Attached files
| file | filename |
|---|---|
| 8-K - 8-K EARNINGS RELEASE Q1'16 - APPLIED MICRO CIRCUITS CORP | a8-kearningsreleaseq116.htm |
| EX-99.1 - EXHIBIT 99.1 - APPLIED MICRO CIRCUITS CORP | exhibit991pressreleaseq116.htm |
Exhibit 99.2
Applied Micro Circuits Corporation
Fiscal Q1 2016 Conference Call Script - with Q&A
July 30, 2015
PRESENTATION
Operator
Good day, ladies and gentlemen, and welcome to the Applied Micro Circuits Corporation first-quarter FY16 earnings call.
As a reminder, this conference call is being recorded. I would now like to introduce your host for today's conference, Ms. Suzanne Schmidt with Investor Relations. Ma'am, you may begin.
Suzanne Schmidt - The Blueshirt Group - IR
Thank you operator. Good afternoon, everyone and thank you for joining today's conference call. On the call with me are Dr. Paramesh Gopi, our President and CEO, and Doug Ahrens, our Vice President and CFO.
Before we begin, I would like to remind you that various remarks that we make on this call including those about future financial results including revenues; gross margins; operating expenses; design wins; product plans; our competitive situation; market trends; statements about future development, production and adoption of X-Gene®, X-Weave®, HeliX® and other products; 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 forms 10-K and 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 them.
I would like to point out that a number of the securities analysts that cover our stock use various measures, and this creates a range of variability relative to Street financial models. When we say Street estimates, we mean consensus of the major financial analysts' models, and they are not necessarily the guidance that was given by the Company. With that, I will turn the call over to Paramesh.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Thank you, Suzanne. This afternoon, we reported results for the first quarter of FY16, and we are making excellent progress on all fronts.
First, our base business is solid, as demonstrated by growth in product revenue from the prior quarter.
Second, X-Gene is continuing to gain momentum with more end-customers, and is a clear leader in the ARM server space. At the Apache Spark Summit in San Francisco last month, X-Gene was the only mature production-ready ARM solution to be unveiled and demonstrated. In addition, over one dozen X-Gene-based ODM and OEM products are production ready. X-Gene has completed several extremely important customer trials that are necessary to displace a
long-standing incumbency, and these X-Gene deployments have resulted in numerous success stories and testimonials over the past year and a half.
Third, we have secured a growing number of design wins for our HeliX product family, further demonstrating that we are driving the transformation of the market to 64-bit ARM as the new emerging standard for embedded applications. We have successfully converted and continue to convert several MIPS sockets from the leading incumbent to HeliX, thereby expanding our leadership position in the embedded market for applications such as network function virtualization, or NFV, enterprise class switches and routers, network security and storage.
Fourth, with respect to our technology leadership, we have taped out our next generation of X-Gene test chips on leading 16-nanometer FinFET technology. We have already demonstrated our competitiveness versus products that are on smaller geometries and our investments in leading edge process technologies further enhance our performance relative to the incumbent x86 suppliers and continue to place us ahead of all ARM 64-bit data center silicon providers. Finally, we are extremely excited to announce the tapeout of the industry's first 16-nanometer ultra low-power high-density PAM-4 converged copper optical DSP device, that extends our proven leadership in the service provider and data center markets and opens up a large new growth opportunity for our Connectivity business.
With regards to X-Gene end-customer testimonials, I would like to highlight our recent announcement that TSMC, one of the world's leading semiconductor foundries, will utilize our X-Gene-based processor integrated within HP's Moonshot platform as a cost- and power-saving option for implementing an open source memcached solution for web-tier front-end optimization. We are excited to see customers like TSMC recognize the performance capabilities and total cost of ownership advantages that X-Gene delivers when compared to the latest generation of x86/Intel products. In fact, X-Gene technology is in production today in a wide range of leading OEM and ODM storage and compute platforms, enabling IT organizations to implement power-efficient, high-performance and significantly lower TCO solutions for their key workloads driving enterprise and cloud data centers.
Having highlighted our X-Gene progress, I will now update you on the growing success of our X-Weave and HeliX product lines within our base Connectivity and Computing businesses, after which I will return to X-Gene. Turning first to our base business and our Connectivity product line: Sales of our connectivity products grew again in the June quarter and are once more driven by robust demand for 10 Gigabit per second OTN ports and DWDM equipment as well as 100 Gigabit per second long haul service provider and intra-data center links.
Our X-Weave products have opened up new architectural opportunities for us in data centers. As a key proof point, a major Tier 1 networking customer is developing new enterprise switches for mega data center applications, which was made possible by the leading-edge technology provided by our X-Weave product family. Our latest X-Weave product is in the unique position of enabling major networking OEMs and data center operators to meet the new security specifications required by internet content providers and government agencies.
As you may recall, we are seeing wide Tier 1 data center and service provider platform adoption of our X-Weave 240 Gigabit per second products, which combine three important pieces of our leading-edge technology portfolio. One, the industry's leading forward data collection algorithms for long haul and metro networks. Two, our high port density, lowest-power analog mixed signal transceivers. And three, 100 Gigabit per second digital signal processing, or DSP. These technologies power the industry's fastest-growing packet optical service provider and router platforms.
Today we are very excited to announce that we have taken the next breakthrough step in extending our leadership in data center interconnect technologies by taping out the industry's first 16-nanometer FinFET ultra low-power and high-density 100/400 Gigabit per second PAM-4 optical and copper physical layer device. To give you some context on why this is important, I would like to elaborate on the technology, its implications for the data center and also share with you its growth prospects within our connectivity business.
PAM-4, which is four-level pulse amplitude modulation technology, converges the inter-data center and telecom service provider metro technology interconnect requirements into one single standard, thereby enabling 100/400 Gigabit per second, small form factor, low-TCO modules. In addition, PAM-4, which is in the process of being standardized by the IEEE, is fast emerging as the interconnect within and between server racks for custom, high-density, hyperscale
data centers and is predicted to be the de facto physical layer for 100 Gigabit per second and 400 Gigabit per second switch port connections.
The market for PAM-4 solutions is estimated to grow to north of $750 million in 2020. We estimate first deployments of this technology to begin in late 2016. With the tapeout of our newest multi-laser and copper capable ultra low-power, high-density 16-nanometer FinFET platform built on our proven technology leadership, we are uniquely differentiated as we can enable a true multi-terabit per second solution within current rack and chassis power envelopes.
Additionally, our proven PAM DSP technology provides an unprecedented level of programmability to accommodate standards evolution and the various interconnectivity options desired by service providers and data centers. We look forward to providing you further updates on our progress as we lead this exciting new market with our breakthrough technology investments.
Moving now onto our computing business, HeliX remains the leading choice for high-volume enterprise networking equipment vendors as they move from MIPS and PowerPC to 64-bit ARM. We remain very enthusiastic about our opportunities in the embedded market as we leverage our existing IP and investments in 64-bit ARM to serve embedded applications.
Last quarter, we indicated that we expected to expand our HeliX design wins into new markets. We are pleased to now report new innovative applications using HeliX 2 in the data center outside of general-purpose compute platforms and storage platforms. In this context, we are very excited to highlight two important new design wins, both of which replace existing Intel processors.
The first design win brings 64-bit ARM to a high-performance Tier 1 data center video platform where HeliX enabled a unique set of features including enhanced surveillance and real-time analytics without compromising capacity at scale. This is the start of an important trend where data center operators pair function-specific, programmable accelerators with existing infrastructure to enable real-time services at the lowest incremental cost at massive scale. This data center video platform is targeted for production release in the first calendar quarter of 2016.
The second HeliX 2 design win brings 64-bit ARM, large memory and 10 Gigabit per second packet processing to enable unique data center network management and software acceleration critical for software defined networks. HeliX 2 enables one of our top leader customers to drive wireless network management at scale for large data center operators and hyperscale enterprises.
Lastly, we are heavily engaged with top-tier networking and storage OEMs who have chosen HeliX over other competing ARM-based products to facilitate their move from MIPS-based architectures to 64-bit ARM.
Now turning back to X-Gene, we made significant progress in the June quarter with market penetration increasing and ecosystem partners advancing. In the June quarter alone, we secured more than 20 new end-customer engagements in targeted vertical markets, and we expect a good number of these to convert to design wins and deployments going forward.
To provide further context on the new and growing enterprise ecosystem for X-Gene, I would like to elaborate on the showcasing of X-Gene platforms at the Apache Spark Summit held last month in San Francisco. Spark is the number one choice of large enterprises for predictive analytics and deep learning, which are the highest growth areas for scaleout computing and enterprises.
We were able to demonstrate two unique configurations running Spark. The first is a cluster of X-Gene servers for predictive analytics featuring applications targeting airlines and metro transport industries.
The second is a high-performance compute cluster consisting of X-Gene server nodes and NVIDIA's GPU targeted at deep learning applications. This comes on the heels of over one dozen production-ready ODM and OEM products based on X-Gene, real end-customer trials and deployments that have resulted in success stories and testimonials.
I would like to emphasize that all ARM service silicon vendors will have to go through a similar cycle to deployment and to revenue. In the scientific and high-performance computing vertical, Cirrascale and E4 both demonstrated and are shipping second-generation production platforms based on X-Gene that interoperate with Mellanox's 56 Gigabit per second and InfiniBand technology and the newest NVIDIA Tesla GPU compute engine.
In the enterprise vertical, we have several major financial institutions beginning to deploy production X-Gene servers for a variety of different applications. This is our first announcement in the financial segment, and this enterprise vertical represents a significant SAM expansion beyond that represented by hyperscale data centers.
Finally, in the hyperscale data center vertical, one of the largest Asia-based mega data center operators has made a commitment to port native applications to X-Gene, and we expect the commitment to result in hardware deployment over the course of the year. As I have mentioned in the past, there is a strong imperative in Asia to move toward alternate open CPU architectures such as ARM, and we are in an excellent position to benefit from this strategic priority.
In terms of our ecosystem, there was excellent progress in the June quarter. We now have the first SDN/NFV port on X-Gene which illustrates the ecosystem momentum around X-Gene in this space.
Additionally, Red Hat, SuSE, CentOS and Ubuntu are all in distribution. Red Hat is in advanced beta, and SuSE and CentOS and Ubuntu are now fully in production. With the addition of SuSE as another major Linux OS provider, more than 80% of the ecosystem is X-Gene ready, and, by definition, we have become the de facto ARM solution provider for a wide range of cloud and enterprise applications.
I am happy to note that our second-generation product, X-Gene 2, is in active evaluations with a number of key end customers. In addition, we have taped out our 16-nanometer FinFET test chip in preparation for our next generation X-Gene products. It is important to note that the X-Gene family is the de facto industry standard for scale-out ARM 64-bit server solutions, and we continue to extend our leadership position by investing in advanced process technologies.
As you know, X-Gene 1 competes and wins against Intel's 22-nanometer and 14-nanometer products that utilize a FinFET process. The tape out of our own FinFET test chips uniquely positions us to now compete against Intel's future roadmap including their premium class E7 platform.
With that, I would like to turn the call over to Doug for a review of the June quarter financials and our September quarter guidance.
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
Thanks, Paramesh. June quarter consolidated net revenues were $37.8 million, slightly above the midpoint of our guidance range. The book-to-bill ratio in the June quarter was above 1.
Strength in our connectivity product revenues, particularly our X-Weave 240 Gigabit per second and OTN products, served to more than offset lower PowerPC business. As we have noted previously, we expect PowerPC to oscillate around current levels in the near term.
Importantly, customer activity around HeliX remains strong as we continue the transition of our embedded product lines from PowerPC to 64-bit ARM which puts us in the position to gain embedded market share. Connectivity product revenues were $23.3 million, 14% higher than the prior quarter. And computing revenues were $14.5 million, 13% lower than last quarter.
In terms of geographical split, sales to North America accounted for approximately 47% of total revenues. Sales to Europe contributed 20%, and sales to Asia contributed 33%.
In the June quarter, four customers accounted for more than 10% of our business: global logistics support vendor Wintec, distributors Avnet and Arrow, and contract manufacturer Flextronics. Distributor revenues for the June quarter were $30 million.
Non-GAAP gross margin in the June quarter was 55.8%, above the midpoint of our guidance range primarily due to mix. Non-GAAP operating expenses were $24.4 million, down from $26.5 million in the prior quarter.
Non-GAAP R&D expense totaled $17.6 million or 12% lower than the previous quarter due to the full realization of savings from operating efficiencies that we initiated earlier this year. Non-GAAP SG&A expense totaled $6.8 million. Our income tax was approximately $200,000.
Our first fiscal quarter non-GAAP net loss improved to $3.3 million or $0.04 per share. Our net loss on a GAAP basis was $7.4 million or $0.09 per share versus a net loss of $15.1 million or $0.19 per share in the prior quarter.
Our non-GAAP financials exclude certain items required by GAAP. A complete reconciliation between GAAP and non-GAAP financials is available on our June quarter earnings release which can be found in the Investor Relations section of our website. Please note there is no reconciliation relating to forward-looking statements.
Channel inventory was 34 days, down from the prior quarter's 41 days and the lowest it has been since 2009. We are continuing our effort to maintain a lean supply chain and enhance our visibility to end customer demand.
Turning to the balance sheet, our cash and short-term investments increased to $76.6 million compared to $75.4 million in the prior quarter. Internal inventory levels were $19.8 million, down from $23.5 million in the prior quarter. Inventory was higher in the prior quarter due to timing of specific purchases.
Now turning to our payments related to the Veloce acquisition, during the June quarter, Veloce payments were comprised of $100,000 in cash and $900,000 in stock. To date, we have paid a total of $169.9 million or about 95% of the total Veloce merger consideration of $178.5 million.
With regard to guidance for the September quarter, revenue is forecasted to be approximately $39 million plus or minus $2 million. We have good visibility it into our September quarter revenue. Our backlog coverage as of the date of this call is greater than 80%. The combination of a lean distribution channel, solid backlog coverage and a strong book-to-bill ratio gives us confidence in our September quarter revenue guidance.
We expect non-GAAP gross margin in the range of 55% plus or minus 2 percentage points. Non-GAAP operating expenses are expected to be in the range of $24.5 million plus or minus $1 million as we continue to remain focused on cost control.
This is our new baseline level of non-GAAP OpEx, but we could periodically experience a bump in OpEx associated with spending for masks. In the September quarter expect a net tax credit of approximately $500,000, and going forward, we expect our tax expense will return to approximately $200,000 per quarter. We expect earnings per share to be a loss of $0.03 plus or minus $0.02 on a non-GAAP basis.
To summarize, our base business has clearly stabilized and we are seeing excellent customer activity around our HeliX product family which should build our embedded computing business over the coming years. With regard to connectivity, the fact that we have taped out the industry's first PAM-4 transceiver platform on a 16-nanometer FinFET process is evidence of our ongoing focus and successful development of leading-edge optical networking products.
Lastly, following numerous prior announcements about product and ecosystem readiness, we are very pleased with the level of end customer adoption of X-Gene such as our recent announcement about the engagement with TSMC to leverage X-Gene technology for web caching applications. Now I will turn it back over to Paramesh for some closing remarks.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Thanks, Doug. To summarize: First, as demonstrated by the TSMC announcement and the PayPal disclosure from the last quarter, we are seeing growing evidence that X-Gene is doing very well in early rack-scale deployments, a precursor to broad market adoption. In addition to these important new engagements, our pipeline is expanding and we have production platforms in use by an increasing number of key customers in every vertical market we address: scientific and HPC, hyperscale, networking and now the enterprise.
With X-Gene 1 and 2, we have a clear time-to-market lead that's enabling us to capture design wins and establish barriers to entry for any ARM competitor. In addition, X-Gene puts us in a strong position to capitalize on China's imperative to break free from the monopoly of x86 server solutions. We continue to extend our technology leadership by investing in advanced process technologies such as a 16-nanometer FinFET tape out, thereby enabling us to compete successfully with Intel's E7 class server products.
Second, we announced the tapeout of the industry's first 16-nanometer FinFET PAM-4 solution capable of addressing both the high-growth data center interconnect market as well as the server rack interconnect market. This market is estimated to be $750 million in 2020 with products from OEMs slated to ship in late 2016.
Our unique solution represents a multiyear investment and is slated to have the lowest TCO per port, thereby putting us in a uniquely differentiated position to enable terabit connectivity within current power system budgets. I would also like to point out that this new generation of PAM-4 capable X-Weave technology will be a very important complement to our industry-leading X-Gene product portfolio as servers move to 100 and 400 Gigabit per second rack and cluster connectivity.
Third, as for HeliX, our ARM 64-bit embedded product family, customer activity continues to be very strong, and we have multiple design wins that should allow us to grow embedded processing share over time taking share from MIPS and x86 architectures. Finally, our base business is being revitalized by strong customer traction with X-Weave in both the service provider and data center markets particularly with the new security specs required by OEMs that our latest generation of products deliver.
With that, we would like to open up the call for questions. Operator?
Operator
Cody Acree, Ascendiant Capital.
Cody Acree - Ascendiant Capital Markets LLC - Analyst
Thanks, guys, for taking my questions and congrats on the progress. Paramesh, can you maybe go back to the TSMC announcements and give us maybe some idea of sizing of that opportunity and maybe timing as well?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Just to give you some context, TSMC is obviously a captive enterprise data center which includes a number of different types of applications. One of the key application sectors for them is essentially the web tier because a lot of their customers interact with them through the web tier. And part of that is how do you have a very large set of data that is cache-based and that you can essentially serve up for very low latency cache-based customer access.
So just to scope it, today that type of tier is enabled by Xeon E5s. And if you look at the papers that have been published, I will take you back almost 2 1/2 to three quarters ago to when HP published the first white papers on what you would save if you were to use a high-density server like an HP Moonshot. And it is almost like a case study where you run
what they need on the platform and you get the type of savings that they talked about in that white paper, which was between 25% and 50% of savings versus an existing incumbent E5 type architecture. So that puts color on the actual application set.
When this will deploy, our expectation is that most of these -- most of the other pieces that we talked about over the last six to nine months, this is the year of adoption. So they will continue to basically drive small deployments this year, and we expect that this will all turn into, what I will call, a change in their deployment to a much higher level next year.
Cody Acree - Ascendiant Capital Markets LLC - Analyst
Paramesh, as we look at X-Gene now that you have met your 10,000 units target in the last couple quarters, are there some new metrics that you might be able to give us a way to track this? I know you are not ready to start breaking out revenue in that segment of computing, but is there something tangible that you can give us as a way to track the progress of the revenue ramp?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
I think you said two things. One is we were going to give you end -- in the end, all of this is going to manifest into end customers buying server products. I just want to make it super clear.
So two years ago we talked about probably almost 200 people signing license use and evaluation agreements with us. And it has taken that long for us to get to end customer testimonials like what we are experiencing today.
I want to say that this is the necessary and sufficient condition for us to generate revenue in the future, number one. Number two, as far as the timing of the revenue, I can say that I've never felt better than I do today about the line of sight to revenue being generated in calendar 2016.
Cody Acree - Ascendiant Capital Markets LLC - Analyst
Doug, the cash that stabilized here, growing a little bit quarter to quarter, inventory dropped a bit. Are we reaching now a level that you are starting to feel comfortable with cash stabilization? What's your trajectory look like?
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
This quarter working capital went our way if you look at the change in inventory and payables and receivables. So as a general rule, you can look at our non-GAAP net income as an indicator of the trend on cash.
So right now, given our guidance, we are in that mode of transitioning: We've got the stable base business, we've got the traction with the ARM products and a lot of good upside here that's come from our new products in the Connectivity space. So you see the trends going in the right direction also combined with our good OpEx controls which is now $24.5 million a quarter. You put all that together, yes, our cash position looks healthy in our minds.
Cody Acree - Ascendiant Capital Markets LLC - Analyst
Lastly, Paramesh, maybe on a different topic, Intel and Micron just announced their 3-D crosspoint memory architecture, a little different approach bringing memory closer to the processor. They are touting obviously much better speeds for performance in servers and data centers. I guess what is your reaction to that architecture, and is there any reason that X-Gene could not eventually pair with that memory structure?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
I can tell you that just to give you some context, about a year ago, I think there's actually a public paper written that we did with SanDisk, and one should assume that where Micron and Intel are talking about bringing those two things closer together, it makes the most sense to do it with ARM. Why? Because we have the highest ability to integrate and not deal with the Windows tax overhead that most data centers in the new age don't want to deal with.
We can essentially custom craft memory interfaces like we've done with X-Gene 1 and X-Gene 2. And with things like Rocky that link chips together, what we're doing is effectively changing the memory density of servers going forward.
Just to give you a data point, I would tell you that a Tier 1 mega data center operator has already customized boards with X-Gene 1. This happened a year ago where they used a different set of Micron-type memories running at a different speed to quadruple the densities of memory hooked up to X-Gene. We are already at a point where we can see that trend having a very large effect, but I think it's funny -- I think that trend started in our industry, in the ARM industry, when X-Gene 1 was first introduced to mega data center operators. So very happy to see that Intel finally caught up.
Cody Acree - Ascendiant Capital Markets LLC - Analyst
I expect nothing less, thanks, Paramesh. Appreciate it, and good luck.
Operator
Hans Mosesmann, Raymond James.
Hans Mosesmann - Raymond James & Associates, Inc. - Analyst
Thanks. Paramesh or Doug, can you describe or give us a sense of X-Weave sales in the quarter, what part of the growth in the September quarter is coming from X-Weave?
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
X-Weave in terms of the newest 240 Gig part is still our newest product in the family. So it is relatively small as a portion of the total, but we are very happy with the growth we are seeing. The bookings have been growing nicely, the shipments have been growing.
We just don't break out any individual products, but I can say we're very happy with the newest 240 Gig part. We saw in the Connectivity space a very stable picture. When I look at the data, the sell-through was healthy, the bookings are healthy.
So we are just in a pretty stable place right now in Connectivity, and I think a lot of that has to do with the fact that we have such a lean channel inventory. It's really working in our favor right now.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Hans, to add a little bit more to that, I would say in the history of the Company, since even before I have been here, what we hear in terms of the statistics are that we have never had this level of design win momentum for our -- I will call it 100 and 240 Gigabit per second X-Weave products from the time they were announced. I think we mentioned that on the last call as well.
So, unprecedented design win momentum. And most importantly, I don't know whether I can put a fine point on this, we are now going to lead the entire transition into the PAM-4 space with X-Weave.
In fact, I think versus any other competitors out there, we put a huge foot forward by taking the technology and moving it directly to 16-nanometer. And we will probably have the lowest power, highest density FinFET and the only FinFET offering within the next two quarters.
Hans Mosesmann - Raymond James & Associates, Inc. - Analyst
That's helpful. Paramesh, on the 16-nanometer front, on the X-Gene tapeouts, when you say X-Gene test tapeouts, does that mean this is not -- the design has not been complete?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
No, I want to remind you if you look at every -- if you go back the last 2 1/2 years and if you examine every one of our test chips, we've always been public about when we tape out a test chip and when we ship a product. If you think about it, it is pretty much signifying that we are now ready to really drive 16-nanometer product through the end phase of development.
Because it is such a huge investment, one of the things we always do is to make sure that we tape out test chips to verify speed, to verify all of the different characteristics that we would require in a process like this, and, most importantly, to make sure that performance is guaranteed. And that's why we feel really comfortable today about the fact that so far we have competed between Atom and E5. I'm happy to say that with the 16-nanometer test chip and what we're trying to do here, we have a great deal of confidence in making sure that we can extend X-Gene up to the E7 range here in short order.
Hans Mosesmann - Raymond James & Associates, Inc. - Analyst
Okay, that's helpful. One last one on the HeliX ARM embedded front, who are the competitors that you are seeing there?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Just to be clear, if you look at who is driving, who has been in these sockets, there are two classes of sockets. One has been with leading MIPS vendors, basically people who've been security processors and those type of functions, specific accelerators. And we essentially completely are on the way to replace them in very, very high volume, large footprint networking applications and security applications.
And the other piece of it is obviously Intel and x86. As you probably know and I talk about, this particular design win that we did is a very high-profile design win where this is essentially a large video analytics and surveillance box that was all based on Intel on 64-bit but the embedded pieces of the Intel roadmap.
We pretty much are able to give them massive difference in terms of density, scale and cost for way better functionality. So what that does is it accelerates the adoption of this technology, bringing service class technology into an embedded space which has suffered for lack of innovation. And you are really changing performance point, price point and security point.
Hans Mosesmann - Raymond James & Associates, Inc. - Analyst
Thank you.
Operator
Matt Ramsay, Canaccord Genuity.
Matt Ramsay - Canaccord Genuity - Analyst
Thank you very much. Good afternoon guys. Paramesh, I wanted to hone in on a couple of things that you pointed out on the X-Gene front. One was a first either design win or engagement with a major financial institution and I wanted to understand a little bit about that opportunity.
And also you mentioned a large web-tier supplier in Asia that you made forays into. Maybe you could give us a little bit more on both of those opportunities and the magnitude and timing, etc? Thank you.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Absolutely. Let me start by the first one. I think this is the first time in the script we’ve talked about the financial sector. I think we are very, very pleased with the type of workload that we are able to show - a 3 to 8 times type performance improvement for the same power with various types of database applications, whether they're things like Mongo DB, whether they're things like DataStax, whether they are things like large in-memory database type applications.
So in the banking and financial sector, the notion of scaleout with deconstrained memory, deconstrained networking and much more scale has caught on very, very well. Very, very pleased with results there.
We have actually not just one but multiple customers in beta right now looking at their specific implementation on these different types of production platforms as they move into 2016. So timing for that on all the financial sector is -- our estimation is in 2016, number one.
Number two, on the mega data center operator when you talked about it, the mega data center operator is one of the largest purveyors of Internet services in the pan-Asia region. And part of their deployment strategy is to make sure that they benchmark against current incumbent Intel E3 and E5.
And essentially they have made a corporate commitment now to ARM on X-Gene because they have now decided to move their entire application base. It is beyond, we are beyond Linux and beyond all of the ecosystem-based software components but now directly into end-customer applications.
And the timeline for that, as I mentioned, that exercise will be complete this calendar year. That will have revenue implications for next year.
Matt Ramsay - Canaccord Genuity - Analyst
Great, thank you for the color there. In a prior question, I think Cody brought up some announcements made by Intel in the memory front.
I would like to get your impression of a couple other announcements that they made as the chief incumbent in some big markets for you guys. One is the push out of their 10-nanometer products a year, and the second one is the acquisition of Altera.
I get a lot of questions from investors these days on customized or application-specific server type products, the ARM ecosystem versus what Intel might be doing longer-term with an FPGA. So any color on any of those two would be appreciated.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Okay, let me answer the second part of your question first, and I will refer you back to some prepared comments that we made today. I think there is going to be a very, very interesting opportunity created for what I call function specific combined software plus hardware single-chip SOCs -- not necessarily FPGAs for a large set of data center applications.
So the premise of basically accelerating certain data center functions, the premise of accelerating certain code fragments and the premise of making sure that you can incrementally add performance to an existing software stack I think is value. It has value.
What I do however think is the price point associated with FPGAs and the volumes that they can attach to have challenges. For instance, we talked about this entire new data center network management offload.
If you are a hyperscale data center provider and you have 100,000 servers and you want to find a simpler way of managing them and performing basic analytics on their usage, you need to have -- touch into every single server. You also need to have requisite compute that is off band with sophisticated software that can be shared between the server and the off band processor. In that case, just the volume that you need and the power efficiency that you need is more aptly suited to a SOC.
So I believe that what has been talked about as a value proposition for certain fixed-function elements and/or programmable high-speed elements versus a CPU exists. I believe that it represents a unique opportunity for people like us and we already have one proof point with HeliX where, believe it or not, that is an exact application where I could see an Intel coming in and saying -- use an Altera FPGA. But the cost points and the power points do not let it sit as a perfunctory element with high volume on a server platform.
Matt Ramsay - Canaccord Genuity - Analyst
Great. That's really helpful. Thanks for the perspective. Doug, maybe you could talk a little bit about gross margin. We've seen a little bit of a recovery in the Connectivity business and there's stabilization there relative to a couple quarters ago, but it seems like the gross margin seems to be still trending a bit downward/flat.
As we get into the back half of this fiscal year and going forward, I don't know how you might project gross margin from here given all the new products that potentially could ramp. Thanks.
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
Yes Matt. So the expectation right now is that this is basically the margin levels we're seeing; 55% represents the mix we have today. And the mix is the main factor here that moves it up or down. The more Connectivity we do, the margins go up, as an example.
So for now, given what we can foresee in the next quarter, we're holding at the 55% level. It's too early to say what the puts and takes will be in the future, so I don't have any further insight to share here today that would change that trajectory.
Matt Ramsay - Canaccord Genuity - Analyst
All right. Thanks very much guys.
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
Thank you.
Operator
Ambrish Srivastava, BMO.
Kunal Patel - BMO Capital Markets - Analyst
Hi, this is Kunal Patel calling in for Ambrish. Thanks for taking my question.
My first question is regarding Cavium, during their call yesterday -if any shipping is done of direct ARM-based products with meaningful volumes towards the end of the year. Can you comment on what you're seeing in terms of competition within the ARM server camp? Thanks.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
So I think that we've said that we've already shipped in meaningful volumes, so I just want to say that. We've also disclosed that we've passed our 10,000 server mark.
We've also disclosed end customers that are using our devices, and I can tell you for a fact that we have full line of sight now into revenue next year from these types of deployments that will convert into production revenue. So I think all that clearly says that we have a clear heads up in this space relative to our competitors.
The other thing that I will also put a fine point on is we have now taped out our first 16-nanometer device for subsequent productization to go out against E7. And the reason we are doing that is primarily to make sure that in the coming years, as we expand ourselves into becoming and maintaining our leadership, we want to also cover the E7 class of servers in the market. So, I think, I am happy to see that other ARM players are here, but wanted to put a fine point on the fact that we were at this for the last four and a half years, we shipped our first production parts in October of last year and we are fully online to making sure that revenue and line of sight is improving almost on a quarterly basis for us.
Kunal Patel - BMO Capital Markets - Analyst
Thanks, that's helpful. And last call you mentioned X-Weave will benefit from Verizon's US Metro network upgrade towards the end of the year. Can we comment on whether that is still on track and when we should see meaningful revenues?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Say that again? I didn't quite understand.
Kunal Patel - BMO Capital Markets - Analyst
The last call you mentioned X-Weave will benefit from Verizon's US Metro optical network upgrade. I think you had a win there. Can you comment on -- is that on track, and when should we see meaningful revenues?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
I maintain what we said last quarter. First of all, that deployment for us will occur in the middle of 2016 is what we said. And we're already seeing early signs of all of that happening.
In fact, if you look at the X-Weave ramp that we expect, I would tell you that we're seeing a little bit of acceleration on the development front for our end customers deploying it. So, very happy to note that that ramp is occurring, and it may even occur a little faster than what we thought.
Kunal Patel - BMO Capital Markets - Analyst
Great, thanks. And just one question for Doug. Is it -- can you remind us what the breakeven revenue number is? Is it still in the low $40 million range?
Doug Ahrens - Applied Micro Circuits Corporation - VP & CFO
The basic assumptions behind that are, given our OpEx levels that we achieved now given the savings we put in place this year and a 55% gross margin level, it puts us at a breakeven revenue level in the low $40 millions. So somewhere between $40 million and $45 million is where that math puts you.
Kunal Patel - BMO Capital Markets - Analyst
Great, thanks guys.
Operator
Suji Desilva, Topeka.
Suji Desilva - Topeka Capital Markets - Analyst
Hey guys. First a question on the embedded business in HeliX. What -- when does the revenue from HeliX kick in to offset the extinct businesses and return that segment to a steady growth business.
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
We already said that HeliX is on track to have first revenues come towards the December quarter of this year. We also in today's call announced a design win. The start of that design win will be in the first quarter of 2016 calendar.
So HeliX revenues--in fact, our confidence in HeliX revenues continues to grow, and the timing continues to actually be drawn in a little bit. So from a revenue point of view, very close. We're less than six months away. So feeling very good about it and feeling really bullish about the fact HeliX has a very wide adoption tier.
Suji Desilva - Topeka Capital Markets - Analyst
Thanks Paramesh. And then the other questions on X-Gene. The X-Gene 2 product, is that something you envision your X-Gene 1 customers upgrading to, or does it bring a new set of customers into the realm of your opportunity?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
It does both, what you said. It's not even about an upgrade. Just to give you a notion, a lot of people have just now deployed X-Gene 1 in their enterprise.
So for them, X-Gene 2 is the next logical step to basically scale out their systems. As you know, it has RDMA on it. And more importantly, anybody who is looking at much higher density compute platforms and much higher density database type platforms will always target X-Gene 2, because you can scale it in a more seamless fashion and you get a better TCO.
Suji Desilva - Topeka Capital Markets - Analyst
Okay, perfect, thanks guys.
Operator
Krishna Shankar, ROTH Capital.
Krishna Shankar - ROTH Capital Partners - Analyst
Paramesh, can you give us some feedback on how the existing deployments and the design wins you've talked about, how they're doing in the field? For example PayPal, HP and -- HP Moonshot and any other deployments in the field. Can you give us some feedback on how things are going so far?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
So I think one thing is clear. When you get one customer after another starting to become a testimonial for you, the assumption is that they're building on the previous customer's confidence of a deployment. So the way it works in the enterprise is, before anybody will adopt you, obviously they do work, but I would like to point out a very interesting trend:
If you look at how InfiniBand got adopted many years ago, it was first part of one enterprise, and that enterprise and/or the government essentially said this is great. People made reference checks, so they actually did less work to get to further deployment. So the fact that we are consecutively talking about people getting small but meaningful rack-level entrees into their data center is very good.
So to answer your question directly, both deployments -- the PayPal disclosure. We also said last quarter that we would expect things like that to convert to production deployments through the course of this year and next year.
No change to that, feeling really good about what we see on that front. I will also say that we're feeling really good about other enterprise customers that we have not disclosed who are using that data and those references to give us very high confidence in our revenue expectations for next year coming true.
Krishna Shankar - ROTH Capital Partners - Analyst
Okay, and then for the HPC market, at the ISC conference in Europe last month you announced some upgraded design wins and referenced designs of people like E4. Can you talk about the HPC market opportunity for you and how that could ramp in 2016 with partners like NVIDIA?
Paramesh Gopi - Applied Micro Circuits Corporation - President & CEO
Absolutely. So essentially we had at ISC announced we had actual testimonials from CERN from the labs, and from other major supercomputing vendors that talked about the power/performance proof points that they were going to use. This was all based on X-Gene 1. The last set of platforms that we talked about at this ISC was all based on X-Gene 2.
X-Gene 2 is uniquely capable because of the scaleout interfaces that have been added to it. Therefore, if you were to look at it, we have E4 and Cirrascale - they did one platform last year with one generation of NVIDIA, they did another platform this year with the new generation of NVIDIA. And we see a lot of end customers being lit up in both platforms. The interesting offshoot of this which is very neat for us is to connect enterprises.
I don't know whether you heard the first part of the call, but we talked about us being the only solution at the leading Spark Conference in San Francisco. In fact, NVIDIA GPUs and X-Gene make for the ideal combination for the future of data analytics as well as machine learning. So a lot of the HPC work, Krishna, is going to get very leveraged over the coming quarters into enterprise deployment.
Krishna Shankar - ROTH Capital Partners - Analyst
Great, thank you.
Operator
I'm showing no further questions at this time. Ladies and gentlemen, thank you for participating in today's conference. This does conclude the presentation, and you may all disconnect. Everyone have a wonderful day.