Seeking Alpha
We cover over 5K calls/quarter
Profile| Send Message|
( followers)

Cavium (NASDAQ:CAVM)

Q2 2014 Earnings Call

July 30, 2014 5:00 pm ET

Executives

Angel Atondo - Marketing Manager

Syed B. Ali - Founder, Chairman, Chief Executive Officer and President

Arthur D. Chadwick - Chief Financial Officer, Principal Accounting Officer and Vice President of Finance & Administration

Analysts

Blayne Curtis - Barclays Capital, Research Division

N. Quinn Bolton - Needham & Company, LLC, Research Division

Anil K. Doradla - William Blair & Company L.L.C., Research Division

Hans C. Mosesmann - Raymond James & Associates, Inc., Research Division

Richard E. Schafer - Oppenheimer & Co. Inc., Research Division

Kevin E. Cassidy - Stifel, Nicolaus & Company, Incorporated, Research Division

Brian T. Modoff - Deutsche Bank AG, Research Division

Matthew D. Ramsay - Canaccord Genuity, Research Division

Sanjay Chaurasia - Nomura Securities Co. Ltd., Research Division

Ruben Roy - Piper Jaffray Companies, Research Division

Joseph Moore - Morgan Stanley, Research Division

Betsy Van Hees - Wedbush Securities Inc., Research Division

Sujeeva De Silva - Topeka Capital Markets Inc., Research Division

Operator

Good day, everyone, and welcome to the Cavium Incorporated Second Quarter 2014 Earnings Conference Call. Today's conference is being recorded. At this time, I would like to turn the conference over to Angel Atondo. Please go ahead.

Angel Atondo

Thank you. Good afternoon, everyone, and welcome to Cavium's Second Quarter 2014 Financial Results Conference Call. Leading the call today are Mr. Syed Ali, President and CEO of the company; and Art Chadwick, Vice President and Chief Financial Officer.

Before we begin, I'd like to remind you that various remarks that we make on this call, including those about our future financial results, including revenues, gross margins, operating expenses, design wins, product plans, our competitive situation, market trends and our anticipated growth and profitability, all constitute forward-looking statements for the purposes of the Safe Harbor provisions under the Private Securities Litigation Reform Act. These forward-looking statements and all other statements that may be made on this call that are not historical facts are subject to a number of risks and uncertainties that may cause actual results to differ materially. We refer you to our most recent Form 10-K and Form 10-Q filed with the SEC, in particular, to the section entitled Risk Factors, and to other reports that we may file from time to time with the SEC for additional information on factors that could cause actual results to differ materially from our current expectations. These forward-looking statements speak only as of the date hereof, and we disclaim any obligation to update these forward-looking statements.

In addition, Cavium reports gross margin, operating expenses, net income from operations, net income and basic and diluted net income per share in accordance with the GAAP and also on a non-GAAP basis. Management believes the non-GAAP information is useful because it can enhance the understanding of the company's ongoing economic performance, and Cavium, therefore, uses non-GAAP reporting internally to evaluate and manage the company's operations.

Cavium has chosen to provide this information to investors to enable them to perform comparisons of operating results in a manner similar to how the company analyzes its operating results. The full reconciliation of the GAAP to non-GAAP financial data can be found in our earnings release issued earlier today, and we ask that you review it in conjunction with this call. Additionally, the information we provide on this call regarding sales by market involves certain management judgment as to which market each sale is assigned, and you should consider that when analyzing such information.

I will now turn the call over to Syed Ali. Syed?

Syed B. Ali

Thanks, Angel, and thanks to everyone for joining us today. In brief, Cavium's second quarter revenue was $90.7 million, up 9% sequentially, and up 22% year-over-year. This was our ninth quarter in a row of sequential growth. Non-GAAP gross margins were 64.5%, and non-GAAP net income was $19.9 million or $0.35 per diluted share. Art will discuss our Q2 financial results and Q3 guidance in more detail shortly.

In Q2, we saw strong growth in our core service provider, enterprise and data center markets. Both the service provider and enterprise data center markets grew sequentially. Sales into the broadband and consumer market were down. In the service provider market, sales into wireless infrastructure continued to be extremely strong. Sales into the wireline and telecom were soft, although we did see the beginning of a turnaround at certain customers. Growth in the enterprise and data center markets was primarily driven by customers in the data center, enterprise networking, and security markets. Sales into the broadband and consumer were down, as sales into discontinued legacy consumer segments continued trailing down.

Now as usual, we will move on to our Q2 design win traction, Q2 product announcement highlights, our product development status and the market environment that we see for Q3. After that, I will cover details of the Xpliant acquisition we announced today.

Q2 was another very strong design win quarter. Design win momentum continued, with contributions from across our product line spectrum. Once again, our design wins were across the enterprise, data center, wired and wireless infrastructure and low-end broadband and control plane applications. Overall, we saw strong continued design win momentum in both our traditional, as well as in new markets. I would especially like to highlight some areas where we are seeing significant increase in design win momentum.

With the widespread sampling of our 28-nanometer OCTEON III 48-core device 78XX and CN -- and our CN70XX, our lower-end 4-core device, goes into production, we are seeing excellent and growing design win momentum for our OCTEON III product line. We observe continued adoption of our low-end OCTEON III 70XX product family across a wide range of applications. In addition, we are seeing a new set of design wins for our high-end 48-core OCTEON III CN78XX product line in both the wired and wireless service provider and enterprise segments.

In the security segment, our design wins continue to increase for NITROX, OCTEON and the highest level security for these [ph] products. Once again, these designs are driven by market demand for a robust line rate high-performance security in next-generation firewalls, unified threat management appliances, DPI blades, cloud servers and secured https web servers. This validates our assessment that the need for pervasive intelligent end-to-end security across the network is essential. And Cavium is providing market-leading solutions required for delivering the highest level of security for the enterprise and service provider infrastructure.

We are seeing continued traction of our LiquidIO product line, where we won 2 new design wins for our current generation product, which should generate meaningful revenue for us in 2015. We also have continuing evals at multiple Tier 1 customers. Strong design win momentum is being driven by strategic initiatives at cloud service providers and data centers to provide high-performance networking and software-defined networking services with increased throughput, high-performance security, and reduced latency. We also had a successful public launch of ThunderX at Computex in June, with very positive press coverage. The ThunderX is a new game-changing category of processors that are workload optimized for a range of applications in the cloud and data center. The 2.5 gigahertz, 48-core ThunderX is the world's highest-performance, low-power 64-bit ARMv8 SoC family of workload-optimized processors, with a range of SKUs and form factors for high-performance volume, compute, storage, secure compute and networking-specific workloads.

ThunderX family is our fifth generation of multi-core processor from Cavium, targeting high-performance volume servers and appliances for large data centers and cloud infrastructure. This product family is based on a highly efficient, full custom processor core, designed by Cavium in 28-nanometer process under architectural license from ARM. It is fully compliant with the ARMv8 architecture, as well as ARM's server-based system architecture standard. While bringing to market dramatic enhancements that include: The first ARM-based SoC that scales up to 48 cores with up to 2.5 gigahertz core frequency; the first ARM-based SoC to be fully cache coherent across dual sockets, using Cavium's coherent processor interconnect; integrated I/O capacity, with hundreds of gigabits of I/O bandwidth; 4 DDR3 or DDR4 72-bit memory controllers, capable of supporting 2,400 megahertz memories, with up to 1 terabyte of memory in a dual socket configuration blade; hundreds of integrated hardware accelerators for security, storage, networking and virtualization applications; standards-based low-latency ethernet fabric, interconnecting thousands of ThunderX nodes in 2D and 3D configurations and enabling fabric monitoring and SLA enforcements, with awareness and policy enforcement for virtualized networks.

Another key feature is the virtualization everywhere, with Cavium virtSOC technology, which enables full system virtualization for low latency from virtual machine to I/O. Overall, we believe that this delivers best-in-class performance per watt and performance per dollar for the target applications. It should be pretty evident from our product announcement that we have a very differentiated product line that is targeting the volume sweet spot of the large data center CPU markets. Concurrent to our announcement, we had a number of ecosystem partner announcements also in Q2. By working closely with ecosystem partners, we are working diligently to enable ThunderX for efficient rack level deployment and application optimization from Day 1, with fully developed firmware, Linux Kernel and operating systems, development tools, cloud infrastructure and scale out application software. Cavium's engagements in the software ecosystem include membership in the UEFI Forum, open source and industry groups such as OpenStack Foundation, the Linux Foundation, the Xen Advisory Board and Linaro.

Cavium plays an active role in upstream's Linux Kernel development that ensures standard support for all our ThunderX product families and product features and demonstrated support for Linux OS distributions. These include Canonical Ubuntu, Red Hat Fedora, openSUSE and MontaVista. Also, we have efficient and scalable virtualization support via KVM and Xen and full support for development environments, including GCC suite and Java. Many of these are already available today for the ThunderX family.

Customer traction for our ThunderX product line continues to be extremely strong, as we have a very differentiated and competitive product compared to both incumbent, as well as other announced ARM vendors in terms of integration, performance and power efficiency for our target markets. We have several ongoing developments with lead customers who are developing boards and systems based on ThunderX. Additionally, we are working with a number of other ODMs to build a range of board and system form factors, which will all be ready next quarter to ship to the customer base.

Additionally, another highlight of Q2 was that we have completed extensive benchmarking of our existing 48-core OCTEON III 78XX product. The benchmark results we are getting are extremely good, and are at or higher than our expectations. This significantly increases the confidence level that our 48-core ThunderX ARMv8 processor will perform at or better than our aggressive design targets.

Now I'd like to give an update on our 28-nanometer product development efforts. As we have talked about before, we have widely sampled our flagship 48-core OCTEON III 78XX product family to multiple Tier 1 customers worldwide. The ThunderX development has been completed, and we are on target to have silicon back off our flagship 48-core ThunderX 88XX line next quarter. Our low-end 2- to 4-core OCTEON III 70XX and 71XX family of products have completed product qualification tests and is starting volume shipments this quarter. Our next-gen LiquidIO2 product line is in full execution mode and is targeted to sample in Q1 2015.

In earlier calls, I have mentioned that the LiquidIO2, which is our next-generation product, is seeing strong customer traction and will address a significantly larger market. Now I'd like to move on and give a brief outlook on the market environment that we see for Q3.

We now expect another strong, broad-based sequential growth quarter in Q3. We expect sequential growth in the service provider market. In the wireless infrastructure market, despite some softness in China, we are seeing good growth in other geographies such as India, Japan and Europe, which are more than offsetting the much talked about softness in China. We are also seeing a nice uptick in the wired infrastructure markets, which had been soft for the last few quarters. We also expect our enterprise data center markets to be up sequentially, driven by our traditional markets. In addition, this quarter, after several quarters of sequential declines, we expect our broadband and consumer business to be up in Q3, driven by sales of our low-end 1- to 4-core OCTEON products, including our first volume production shipments of our 28-nanometer OCTEON III family. So in summary, for Q3, we expect all segments to be up.

I would now like to turn to the acquisition announcement we made today. We are extremely pleased to announce that we have entered into a definitive agreement to acquire privately held Xpliant. Xpliant is a fabulous chip company based in San Jose, California and Bangalore, India and is a provider of a family of high-performance, high-density switching products, targeting a broad range of switching applications for the data center, cloud, service provider and the enterprise markets. The Xpliant team is a world-class team of switching industry veterans who have previously delivered several generations of industry-leading switching solutions to market at a range of leading switching system and silicon vendors. The rationale behind our acquisition of this exciting technology and products is compelling. The steep growth of data over both 16 mobile networks, which is being fueled by consumer and enterprise mobile applications, social networking and content streaming, is driving the need for increased brand -- bandwidth, intelligence, and flexibility in both networking infrastructure and data centers. We see a major architectural revolution playing out in data centers, service providers and enterprises, with a move away from fixed function architectures to flexible software-defined everything: compute, networking and storage. We are addressing the compute and data processing elements for this new architecture with our recently announced Thunder and our portfolio of existing products, such as OCTEON, LiquidIO, et cetera. The other important part of this solution are the switches, which connect all the elements together, both within the system and from system to system. Current switch solutions are built on legacy fixed function architecture that will not be able to address the new market trends and requirements in data center, service provider and the enterprise markets. Current high capacity and bandwidth switching silicon is hardwired to only offer a limited set of fixed function capabilities that do not enable the software flexibility required to deploy true software-defined networking, resulting in extremely inefficient networks. Another disadvantage of the current legacy fixed function architecture is that system vendors cannot differentiate and optimize their product features and capabilities for a range of target market applications.

Xpliant has developed a family of switching silicon solutions with world-class speeds and feeds of 10, 40 and 100 gigabit, as well as world-class port density and throughput, ranging from multi-hundred gigabit to multi-terabit. But in addition, these products have been designed from the ground up to specifically address the challenges of throughput and flexibility, driven by the needs of next-generation networking infrastructure and software-defined networking. Xpliant's innovations in switching architecture will enable increased intelligence and flexibility beyond what is available today in commodity Ethernet switch silicon, while delivering market-leading bandwidth, throughput and scalability, along with unprecedented flexibility and programmability. All of these characteristics are highly desirable to efficiently enable virtualized data center and software-defined networking without compromise. Xpliant's flexible and programmable architectural solutions are designed to enable OEM customers to bring to market customized and differentiated products optimized for their target markets. The TAM for the switching silicon market is over $1 billion.

Xpliant's family of switching solutions are also very highly synergistic with Cavium's existing infrastructure product offerings. And we will be able to leverage our strong customer relationships with leading infrastructure OEMs and end-users to drive product adoption. We also believe that the Xpliant switching solutions will enable Cavium to significantly increase our share of bill of materials in the data center, service provider and enterprise markets. Additionally, these products will allow Cavium to offer complete end-to-end solutions for compute, networking and storage, optimized for best performance, cost and power for next-generation virtualized software-defined infrastructure.

A classic example of the type of systems we can enable is under development at one of our large customers. The system uses multiple OCTEON III chips for low-latency packet processing, multiple Xpliant chips for switching, and multiple ThunderX chips for cloud software applications. Another example is in standalone enterprise data center switches, where we could also provide CPU for the control function. Xpliant's products are seeing strong customer traction with solid lead customers and a very robust opportunity pipeline. The products are targeted to start sampling in Q4, which is next quarter, and start production shipments in the second half of 2015. In summary, we are continuing to execute on our existing products. We are rolling out next-generation products like OCTEON III. We are in development of next-generation products like LiquidIO2. Our design win rates for our existing products are at record levels, which will drive higher market share in our currently served markets. Additionally, we now have 2 new exciting product lines: ThunderX and the Xpliant line, that address large markets and will enable us to offer a world-class portfolio of infrastructure solutions to our customers.

On that note, I would now like to turn the call over to Art Chadwick, who will provide a detailed discussion of Q2 financial results and guidance for Q3. Art?

Arthur D. Chadwick

Great. Thanks, Syed, and thanks to all of you for joining us today. I'll first go through Q2 financial highlights and then provide guidance for the third quarter of 2014. As Syed mentioned, Q2 was another record quarter for us, with revenue of $90.7 million, up 9% sequentially, and 22% year-over-year. Sales into the enterprise and service provider market were $84.0 million, or 93% of sales, up 11% sequentially, and 33% year-over-year. Sales into broadband and consumer were $6.7 million, or 7% of sales, down 12% sequentially, and down 40% year-over-year, due to our exit from certain consumer businesses.

In Q2, we again had 3, 10%-plus customers. Sales to those 3 customers accounted for 51% of total sales this quarter. Non-GAAP gross margins were 64.5%, which was on the high end of our guidance. Non-GAAP operating expenses were $38.3 million, a 3% sequential increase over Q1. Non-GAAP R&D expenses were $26.4 million and non-GAAP SG&A expenses were $11.8 million. Non-GAAP operating income was $20.2 million, and that was up 17% from Q1, and up 46% from the same quarter last year. Non-GAAP operating margins were 22.3%, up 160 basis points from Q1. Income tax expense was $311,000, less than 2% of non-GAAP income.

On June 30, we exercised our option to acquire Xpliant. And today, we signed a definitive agreement to acquire the company. This triggered Xpliant to take a $13.9 million noncash charge to revalue their investor notes payable, reflecting the increased amount that will be paid to note holders when the company is acquired, and that charge is included in our non -- I'm sorry, that charge is included in our consolidated GAAP results. Our consolidated net GAAP loss in Q2 was $11.0 million.

Q2 non-GAAP net income was $19.9 million or $0.35 per share, up 16% sequentially, and up 58% from a year ago quarter. Q2 non-GAAP results exclude $30.9 million in non-GAAP adjustments, which include Xpliant charges, stock-based compensation expense, amortization of acquired intangible assets and other expenses as detailed in our reconciliation between our GAAP and non-GAAP results, as reported in our press release. Quarter end AR was $52.1 million, that's down $4.7 million from Q1, even though sales were up 9% this quarter. DSOs in Q2 were 52 days, down very nicely from 61 days in Q1. However, our longer-term DSO targets still remains at 60 days. Inventory at the end of the quarter was $56.7 million, up $7.7 million from Q1, from inventory being staged to support expected sales growth during the second half of this year.

We ended the quarter with a strong balance sheet. Our net cash balance was $140.6 million, and that was up $16.5 million from Q1, and that is after we made an additional $5 million cash investment in Xpliant during the quarter.

We issued a press release today with details regarding our planned acquisition of Xpliant. The total cost to acquire Xpliant is projected to be approximately $90 million, which includes payments to stockholders, note holders, employees, assumed liabilities, interim operating expenses and other costs. Cavium was an earlier investor in Xpliant and provided $15 million in funding through June of 2014, which is included in that $90 million cost estimate. The remaining $75 million will be comprised of approximately $40 million in cash and $30 million worth of Cavium stock. Based on our current internal forecast, we believe that a large majority, if not all, of that cash amount can be funded from Cavium positive cash flow between now and close, which we expect will happen by the first quarter of next year. By the time the deal closes, Xpliant operating expenses will have decreased from current levels, as they phase out a large number of contractors and back-end services used to develop their current family of products. We will integrate the company and gain cost efficiencies by leveraging our existing infrastructure. We will hire their employees in California and India. The acquisition of Xpliant will add incremental expense to our run rate. However, a significant portion of those reduced expenses will be absorbed as part of our planned operating expense growth next year. Additional incremental expenses are expected to dilute non-GAAP income in the first half of next year by just a few cents. But Xpliant revenue should make the acquisition incrementally neutral to slightly positive in the second half of 2015.

I'd now like to provide more specific guidance for the third quarter of 2014. We are starting Q3 with record backlog, and expect continued strong growth in our core enterprise service provider and data center markets. We expect third quarter revenue will be between $96 million and $98 million, which at the midpoint, would be 7% sequential growth. Q3 non-GAAP gross margins are expected to increase 50 to 100 basis points sequentially, due to product cost reductions and improved product mix. Q3 non-GAAP operating expenses are expected to increase between 3% and 5% sequentially, which at the midpoint, would be approximately $40 million for the quarter. Interest and other nonoperating expenses are expected to be approximately $200,000. Income taxes in both Q3 and Q4 are expected to be approximately 4% of non-GAAP income, increasing to between 7% and 9% of non-GAAP income in 2015. The Q3 non-GAAP share count is expected to be approximately 58 million shares. And based on those assumptions, Q3 non-GAAP EPS should be between $0.38 and $0.40 per share.

And on that note, I'd like to hand the call back to the operator for our Q&A session. Operator?

Question-and-Answer Session

Operator

[Operator Instructions] And we'll first hear from Blayne Curtis with Barclays.

Blayne Curtis - Barclays Capital, Research Division

Maybe just from a high level, Syed, if you can talk about the strength in the guidance. You said pretty much growth across the board. You mentioned offsetting some of the China weakness elsewhere. Is that 3G ramps on the wireless side that are offsetting, you mentioned some other regions? And then if you had any perspective on when this China pause, how long it will be, maybe start there?

Syed B. Ali

Blayne, now essentially, this quarter, our growth in wireless infrastructure is being driven by some new 4G deployments in countries like India, Japan and Europe. But also, as I talked about in the past couple of calls, we really started shipping 3G in volume at the beginning of this year, it was roughly in Q1. So it's really a combination of the 2 events that is helping to more than offset the weakness in China.

Blayne Curtis - Barclays Capital, Research Division

Okay. Then maybe on the enterprise side, that seems also strong. You mentioned last quarter that the new products were ahead of that plan to be 10% of revenue. Is that still on track, and how is LiquidIO progressing within that?

Syed B. Ali

In the enterprise and data center markets for Q3, it's a much more significant, wider range of customers and products that are providing the growth. So it seems pretty healthy for us in Q3.

Blayne Curtis - Barclays Capital, Research Division

Okay. And then maybe just finally, on Xpliant, you mentioned that you already had some customer interest. It seems like silicon in Q4, could you just talk about -- I'm assuming you're able to show them in a PLD or such, just customer buy-in, is the new approach, what kind of challenges are there to get people on board? And just a perspective on the design cycle. I'm assuming these are more dynamic cloud guys and such. You mentioned revenue is really a second half next year. If you could just -- any comment on there would be appreciated.

Syed B. Ali

Yes, that's a good question, Blayne. So one of the things that I'd clearly like to point out is first and foremost, this is a family of world-class switches, that are -- have great speeds and feeds and have great throughput. So they are world-class in terms of both port density and throughput. So they will -- even if you keep the functionalities similar to the functionality that they're using today, these will be world-class products. On top of that, they have a whole range of programmability and flexibility that will allow individual OEMs to actually customize the product, and really optimize it on a per end market or per end application basis. While the design cycle for this, we do -- as I said, we've got our lead customers, plus we have a very robust pipeline. And we expect our first revenues for this product line to start in the second half of 2015. So the typical design cycle, I would say is somewhere in the 1-year range, kind of give or take, 9 months to a year range, but the early customers are already -- will have boards ready, will have software ready, very, very similar to the ThunderX model, where we have worked with lead customers even before silicon has come, we've given them SDKs, so that their software development can progress. So it's -- the cloud doesn't start just when the silicon comes for the early customers. It has started a few quarters before.

Operator

And next we will hear from Quinn Bolton of Needham & Company.

N. Quinn Bolton - Needham & Company, LLC, Research Division

Just wanted to follow up on Blayne's question there on the base station side of the business. Obviously, you talked a little bit about China weakness, but it seems to me that there may be a perception that your exposure to China LTE may be bigger than it is. Is there any way you can quantify what percent of your base station business actually sells into China LTE? I mean, my sense is that it's just a relatively small percentage, but I was hoping you could -- you might be able to clarify that.

Syed B. Ali

That's a good question, Quinn. Essentially, when you take a look at our entire wireless infrastructure business, especially in 2014, it's a very nice mix of both 3G and 4G. It's a very nice mix between China and the other geographies. So it's very difficult for us to actually give an exact number, because for the China Mobile section, primarily because we have similar chips or the identical chip, being shipped into Europe, into India by the same OEM. So it's a little bit difficult to come up with an exact number. But if I take a slag at it, it's probably in the 1/3 range of our total wireless infrastructure revenues.

N. Quinn Bolton - Needham & Company, LLC, Research Division

Okay, okay, great. And then just wanted to clarify on the ThunderX. I think you said in your prepared script that you have silicon back in Q4. Has that already taped out or if it hasn't taped out, can you give us a sense of what that tape-out schedule is?

Syed B. Ali

Yes, so essentially, we have sent over the database, it's going through final checks at the foundry and we will have silicon next quarter.

N. Quinn Bolton - Needham & Company, LLC, Research Division

Okay, great. And then still on track for sort of design win and announcement, probably late this year, early 2015 for ThunderX?

Syed B. Ali

Yes, I think that will be an approximate timeframe, yes.

Operator

And next we'll hear from Anil Doradla of William Blair.

Anil K. Doradla - William Blair & Company L.L.C., Research Division

Syed, a couple of questions. I think Broadcom has their Trident platform. We saw some news coming out of the acquisition of Cortina by Inphi this morning. So there's a lot of stuff going on in the, kind of the switching side. You talked about a $1 billion in addressable market. Can you just frame what's going on from a, kind of competitive landscape? How long would this take to say get to $100 million in revenue, and how would you handicap it based on many of your products that you've had? Would this be on the quicker side to get to 0 to $100 million? Any color would be very helpful.

Syed B. Ali

Anil, that was 7 questions, I calculated. Let me just kind of give you the high level of puts and takes on this. Essentially, this product line ships into the standard switching market, obviously. And the major players there today are primarily Broadcom at the mid-range and high end, and Marvell at the low end. And obviously, there is some Intel through their Fulcrum acquisition that is there in somewhere in the mid-range of the products. So our product line primarily focuses on 10, 40 and 100 gig, which really is kind of the mid-range to high end of the market. We are not going after kind of the low-end, 1 gig type of commodity silicon market here. So we believe that we have -- that the Xpliant product will be a very compelling product and our customer feedback has attested to that fact that this is a very competitive product in the market. So essentially, we expect this, like I said, to start revenue shipments in the back half, again, on the back of our lead customers. And really start accelerating in '16 and '17. So since this is an infrastructure product, but the design cycles, like I said, are in kind of the 1-year range instead of, in our traditional kind of main CPU, big iron type infrastructure, it could be as long as 2 to 3 years. Here, if the product is a standalone switch, design cycles are a lot shorter. But if it is used in an embedded product, like a router or a cloud RAM type system, in that case, the design cycles will be like our traditional cycle. So the answer is probably something in between, based upon the exact end application.

Anil K. Doradla - William Blair & Company L.L.C., Research Division

Great. And Art, did you say it's going to be a couple of pennies of dilution in 2015, Xpliant, did I get that right?

Arthur D. Chadwick

So what we said in the first half of next year, we believe it will be a few cents dilutive on a non-GAAP basis. In the second half of the year, we believe it will be neutral to, perhaps slightly positive, because we'll start shipping some revenue -- excuse me, revenue in the back half of the year.

Operator

Our next question comes from Hans Mosesmann of Raymond James.

Hans C. Mosesmann - Raymond James & Associates, Inc., Research Division

Syed, can you give us a little flavor of what's happening with NEURON and Fusion? You didn't mention it, or I didn't catch it, in your prepared remarks.

Syed B. Ali

Yes, we have so many things to talk about that -- the script -- my script, it was probably the longest script I've read in quite a long time. But yes, so essentially, on the OCTEON Fusion side, we are in trials at several service providers and we expect to start seeing more meaningful revenues for both the OCTEON Fusion and the NEURON starting 2015. So we're getting design wins in both sides of the equation. With NEURON, the design cycles are much longer, primarily because it goes into very high-end equipment. Whereas with the Fusion, they're a bit shorter.

Hans C. Mosesmann - Raymond James & Associates, Inc., Research Division

Okay. And then just going back to this Xpliant, how does the -- your OCTEON III and ThunderX portfolio of products play into that, the adoption of this particular switch technology? The reason I'm asking is that your larger -- your competitor may not have an ARM product in the market until 2016 or who knows when.

Syed B. Ali

Yes, so the way to look at it, Hans, is that essentially, let's take 2 or 3 different classes of systems, right? One is just a standard switch, right? With the 10, with multiple 10 gig, 40 gig, 100 gig ports and it could be a single pizza box type 1U, 2U type system, or it could be in a chassis-based system. So obviously, in those type of systems, you have both the CPU and a memory. In a lot of the high-end systems actually, X86 is used as the control for these -- for the high-end switches. So there, obviously, ThunderX has a very, very good opportunity to play. The second is, in more kind of networking centric systems, where there are EDGE routers, CMTS headends or what have you, essentially, the switches are an important element and a piece of the total solution. So -- and in the data center, obviously, where we're trying to drive ARM, there'll be a lot of switching silicon that is also used. So this -- plenty of synergies that we see across our end markets to really deliver total solutions. And also, a lot of the potential customers that we have, we are already selling into, so we have very good customer relationship there, which should help the market adoption.

Operator

And next, we'll hear from Rick Schafer from Oppenheimer.

Richard E. Schafer - Oppenheimer & Co. Inc., Research Division

I had a quick follow-up on the base station question. Can you guys say if you're still on track to exit this year at that sort of $100 million combined run rate you've talked about? And what does that run rate look like if we look into 2015? Do we kind of cap off at $100 million? Or does it continue to grow and maybe an idea where it grows to?

Syed B. Ali

The wireless infrastructure market for us consists of more than just a base station component. There is base stations, and for 3G, there is radio network controllers. There is GSM, which is kind of the backhaul, if you will. Whereas in 4G, your have base stations and EPC. So overall, for our wireless infrastructure market, hitting those type of numbers this year is definitely on the cards, but there is not a cap to that, because as overall -- second piece of the equation, which is kind of the backhaul aspect of it, the cloud RAM aspect of it, where we have very good designs also.

Richard E. Schafer - Oppenheimer & Co. Inc., Research Division

Got it. And then just a quick follow-up, also on the NEURON Fusion comments you just made. If we're looking at 2015 and we've got sort of a, maybe a potential proxy in the way that LiquidIO launched in the third quarter of last year, and we've obviously seen that ramp. Is there a way to kind of gauge -- I mean, could Fusion or NEURON or both kind of follow a similar type trajectory if we're trying to model next year's contribution?

Syed B. Ali

Yes, I think no, not as fast as -- I don't think either one of them will be as fast as LiquidIO, because LiquidIO is a card solution. You just kind of crack it in and the development times are very short and the time to revenue is very short. So these 2 will be generally longer, but we expect Fusion revenues to ramp faster than NEURON, because NEURON again goes into very large equipment, where they take a long time to go to production and the ramps are not quite as fast as a server in a data center.

Richard E. Schafer - Oppenheimer & Co. Inc., Research Division

Got it. And then my last question is just on the low core count OCTEON III. Can you size or give us an idea of the size of that TAM or that market potential opportunity next year for you guys, and maybe an idea what you expect your share to kind of roughly be in that new opportunity for you?

Syed B. Ali

We believe that this market -- and I think we have talked about this in earlier calls -- as over $500 million. It's -- that is the existing market today. And obviously, for a product line like this, our goal is -- first is, can we get to $100 million in revenue? And we believe that over time, the answer is, yes.

Operator

Kevin Cassidy, Stifel.

Kevin E. Cassidy - Stifel, Nicolaus & Company, Incorporated, Research Division

On the broadband and consumer, it seems to hit a bottom in the second quarter, you're saying growth in the third quarter. Is that just going to be seasonal or should we expect revenue to grow from here?

Syed B. Ali

No, like I said, the consumer has kind of turned off. Obviously, in any product, there is -- there can be some slight seasonality. But from the bottom in Q2, if you kind of take a look at back half of '14 versus back half of '13, or even back half of '14, we should see -- or the front half of '14 -- excuse me, front half of '15, we should show growth on that area.

Kevin E. Cassidy - Stifel, Nicolaus & Company, Incorporated, Research Division

Okay, great. And also just to -- the ThunderX and LiquidIO devices, understand how that works, is ThunderX, does that replace LiquidIO? Is that cannibalizing? Or do they still play together? And do you ever plan on going into the switch system business?

Syed B. Ali

First of all, the answer about switch system is a big N-O, so let's start off with that. Essentially, the LiquidIO functionality is already built into Thunder. So LiquidIO goes into non-Thunder servers, likely the x86 servers, for example, where pretty much all our deployment is right now, from a data center viewpoint. So essentially, with Thunder, the LiquidIO functionality is built in. Where you don't have some of that functionality, you can use a LiquidIO card to add to a standard x86 system.

Kevin E. Cassidy - Stifel, Nicolaus & Company, Incorporated, Research Division

Okay, so are you expecting LiquidIO revenues to trail off next year then, with ThunderX going up?

Syed B. Ali

No, they're mutually exclusive. If you look at the current x86 market, it's probably 98% market share. So the LiquidIO attaches to that, whatever 97%, 98% market share. And whatever Thunder market share it gets, obviously, LiquidIO will not be there, but a higher ASP Thunder will be there.

Operator

Brian Modoff of Deutsche Bank.

Brian T. Modoff - Deutsche Bank AG, Research Division

Couple of questions for you. First Art, can you talk about kind of the visibility you have for Q3? And maybe some visibility into Q4, what kind of orders do you have backing your guide? And then Art, can you talk a little bit about -- can you give us any feedback? Obviously, as part of your due diligence on Xpliant, you talked to potential customers of the product. Can you give us any feedback you might have had out of Cisco or Arista or any of the other switch vendors, what they think of the architecture or what they think some of the use cases might be, what's their view? Appreciate that.

Arthur D. Chadwick

Sure. So talking about our confidence for Q3, I will say this. We are entering Q3 with record backlog and extraordinarily high coverage for the quarter. So we are very confident in our top line guidance for Q3.

Syed B. Ali

And regarding, Brian, the kind of use cases, obviously, Xpliant has talked to a lot of customers and the use cases range from, kind of like I talked about, answering a previous question, standalone pizza box or chassis-based switches to service provider type boxes to enterprise boxes. So there's a very, very wide range of applications. And since this product is programmable, we are able to attack multiple markets with one common architecture, rather than have an architecture per segment.

Brian T. Modoff - Deutsche Bank AG, Research Division

Okay. And then what was some of the feedback customers might have given you on the product?

Syed B. Ali

Well, I think it's evident from the fact that we decided to acquire them...

Brian T. Modoff - Deutsche Bank AG, Research Division

I get that, but I just...yes. I get it.

Syed B. Ali

No, the feedback is very, very -- been very, very consistent. First, your speeds and feeds are world-class. So definitely, it's a world-class switching product. But second, what intrigues them is obviously, the level of customization and the level of special features that they can add on. So it's becoming a very attractive value proposition for customers.

Operator

[Operator Instructions] We will now hear from Matt Ramsay of Canaccord Genuity.

Matthew D. Ramsay - Canaccord Genuity, Research Division

Syed, maybe you could add a little bit, sort of the future of how you're thinking about the roadmap of Xpliant as you bring it in. To me, it seems like there's -- you spoke a lot about the software configurability and different features of their product roadmap for SDN and other applications. I mean, is that something that you feel you can then take and wrap around some of your own custom core work, either MIPS or ARM, going forward? Or is it going to remain 2 separate product lines in your view?

Syed B. Ali

At the beginning, obviously, there are separate product lines, but there's a lot of cross-pollination that can occur, because as you see what is happening in the cloud, everything is really networked out, right? So having processor and switching solutions, whether locally within the system or from system to system, this is going to be a very, very attractive piece of technology. So I think if you -- on both sides of the equation, the cross-pollination between kind of the CPU side and the switch side will really lead or will really drive our solutions to be very, very world-class.

Matthew D. Ramsay - Canaccord Genuity, Research Division

And then one follow-up on Thunder. Obviously, you guys announced, when you did the announcement in Taipei, the multiple different SKUs of ThunderX part. Maybe you could give us a little bit of color into different customer feedbacks on the different SKUs and which, I guess, workload tiers you're more or less excited about for that roadmap as it rolls out?

Syed B. Ali

The really interesting thing that has happened here, Matt, is we have customer interest in all 4 of the SKUs. And I'm talking about, from kind of name brand customers. So there are certain class of applications where they just want very high-density compute, and that goes with our compute SKU. A lot of web front-ends and others that need security that goes for the secure compute SKU. Things like root clusters, storage, and things like media streaming, for example, go with our storage SKU. And finally, things like NSV, cloud RAM, they are very attracted towards the ThunderX NT SKU. So it's been very encouraging for us that we are seeing interest across the board on these products.

Operator

Sanjay Chaurasia, Nomura.

Sanjay Chaurasia - Nomura Securities Co. Ltd., Research Division

Syed, one question. Could you give us some background how the deal evolved for you guys? And given the way you have described these products, it seems like a very attractive acquisition price. And I was just wondering if there were any other bidders for this asset.

Syed B. Ali

Yes, essentially, the founders of Xpliant, when they were doing their rounds of funding in early 2012, they also came to us. We looked at the technology, at that specific point, felt it had a lot of potential in next generation type networks. So that's why we were a lead investor. There were obviously other investors in them also. And as the company went and started talking to everybody from kind of enterprise to service provider to hyper-scale type guys, the interest level and the fascination with this product was very, very strong. And that's what kind of led us to go ahead and do it a little bit sooner than later on this. And overall, from our viewpoint, because we were a early investor in Xpliant, this definitely helps. And there was a fixed price, if you will, a call option to buy the company over a certain timeframe for a fixed price. And that obviously helps keep it in a more reasonable viewpoint. And the other aspect of this equation was -- yes, there was some other interest also from other strategics for this technology.

Sanjay Chaurasia - Nomura Securities Co. Ltd., Research Division

And as a follow-up to Art. Art, could you remind us that what is your planned increase in the OpEx for next year, which you indicated that will be sufficient to absorb the cost of this acquisition?

Arthur D. Chadwick

Sure. So what we've said many times is our longer-term goal is to increase OpEx at about half the rate of our top line growth. So for example, we're halfway through this calendar year, our top line has grown about 12% since the December quarter, and our OpEx has grown about 6%, that's over the last 6 months. So we've been maintaining kind of that 1/2 of top line growth. Not all quarters are created equal, but going into next year, our goal is to increase OpEx somewhere around half of our top line growth. That has not changed.

Sanjay Chaurasia - Nomura Securities Co. Ltd., Research Division

And any kind of guidance? Because that's -- given you have so many drivers next year, Project Thunder, back half, some other products ramping, including the switch -- Xpliant switch, how do we model this? Like we -- it could be...

Syed B. Ali

Yes, I think, Sanjay, as Art said in his prepared comments, we expect this to be a few cents negative in the first half of the year, and positive to neutral after that. So you can figure out that the incremental OpEx above and beyond what is already there in our plan is very small.

Operator

Next, we'll hear from Ruben Roy of Piper Jaffray.

Ruben Roy - Piper Jaffray Companies, Research Division

Syed, on the LiquidIO design wins you discussed, 2 design wins, are they both, or each, I should say, unique customers? And it sounds like they're shipping meaningfully in 2015. For 2014, is LiquidIO still -- the revenues mostly related to one customer?

Syed B. Ali

The last portion of 2014 revenues is related to our lead customer, but we have shipping to other customers also. But obviously, the per customer volume shipments are much smaller than the -- our biggest customer. In 2015, we expect meaningful revenue from these new design wins, these are new customers, new applications. So they are slightly different in the deployment model compared to what our current large customer is doing.

Ruben Roy - Piper Jaffray Companies, Research Division

Okay. Thanks, Syed, for that. And then quickly for Art, nice job bringing the gross margins up as promised in terms of the outlook for Q3. As broadband, obviously, is still very small, but I'm thinking if you get some traction with the lower-end OCTEON III, are there any moving parts to gross margins as we look forward? Or is the level that you're forecasting for Q3 a good level to think about over the next several quarters?

Arthur D. Chadwick

Yes, as Syed mentioned, we do expect to ramp revenue with the lower-end OCTEON product, but we also expect to ramp revenue with the higher-end OCTEON. And we get great gross margins there. So to answer your question, we expect some gross margin expansion in Q3, as I mentioned, and we expect to continue to expand gross margins to a modest amount, in the following couple of quarters.

Operator

Next, we'll hear from Joe Moore, Morgan Stanley.

Joseph Moore - Morgan Stanley, Research Division

Customer concentration, I guess, you said 51% of revenues from your top 3 customers. That was 42% last quarter, if I remember right. Can you talk about where you think that number trends over time? I would think some of these newer businesses would tend to diversify that a little bit. We know -- do you see those 3 customers growing faster than the rest of your business?

Syed B. Ali

I think your assumption is correct, Joe. We expect this percentage to come down, because like I said, we are starting to fire on all cylinders across multiple markets. So the percentage of that, we expect to be trickling down.

Arthur D. Chadwick

And we expect it to go down in Q3 specifically.

Joseph Moore - Morgan Stanley, Research Division

Okay, great. And are the 3 customers approaching kind of similar size? Or is still Cisco kind of the biggest by [indiscernible] margin?

Syed B. Ali

Joe, it's a range. And as you know, in kind of the service provider type markets, they -- it tends to be lumpy. So there'll be a specific quarter, where one customer kind of shoots up, another one comes down. The next quarter, it could just kind of switch sides. So overall, there's a range, obviously, of customers, but it's not a huge wide range now.

Operator

And then we'll take our next question from Betsy Van Hees of Wedbush Securities.

Betsy Van Hees - Wedbush Securities Inc., Research Division

Art, you mentioned in your prepared remarks that the inventory that grew -- it's been a long time since you guys have had a kind of a jump in inventory greater than 15%, I had to go back to Q2 of 2011 -- was to prepare yourself for the second half. And your -- midpoint of your revenue guidance is 7%. I was wondering if you could help us understand the -- what type of products that are making up that increased revenue and the end markets that you are looking at that are going to be so robust in Q4.

Arthur D. Chadwick

Great question, Betsy. Yes, inventory went up this quarter. We've got a lot of inventory that we're building, staged for sales growth -- expected sales growth in Q3 and Q4. We've got a lot of products these days and we continue to add to those products. OCTEON III is ramping, as Syed mentioned. We've got Thunder coming out, that's not included in this particular inventory, but we've got a number of new products that have been layering on, including LiquidIO and others. So it really comes down to the fact that we've got more products than ever, which is a good thing. That sales are going up nicely, which is a good thing. And we have to order in advance of our customer orders. It takes that long to get through the factories, so we need to make sure that we've got a full array of what our customers are projecting to buy.

Syed B. Ali

One more kind of add-on to that, Betsy, is, for example, as Cisco grows, there is -- to Cisco, we ship in a hub and we own that inventory. So as Cisco, for example, grows, that hub inventory grows, but it adds on to our inventory line, number one. Number two, also, for some of the more larger customers, we work with them to keep a certain amount of buffer stock, so we can actually go ahead and support some quick upsides that they may have with short lead times. So kind of a combination of these 2 lead to slightly higher inventory rates.

Betsy Van Hees - Wedbush Securities Inc., Research Division

That was very helpful. And it sounds like that -- you said you guys had record backlog again. And you said that in the prior quarter, so congratulations. But it sounds like you actually have greater visibility into the Q4 quarter, given your confidence level to bring in that much inventory. Is that a fair statement?

Syed B. Ali

I think, Betsy, here, when you take a look at it, visibility, obviously, into Q3, is very good. And so far, only a month has gone into Q3, our visibility also into Q4 is kind of approximately what we hope for. So we -- so far, we have not seen any surprises.

Betsy Van Hees - Wedbush Securities Inc., Research Division

That's great. And then I just had one last question on the acquisition. Congratulations. You said it was going to close in Q1 and that seems, given that it's a privately held company, kind of a long time. Did I miss something in terms of some regulatory hurdles that you have to get over as to why the length is so long in terms of closing the acquisition?

Syed B. Ali

I think what Art has said is that, by Q1, so he's kind of giving a worst case type scenario. So Art, if you want to add on to that?

Arthur D. Chadwick

No, I mean, that's simply the timeframe, Betsy. As Syed mentioned, there was a lot of interest in this company, so I think we ended up signing this acquisition deal maybe a little earlier than we would have had, had that other potential competition not been out there. We would've given them just a few more months to get the silicon out, but we decided to sign the agreement now and we expect it to close by Q1. So that's the timing.

Operator

And next, we'll hear from Suji De Silva of Topeka Financial.

Sujeeva De Silva - Topeka Capital Markets Inc., Research Division

I don't know if you provided this before, but did you provide the enterprise versus service provider mix in the quarter? And more importantly, what do you think that mix -- what happens to that mix in the next 1 year or 2 -- 1 to 2 years? Does enterprise start to grow because of the products you're bringing out?

Syed B. Ali

I think we've talked about this before Suji. A couple of years ago, we used to be like 2 to 1 enterprise data center versus service provider. That number is coming down, and over time, we expect it to become -- for our existing products. The new products, like Thunder or others, will have a completely different dimension, but for our existing products, they will more come closer to parity.

Sujeeva De Silva - Topeka Capital Markets Inc., Research Division

Okay. And then Art, I think I heard you mention that acquiring Xpliant gives you the opportunity to forego some extensions you had in your internal plans. Was that correct? And did you have an internal effort similar to theirs underway that this obviates the need for?

Arthur D. Chadwick

No, I think my comments are the reverse of that. We have got a certain infrastructure within Cavium and that Xpliant has a current run rate and that expense run rate, we expect the incremental expenses that we will add on will be substantially less than their current run rate. In other words, we can leverage our infrastructure and not take on 100% of the expenses they are currently spending, because we already have an infrastructure within the company.

Sujeeva De Silva - Topeka Capital Markets Inc., Research Division

Okay, so you're describing the synergies. That's clear. And then lastly, Syed, what's the pain point for the cloud and hyper-scale guys that the acquired technologies solution addresses differently from the existing market solutions?

Syed B. Ali

So when you take a look at many of the -- this is specific to hyper scale. So first -- then I take a look at kind of standard enterprise data center type switches. This has world-class speeds and feeds. It is a best-in-class product and we are pretty confident that this product will gain market share in that segment. But the second more -- very interesting piece is that when you take a look at hyper-scale data centers, a lot of it is a closed system, right? Where the hyper-scale vendor pretty much owns the entire thing. So essentially, what he wants to be able to do is do things sometimes in a nonstandard way to fit his architecture better and to improve performance of the way he has deployed it. So that is what is driving a lot of interest in that -- in this product from the hyper-scale guys.

Operator

And there are no further questions at this time. That does conclude today's conference. Thank you all for your participation.

Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.

THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.

If you have any additional questions about our online transcripts, please contact us at: transcripts@seekingalpha.com. Thank you!

This Transcript
All Transcripts