to be driven by revenue level.
When our revenue level recover, it helps that you pull out gross margin further about that level, and of course, product mix is another one. Typically, we'll see how the product mix goes each quarter, but we are quite confident we'll be able to drive the gross margin about 63% quickly probably in Q1 next quarter – next fiscal year and go beyond that.
On the OpEx side?
On the OpEx side, I think really – you're absolutely right, when you look at the bonus reset and merit to increase the payroll tax all added together in Q1. That's the 5% sequential increase that we discussed, but then we'll take down that OpEx to the $300 million level access in Q4 fiscal 2021.
Thank you.
Thank you. Our next question comes from Gary Mobley of Wells Fargo. Your line is open.
Hi, everyone. Thanks for taking my question. Let me apologize in advance if there's any background noise. Wanted to ask about sort of the revenue breakdown, post sale of the Wi-Fi and Bluetooth business? It sounds like the baseline there for that business in the fourth quarters, excluding that business $675 million. And how should we think about the seasonality off of that as we start out fiscal year 2021 and progress through that fiscal year?
Yes. Okay. Gary, I'll try to give you some color. We're certainly not guiding next fiscal year, but typically the way to think about our revenue seasonality without Wi-Fi business is Q1, sequentially versus the Q4 they will be slightly low single-digit decline that's just regular seasonality.
And then when you migrate into Q2 will increase slightly. I think the most important thing in the second half as Matt mentioned, we have some very significant revenue drivers, so that related to our own product cycle despite overall normal seasonality or macroeconomic situation, those drivers are coming from the 5G from DIY SSD solutions. So those things are going to drive our revenue significantly higher in second half. So that's the only color I can give you right now.
Okay. Well that's helpful. And just to confirm some of the stats for the connectivity business, was it being down seasonally in the fourth fiscal quarter, presumably somewhere in the ballpark of $65 million. Can you confirm that this business is still 50% gross margin business in more than likely operating breakeven in the current quarter?
I think the right way to think about this, what we gave to you guys before right. This business operated around the $300 million revenue with some seasonality and gross margin around the 50%. So that's what we provided in the past.
Okay. All right, fair enough. Thank you, guys.
Thanks, Gary.
Thank you. Our next question comes from C.J. Muse of Evercore. Your line is open.
Hi, good afternoon. Thanks for taking the question. I guess as you learn more about the plans for your lead customer on the 5G side and you gain a better understanding of kind of geographically outside of Korea, the build plans. Can you speak to how you think about your market share there as well as content as we push forward?
Yes. Just to make sure I got this, you're really saying hey beyond Korea for our lead customer, what is the deployment look like, and then what is our content relative to maybe some of those different regions? Did I get that right, C.J.?
Yes, exactly.
Yes. So I think the way to think about it is, first, just on the content side, this tends to be one platform. The only differences I would note is to the extent that in certain regions that massive MIMO is deployed more extensively, then that's certainly a positive thing for us because we have more content in those radio heads. And as I mentioned, we already have some content today in radio heads and in the future we would have in a more customized purpose built solutions.
Beyond Korea, we see certainly Japan deploying next year. A lot of that's driven by the country hosting the Olympic Games. Certainly, for those of us here in the United States, you can see the pretty strong momentum among the various carriers about their plans to deploy 5G and that's probably more robust later in 2020.
I think there is certainly with respect to the device side, there is a slew of devices out there, but the sort of the Holy Grail that people have in mind is when there is an iPhone 5G and to the extent that that the timing of that product becomes clear, I think we certainly think that that will help overall adoption of 5G and drive demand for the service, probably very similar to when the product initially came out and made the transition to 3G, that drove a big cycle and we've seen cycle subsequently from 3G to LTE to 5G.
So I think there is a number of positive data point, certainly we get more and more encouraged about the adoption of the technology. And then there are other things going on. There is a spectrum licenses in India and other countries, and I think we're just going to see a continued sort of evolution of various countries around the world deploying, but for us it's really Korea, Japan, United States and then to some extent depending on when India decides to kick in.
Our lead customers always had a strong presence in that region. But again, content, you should assume unless there's some very high adoption of massive MIMO in a particular region should be very similar to the numbers we've been talking about.
Okay, very helpful. I guess as my follow-up, now that Avera disclosed. Can you talk a bit about the strategy to try to bring there, I guess your number one customer there in-house on the ASP side?
Well, what I'd say is and what's resonating well is that the – across all of – quite frankly our combined Tier 1 5G customers and base station customers is that our end-to-end portfolio and our offering is second to none, okay. I mean, there is not another company that has all of these various pieces under one roof, right from Baseband IP to transport processing IP, front haul connectivity. Obviously, DFE, now we're getting very strong DFE experience and capability from Avera and plus some of the massive MIMO efforts we're putting in.
So that I would say across the board has resonated extremely well that entire data chain being able to be provided by Marvell, and so certainly we're engaged both with the Avera customer base as well as the Marvell customer base and presenting it as one unified strategy.
There is no specific update on an existing customer, but I will say just having brought Avera on board, recently, I mean we've already gotten RFQs from several customers that – new customers for Avera right and we're bringing those Marvell relationships and so we're still early days, but certainly to be effectively a month in and be already engaged in these types of discussions is very encouraging to me.
Very helpful. Thank you.
Yes.
Thank you. Our next question comes from Joe Moore of Morgan Stanley. Your question, please.
Great. Thank you. I wonder if you could talk a little bit about the Thunder opportunity, you talked in the main remarks about the update from Microsoft, which was good to see. How do you size that potential? And what are the prospects for other customers, both with domestic cloud and also in China for that?
Sure. So let me give a big picture first, and then we can give – frame our opportunity. So I think one positive thing certainly out of the Super Computing Conference this year and you can even look as recently as the last couple of days with Amazon's events. It's pretty clear that the ecosystem for ARM-based server class CPUs continues to grow.
I mean, certainly in the couple of years, we've had the business or even less. There has been a tremendous progress there and so we look at it, one, as a positive that there is a number of different solutions, now they are coming to market. The partnership with NVIDIA, it's been significant. And I think the ability to go work interoperate with them, certainly in the exascale type of opportunities is a big deal.
We were pleased to announce that we progress on the Microsoft front and that we put out the joint press release and we certainly see that those deployment starting initially in the early part of next year and then ramping in the second half. I think it's a little bit early to call the size and the slope. But clearly, we're encouraged that we've gone through the gauntlet of qualifications and joint development there over the last couple of years, it's been a great partnership, and we are looking forward to this ramping into production in the second half.
And I think everybody knows the sort of size of the server CPU market is deca billions of dollars. I mean at our own Analyst Day, we said the subset of that that we were going after was probably in the range of $4 billion or so of SAM, just as a subset. So it still remains a big opportunity, still lucrative, but we'll probably give more color as we approach production ramps to be able to give you a sense of what that looks like.
Great. Thank you.
Yes. Thanks, Joe.
Our next question comes from Harlan Sur of JPMorgan. Your line is open.
Good afternoon. Thanks for taking my question. On your second Fusion customer for 5G base stations, they've made some commentary on wanting to pull in their development schedules and accelerate their time to market, with their 5G platforms. The Fusion program is a semi custom program.
So I think that there's only so much that you guys can do to kind of pull in these type of projects, but do you see the potential for possibly pulling in the production ramp into fiscal Q3 and then on future programs with this customer, which you already seem to be engaged with, what's the potential of increasing your dollar content to include things like OCTEON-embedded for control and data plane, your Ethernet products, your radio head content from both core Marvell and Avera.
Sure. Well, thanks for the question. So I think the way I'd frame it at a high level, is if you look at our lead customer and I noted some of the statistics about how fast we were able to bring them to production. And just the overall, development cycle time. I think that was one compelling reason why we were chosen.
And the second opportunity was really the confidence in the team's ability to dramatically improve the historical cycle times of getting these types of ASICs out in general in the base station market. And so we've got some really strong proof points I think and we're executing extremely well to the aggressive schedule that we had committed to upfront for the second customer. We certainly approved our team's capability to the extent that people want to try to go earlier than we would certainly do everything we could.
We are currently on track to sample the product, tape out sample production in line with what we've been saying and that's currently the plan of record. But we're always very sensitive when our customers want to go faster and to the extent that they – we could mutually align and that we would, but at this point, we are committed – we already committed to pretty aggressive schedule to begin with. And so that's what we're sticking to.
On the second part of the question. At least so far the – because the partnership has been strong and the engagement and we've been delivering to what we said. Anytime you do that with the customer it opens up opportunities. And so again, you kind of nailed it between OCTEON for the transport, full ASIC capability for DFE and the ability to add even some Marvell IP and their massive MIMO Ethernet front-haul connectivity, it's a pretty compelling portfolio.
So we're hopeful to continue to win more business across the board and certainly, the best way we can do that is live up to our promises and execute really thoroughly to what we commit to, and that's what we're intending on doing at least with this current opportunity.
Well, thanks for the insights there. And then on storage, last quarter, I believe you guys were still shipping under consumption in the storage business and so still some room to get back to normal consumption levels on top of that I think that you're ramping pretty strongly into your two customers, nearline 16 terabyte HDD programs and that's looking to be a pretty strong tailwind for them in the first half of next year. Your enterprise SSD platforms are doing well, cloud spending is accelerating. So help us understand why the business on a sequential basis doesn't seem to be growing in the January quarter.
Sure. Yes, I think you nailed the growth drivers pretty well and certainly, I think the effort candidly that we put in over the last three years to really repivot the storage business and focus on the cloud and enterprise and get those design wins and get them – products developed and get them ramped that's all starting to play out, and that's helping offset what we see is pretty continued weakness both in the client side.
I think there's been some pretty well-telegraphed challenges on the CPU front in terms of ability to build enough notebooks that hasn't helped anybody certainly the drive companies, and the end OEMs. I think also the export restrictions in the bands relative to some of the Chinese OEMs and the video surveillance area. Clearly those video surveillance cameras and devices have a tremendous amount of storage attached to them, and that's also got some impact that's a derivative impact. We don't sell directly into those companies, but we sell to drive companies that may or may not have business there and that's certainly a weakness too.
So I think, we see these edge and client markets being fairly soft and then offsetting it or even in the case of Q3 more than offset it was strong growth on these product cycles and the tailwind that you mentioned. And so what we're going to continue to do is continue to execute our strategy. We've been putting all of our R&D efforts into the markets that are growing today, and where we can control our destiny on research and development and product execution, we will. Some of these other things like PC CPU shortages or government bans on, it's hard for us to influence those. All we can do is manage around it, and where we can control our destiny.
Thanks, Matt.
Yes.
Thank you. Our next question comes from Timothy Arcuri of UBS. Your question, please.
Thanks a lot. I guess the first question, Jean is maybe if not some qualitative at least some qualitative commentary around what the, what the long-term financial model will look like for the new company. You were saying before you could get about 66% gross margin and sort of 30% to 32% OpEx target. I know gross margin, you said, will be back above 63% exiting fiscal 2021, but where should we kind of think about sort of the upside beyond the 63% and sort of where OpEx can go in with sort of the new company here.
Yes. Tim, thanks for the question. I think in the longer-term, we are quite confident we will continue to drive the gross margin expansion with the combined portfolio. Especially, we are very much focused in the infrastructure market as Matt mentioned, we have very broad portfolio.
Overall, as I said earlier, gross margin impacted by both level of revenue and the product mix. When we recover from the current revenue level and continue to expand, especially with our 5G ramp and other new product cycles unique to Marvell in the second half next year, we do see with the revenue level going up that will help improve our gross margin and the progress is toward that target level.
On the operating margin side, we continue to focus on operating margin expansion. So our focus is definitely is the same operating margin target like before. We continue to manage the SG&A costs to make sure it's below 7% of revenue, and R&D, the investment is very important for us to drive the longer-term profitable growth, and so we continue to invest in our business. So once our revenue level going up, our model is highly leverageable, we will get to our operating model quickly.
Got it, Jean. Thanks for that. And then I guess, Matt, maybe can you give us some sense, I know maybe you don't want to give us a number, but of the $288 million in storage, can you give us a sense of how much is SSD sort of on a relative basis and maybe from a competitive point of view, how much do you worry about the threat of your customers who continue to try to actively displace you there, I mean how real is that? Thanks.
Sure. Yes, Tim, I think the way to think about our storage business is, we actually – I think you followed us. We made the pivot which we felt was more appropriate back when we closed Cavium. We added fiber channel in. We did our Analyst Day, and we really decided to talk about the business more in terms of end market orientation, enterprise and datacenter versus client just because from an R&D standpoint, that's where we invest.
We look at it from an end market application point of view, not so much how much do we want to spend in X, Y, Z product line. So again, I think the commentary, I gave to Harlan, a few questions ago was probably the way you should think about it, which is we're executing well to our strategy on the enterprise and the datacenter side and that continues to grow as a percentage of our total storage, which is encouraging.
On the competitive threat side, I'd say a couple of things, one is at the Analyst Day, we did about a year ago, what we said was look, we acknowledge that verticalization is occurring particularly in the client area and in particular, in the – we believe that the low end portion of the PC market was probably the place where we would see the most traction from our NAND customers, who wanted to go vertical, and we've seen that play out. Now, we weren't necessarily putting much if any R&D in that area, and so we've just been managing that trend.
What we're encouraged by and that's actually why we purposely called out the product we just introduced our first 12-nanometer product is, it's a standalone controller and I would call it for the higher end of the market, but really, what we are able to leverage is for the DIY opportunities.
It's the same process node, same IP set just customized very specifically for an application and that market and that business this DIY concept, which was really more of a concept actually, when we first talked about in 2017 that become real, and now we're even saying in Q3 of next year, we expect a sharp revenue ramp as a result of that business model catching on in the cloud and beyond.
And I think that trend, we're going to continue to see and we look like a really good choice there because the end customers want a supplier who actually has deep experience working with other people's NAND. And so when you – if you are end OEM and you wanted to get a NAND supplier with their own controller team to go start working with everybody else's NAND that probably isn't going to happen because nobody really wants to share their inner workings of how their product interoperate.
So we think the DIY trend is real, it's growing in importance to us. And so we're going to take a standard product approach that gets us the IP we need so we can go do that. But to the extent we can produce a standard product out of it, that's got broad market appeal that can fit certain swim lanes or SKUs of our end customers where they don't have the capability or the capacity of the design team internally to do it than we would be a good supplier.
So the bottom line as we see our SSD business growing through DIY and others on an ongoing basis from starting at this point going out through next year and beyond, driven by some of these trends that I articulated.
Awesome, Matt. Thanks so much.
Okay. Thanks.
Thank you. Our last question comes from Chris Caso of Raymond James. Your line is open.
Yes. Thank you. Good evening. First question is regarding the Fusion radio head win that you talked about, if you could give a little more color on that and assuming that's incremental to the revenue run rates you're running in 5G right now, and perhaps talk about some timing and magnitude of when that ramp occurs?
Yes. Sure, Chris. So the way to think of it is we had a customer take our prior generation fusion and because of the architecture and the scalability of the programmability, right they were actually able to implement it and go into production fairly quickly, and so that's in our current run rate model today.
Just to clarify the next-gen that we won adds more functionality, and so as a result, actually is a higher ASP, so when that product kicks in, which would be in the early part of 2021 in production, I mean, we're going to get the whole development done in calendar 2020. So when that ramps in calendar 2021, certainly there is some incremental revenue to us.
What I would say is the dark horse is what's the adoption of massive MIMO going to be. And I think to the extent that massive MIMO does better that's certainly going to be an opportunity. And then beyond that, this is a trend we see. So this wasn't just one customer who had their own kind of concept about how to do this.
We very much see the trend of purpose-built optimized silicon in the radio head not only at the customer that we're engaged with today, but at multiple other customers and including a very strong engagement we have already with a second customer today to do something very similar. So it's a trend we see it, and we think we're very well positioned to take advantage of it with future design wins beyond our lead customer with this type of architecture.
Thank you. And as a follow-up on the Aquantia business, you talked about getting back to $100 million revenue run rate in that business. Is that just a function of the customers burning off the inventory and getting back to normal shipments, and what are the prospects of growing that business beyond the $100 million level?
Yes. Chris, I think it's a combination. One is inventory digestion and getting back to a normalized run rate. And then the second is, is new design wins that Aquantia had already secured prior to the acquisition, their pipeline was pretty strong. When we acquired them, and certainly, I think it goes without saying with the reception for the combination has been very strong both by the enterprise customer base as well as the automotive customer base.
I mean in the enterprise side, the view is it creates a much stronger scaled up PHY team with best-in-class technology from both Marvell and Aquantia, and we've got that team fully integrated now under Faraj's leadership. And on the automotive side, where we have really by far the most compelling end-to-end portfolio from all the way from 100-megabit PHYs to roadmap to 10-gig and beyond over copper with switches and multiple configurations, port counts and a substantially scaled up R&D team to go deliver to what that market needs.
So I think, I'd just say it's a combination of wins that they already had, new wins we're getting, and then certainly the inventory correction that a lot of folks went through from the first half to the second half that certainly plays itself out as well as we head into the end of the year here.
Got it. Thank you.
Yes.
Thank you. At this time, I'd like to turn the call back over to Ashish Saran for closing remarks. Sir?
Thank you, everyone for joining us today, and we look forward to seeing you at the upcoming Barclays Technology Conference in San Francisco.
Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.