Good day, ladies and gentlemen, welcome to this conference call to discuss our agreement to acquire Altera. My name's Dave. I'll be your operator for today. At this time, all participants are in listen-only mode. We will conduct a question and answer session toward the end of this conference. If at any time during the call you require assistance, please press star zero and an operator will be happy to help. As a reminder, the call is being recorded for replay purposes. I'd now like to turn the call over to Mr. Mark Henninger, Head of IR at Intel. Please proceed, sir.
Great. Thank you, Dave, welcome everyone to Intel's conference call to discuss our agreement to acquire Altera. By now you should have received a copy of our press release, but it's also available at our transaction website, intelacquiresaltera.transactionannouncement.com. I'm joined today by Brian Krzanich, our CEO, Stacy Smith, our Chief Financial Officer, and Steve Rodgers, our General Counsel. In a moment, we'll hear brief remarks from both Brian and Stacy, followed by Q&A. Today's presentations contain forward-looking statements. All statements made that are not historical facts are subject to a number of risks and uncertainties, and actual results may differ materially. Please refer to the risk factors set forth at the end of this presentation for risk factors specifically related to this proposed transaction and the business combination.
This presentation, including the risk factors, will be posted on our transaction website, which is the site that Intel and Altera will use to post all information about the proposed transaction. Please also refer to Intel's and Altera's most recent earnings releases and Forms 10-Q and 10-K filings for more information on the risk factors related to our respective businesses. If we use any non-GAAP financial measures during this presentation, you'll find the reconciliation to the most directly comparable GAAP financial measure on our transaction website. Altera will file a proxy statement with the SEC in connection with this solicitation of proxies from Altera's stockholders. Stockholders are urged to read the proxy and any other relevant documents that Altera files with the SEC when they become available, because they'll contain important information.
Proxy statements and any other documents filed by Altera or Intel in connection with the proposed transaction will be available at the SEC's website, sec.gov, and at the transaction website. With that, let me hand it over to Brian.
Thanks, Mark. I want to start by just saying I'm really excited to be here today announcing our agreement to acquire Altera. I'd like to begin by offering some context from my perspective to illustrate how closely this acquisition is to our growth strategy. This slide that we're showing right now is from our last investor meeting in November of 2014, when I laid out our growth strategy, which is to use our core strengths to enter profitable and complementary markets. Specifically, Intel continues to drive Moore's Law, which has propelled the technology revolution in improving semiconductor performance and power and cost reduction, enabling an integration of more and more capabilities into our devices and leveraging shared IP across the product family. This acquisition is a perfect extension of this strategy.
By bringing together our leading processors with Altera's hardware of programmable FPGAs, we can make the next generation of semiconductors not just better, but truly able to do more.
Good morning, everybody. This is Stacy. I'd like to just add my welcome to Brian. I'm going to take a second to just give you a very high-level snapshot of the transaction details. This is going to be an all-cash transaction at $54 per share. We intend to fund the acquisition with a combination of cash from the balance sheet and debt that we expect to raise. We're targeting a close in six to nine months, and we expect the transaction to be accretive to Intel's non-GAAP EPS and free cash flow in the first year after close. The transaction has been unanimously approved by both the Intel and Altera boards of directors. Now let me turn it back over to Brian to go through some of the strategic rationale of the deal.
Thanks, Stacy. As I mentioned earlier, this acquisition is very well aligned to our strategy. Just as importantly, Altera is very well positioned as a market leader in the FPGA space. When we combine that product leadership with our IP portfolio, we'll be in a position to enable new classes of products that meet emerging customer needs in the data center and the Internet of Things. We believe that this combination will create significant stockholder value in the process. I want to give you just a quick overview of Altera and just kind of give you a little bit about who they are and what they do. As we've talked about, they are a world leader in programmable logic devices. They're headquartered here in San Jose, in fact, only a couple of miles from our headquarters.
They have a little over 3,000 employees in more than 20 countries and have about 12,000 customers served through direct sales force and distributors. Their fiscal year 2014 revenue was about $1.9 billion and a gross margin of 66%. I want to talk a little bit about the current Intel foundry relationship. It was established in 2013 as a multi-year agreement, and the idea was to build world-class PLDs on our world-class silicon technology. The products that we were going to manufacture are Altera Stratix 10, and those SoCs will be using Intel's 14 nanometer Tri-Gate process and packaging technologies. We believe that the result is using world-class silicon and combining the world-class engineering teams of the two companies. Over time, these products were going to become leaders in the marketplace. Additionally, Altera has a very attractive base business.
We serve a wide range of customers which are, in large, high growth market segments. The telecom and wireless customers represent approximately 44% of Altera's 2014 revenue, with industrial, military, and automotive filling out about 22%, and networking and computing filling out the remaining at about 16%. We currently have strong market share segment in about 40% of the FPGAs. I also want to highlight here that we know many of Altera's customers. In fact, we do business with eight of the top 10 customers that Altera has already as Intel. Finally, we increasingly see FPGAs and the integrated CPU plus FPGA products as alternatives to ASICs and ASSPs. The green chart on the right of this page makes this clear, the opportunity in both of these market segments. Part of what we talked about is that this is a growth segment.
When you look at this growth segment, we wanted to look at it both backwards in time, but forwards projecting. We've built this model and we've built this acquisition on pretty much a normalized 7% compounded annual growth rate. This is across a wide range of market segments, as you can see from the chart there. We've talked so far about the attractiveness of Altera's existing business. I'd like to now spend a little bit of time sharing with you some of the synergies we expect to realize. We've broken these out into two categories. Now I'll speak about the product synergies, then I'll ask Stacy to address the cost and manufacturing piece. There are two segments within our portfolio in which new classes of integrated FPGA products are expected to be especially impactful, and those are the data center and the Internet of Things.
Within the data center, we see an opportunity to combine our Xeon microprocessors with FPGAs to significantly improve performance and, through integration, reduce the cost. That combination will position us to address emerging workloads in creative new ways, adding value to our customers. These are products that our customers have been looking for. In IoT, integrating Atom with FPGAs will allow us to pursue segments that are now served primarily by ASICs and ASSPs, expanding our serviceable market by roughly $11 billion. Combined, we expect these product synergies to drive roughly 60% of the value we create with this acquisition. I'll share more specifics on each of these value drivers in just a minute, but before I do, I wanted to ask Stacy to touch on the cost and manufacturing section.
Thanks, Brian. On top of the benefit that we get of bringing out this new class of integrated products, we also expect meaningful reductions in OpEx and improved competitiveness in Altera's existing product line by bringing them onto our manufacturing processes and using our IDM tools. The OpEx reductions will be primarily focused on G&A, and those cost savings will grow over time. On the manufacturing side, we expect to create significant value by working with the Altera team to ensure that their products are consistently first to node, taking advantage of the best process technology in the world. The OpEx reductions and the benefits of manufacturing leadership combine to produce the other 40% of the value that we expect to create.
All right. Let's take a minute now on some of the specific applications, and let's talk about the data center and give you a little bit more detail of just how these products are going to improve the marketplace. For some workloads in the cloud especially, and I've provided a few examples here on this page, FPGAs can significantly improve performance and cost. What's especially interesting is what happens when you integrate FPGAs with our Xeon microprocessors. That combination, by integrating the two products together, improves performance by 2X, and at the same time reducing cost even further. FPGAs also significantly improve the flexibility for our customers, allowing them to quickly implement and update their algorithms. We expect limited shipments of co-packaged Xeon microprocessors and FPGAs in the latter half of 2016, and those will be followed over time with on-die integrated solutions.
By 2020, it's estimated that up to one-third of cloud service provider nodes may use these FPGAs. The next example I wanted to do was turn to the IoT and talk about how FPGAs will help in that space. Using the FPGAs in the Internet of Things will create new growth opportunities by integrating FPGAs with our Atom microprocessors. Here again, integration will provide our customers with improved performance, cost, and flexibility. More specifically, with integration, FPGAs become cost competitive against ASICs and ASSPs, opening up an $11 billion incremental SAM, we believe, by 2020. We'll have the option to preload accelerators for specific verticals, like the industrial automation example I've shown here. These pre-programmed SoCs will combine the benefits of a familiar Intel architecture and workload-specific accelerators that might otherwise be found in ASSPs.
We can take that very same part, that same integrated Atom plus FPGA SoC, and offer it as an ASIC replacement for implementations like driver assistance systems that I've highlighted here as well. That would allow the automaker to define, load, and update their unique IP on a real-time basis. Stacy earlier mentioned the value of advanced manufacturing in the FPGA market segment. Altera has an already strong product portfolio that we think that the product portfolio gets even better when manufactured on Intel process technology. We also anticipate improvements in design speed and effectiveness as a result of having Altera and Intel's engineers all under one roof, working together even closer. Just as our own products benefit enormously from our IDM business model, we expect similar benefits for Altera.
While a foundry relationship provides some of the benefits of manufacturing leadership, this transaction unlocks greater value and ensures that the value accrues to Intel stockholders. That covers the primary value drivers. Before I summarize and move on to a Q&A, I'd like to talk about our integration plans for the company. One of the things that gives us confidence about merging our two companies is that we already have a history together via our foundry relationship. That relationship has resulted in strong engineering engagement. Organizationally, we'll be setting Altera up as an integrated Intel business unit with a dedicated sales force and engineering team. Adding to our confidence, Altera has retention plans in place that will help us to ensure a smooth transition. And once that transition closes, the transaction closes, we'll implement longer-term retention plans from Intel. I'd like to move to the summary now.
This is a big day for both Intel and Altera. I'm looking forward to the opportunities that we can create together. We're buying an already strong business and a strong team, and we think with our technology, we can make that even better. That, combined with the innovative new products in the data center and IoT, will enable us to create significant value for our owners. With that, I'd like to turn it back over to Mark for Q&A.
All right. Thank you, Brian and Stacy. Moving on to the Q&A, because our time is limited here today, we'll ask each participant to limit themselves to just one question. Dave, go ahead and please introduce our first question.
Thank you. The first question comes from C.J. Muse at Evercore ISI. Please go ahead.
Good morning. Thank you for taking my question. First question, thinking through the data center, can you talk how this grows your SAM there, and then why you made a decision to bring Altera in-house as opposed to simply continuing to maintain your current joint venture with them? Thank you.
Sure. Think about that when you integrate the FPGA with the Xeon processor, you're creating a new class of products or a set of products that don't exist today, where you're able to move the software algorithms, things like facial search or encryption, down into the silicon and improve performance, as you said, greater than 2x. That's really what this is creating. That's a new class of products, and that's allowing us to provide a set of products that don't exist today. You said why move them inside versus continuing that in just a foundry relationship?
No, sorry, in terms of the work that you were doing, I believe the HARP research platform that you were working with them.
I'm not going to talk about specific products like that. Let me just talk about the sequence of events. As we said, each time you integrate further and further from being on board on the same motherboard to building a co-packaged product, which is what that product that you just mentioned was, to building a piece of silicon that's a monolithic piece of silicon where the Xeon processor and the FPGA are one piece of silicon. You get significant performance improvements in each one of those integration implementations.
We believe that in order to go to the full extent where you build a monolithic piece of silicon to get the maximum performance cost, power reduction, footprint reduction, all of those improvements that our customers really want and the market needs, we need to integrate the company because you're going to have such a tight relationship between those two intellectual properties, those two pieces of IP, that you have to work within the same team. That's why we wanted to move them inside.
Very helpful. Thank you.
Thanks. Your next question is from the line of Ross Seymore at Deutsche Bank. Go ahead, please.
Guys, thanks for letting me ask a question. Stacy, a couple for you. Any size on the cost synergies? Just generally, can you talk about how Intel thought about the difference between the accretion you'd get from using the same amount of cash to buy back your own stock versus doing this deal?
Sure. Happy to talk about both of those, Ross. As we went through, the combination of cost and manufacturing synergies is about 40% of the value that we expect to create, then the other 60% come from this new class of products that we can bring to the market. Looking specifically at that 40%, the cost synergies is a piece of it. It's not the majority, but it's a piece. For us, the cost synergies are primarily in the G&A space. That's where we expect to realize most of them. Then the other portion of that bucket, that 40% bucket, is as we advance from node to node, we'll continue to make their product line better, and we think that just gives them a stronger competitive position in the marketplace.
On the priorities of capital, this is very consistent with what we've talked about over the last several years, which is, first and foremost, invest in our business. When we look at this deal, this is a great example of that. We believe that we can create significant value for the shareholders, it will result in a good return for you and a thing that only we can do because of the bringing them into our manufacturing facilities and the ability to integrate into a product that's coupled with some of our other IP blocks.
Thanks, Ross.
Sure.
Thanks. Next question is from John Pitzer at Credit Suisse. Please go ahead.
Yeah, good morning, guys. Congratulations on the deal. Just quickly, Stacy, how much debt to do the deal? Brian, you were very clear about the SAM expansion opportunity that Altera brings in IoT, a little bit less clear in the data center. I guess I'm trying to get a better understanding relative to the data center, to what extent this is offensive and TAM expansive versus defensive, trying to protect the footprint that you already have as FPGA gains more and more traction on workloads within the data center. Thank you.
Sure. You were a little faint there, John. Was your question how much of this is debt versus cash?
Correct.
Yeah. We're not breaking it down. As you know, we have significant cash on the balance sheet, significant cash available to do this. It'll be a combination of debt and cash. Beyond that, we're not breaking it down to be more precise.
On your second question, John, we do not consider this a defensive play or move. We look at this in both the IoT and in the data center as expansive. These are products that our customers want built. We said that approximately 30% of the cloud workloads would be on these types of products as you exit this decade. That's a guesstimate on our part based on how we see trends moving, and where we see the market going. This is about really providing the capability to move those workloads down into the silicon, which is going to happen one way or another. This, we believe, is best done with the Xeon processor FPGA combination, which will clearly have the best performance, cost, and footprint for the industry.
This is truly about, in IoT, expanding into new available markets that are currently ASIC and ASSPs, and in the data center around moving those workloads down into silicon and really continuing the growth of the cloud overall. I do not consider that a defensive direction.
Thanks, John. I'll remind everyone again that we're asking everyone to limit themselves to just one question.
Next question comes from Christopher Danely at Citigroup. Please go ahead.
Hey, thanks. I guess good evening, guys, since everyone's in Taiwan. In terms of the price, $17 billion, can you just talk about, I guess, the internal metrics you used to justify the price of the purchase?
Yeah. Can we create value? At that price, we believe we create significant value for our shareholders. Just to reiterate what Brian said at the beginning, the value drivers here are, this allows us to build a new class of products, both for the data center and the IoT business. Both of those businesses are pretty exciting to us. We get some cost synergies, and we get this manufacturing advantage that builds over time as we continue to bring out new generations of FPGAs on our leading-edge process technology. Thanks, Chris. Operator, please go ahead and introduce the next question.
Next question is from Jim Covello at Goldman Sachs.
Great.
Please go ahead.
Thanks, guys. Appreciate the opportunity to ask a question. Brian, question for you. Obviously, there's tremendous value as you're highlighting in the accelerator market, and you're talking about 30% of the workloads being applicable here. There's FPGA accelerators, there's graphic accelerators. How do you think about what your combined offering is going to be compared to what the competition would be going forward then in the form of a graphics accelerator? Thank you.
Sure. There's different workloads that are benefited, I'd say, from different kinds of accelerators. You really have to look at what's your adjustable market. Graphics accelerators, because of the way that architecture works, tend to be more in the high-performance computing space, and they tend to give you more of an acceleration in that space, just because of the way graphics works and how the graphics architecture is laid out. FPGAs are going to be used for accelerators in places where it's more of a logic kind of accelerator, so things like facial search, encryption, that aren't linear in nature. They're very different workloads that are going to be applied. The other thing that an FPGA gives you is the ability to program, so you can adjust the software, basically, that you're using and the acceleration model.
You can do that both in something like a facial search algorithm. You could, over time, make improvements to your facial search algorithm, and as a result, program that into the FPGA. With a graphics fixed die, you're not going to be able to necessarily do that. All of that software occurs on the outside. The other thing is you can actually move from workload to workload with an FPGA. You can move with the same system from accelerating facial search to accelerating an encryption model, and you could literally do that between sequences, basically on the fly. It would be very difficult, again, to do with a graphics system. I look at them as two different kinds of workloads. By the way, there are workloads that are still better done by ASICs and other products.
You saw an announcement we made a couple weeks ago with the ASICs. Same kind of thing, where there are going to be workloads that are better driven by specific algorithms like that.
Really helpful. Thanks so much. Good luck.
Thanks, Jim.
Next question is from Timothy Arcuri. Go ahead, please.
Thank you very much. I actually had two questions. First of all, does this mean that Altera will be fully on Intel at 10 nanometer? I'm also wondering whether the CapEx requirements that you've talked about for 2015, whether this changes those. Thanks.
Sure. They're somewhat related. The CapEx requirements for 2015 should not change because those CapEx requirements are kind of laying out the capacity for 2016 and 2017. We already had Altera built in as a foundry customer. That was one of the real strengths in doing an acquisition of Altera. They were already in our silicon, in our design methodologies. It doesn't really change the capital requirements as a result. The new products that we've talked about start to occur in 2016 as a co-package part, in the second half of 2016, then really ramp in 2017 and 2018. Again, that's kind of outside this capital cycle. You asked about 10 nanometers. Again, that's an independent decision. Altera is an independent company until this deal closes.
You'll have to talk to them about their decisions to make, where they choose to implement 10 nanometers, and we'll support those decisions. We think our technology is good, and we have a good shot at being the chosen one for that, but that's going to be their decision on that.
I'd have to say, I like our chances to win it. They're an independent company, but I think we're well-positioned.
Thanks, Tim.
Okay. Next question is from Joseph Moore at Morgan Stanley.
Great, thank you. I wonder how you're going to think about the gross margin of the business, particularly when you start to foundry the product yourself. You'll stack a foundry gross margin on top of the already high 60s gross margin. Will you therefore have kind of an 80% gross margin on that type of business, or will you use that incremental margin to price more aggressively?
Yeah. I'll talk philosophically. This is Stacy. To the specifics of your question, it's going to take a while for this deal to close. Likely, post-close, you'll see us being pretty explicit about the implications of this acquisition once it closes on our ongoing financials. I'm not going to go there yet. In general, though, when we look at Altera, one of the things that was attractive to us is we like the business. It's a business that benefits from technology. We like the profit profile of it. We're not planning on significant changes there. We think that we can enhance their business by landing them on our process technology, advancing their product line even faster. We think over time that gives us a competitive benefit in the marketplace. We're not planning to come in and massively change the business model.
We think that's one of the things we bought.
Great. Thank you very much.
Next question is from the line of David Wong at Wells Fargo.
Thank you very much. Coming back to your point about they remain an independent company, are you able to have teams start working on single chip designs with PLD circuits on the same chip of processors now, or do you have to wait till the acquisition closes before this happens? Would we expect to see the first single chip processor PLD products begin to come out?
I'll let Steven Rodgers talk about what we can and cannot do prior, then I can talk about the road map.
David, during the closing period, the companies will continue to operate as independent companies with an arm's length relationship. Anything we would do with them would need to be handled in negotiations in arm's length capacity.
Your comment about, or your question about when will we see products. We said that the first products, which are co-package, where two pieces of silicon in the same package would start to occur in the second half of 2016 and really ramp in 2017. Limited availability in the second half of 2016 and then a strong ramp in 2017. We haven't given an exact date for when we'll have a single monolithic piece of silicon, but you can think about it being a little bit after those kinds of dates.
Great. Thanks very much.
Thanks. The next question is from Christopher Rolland at FBR. Please go ahead.
Hey, guys. Congrats on the transaction. Altera played at the lower end of the market where Xilinx didn't play. You guys mentioned IoT. It looks like it's industrial automation and ADAS there. To further penetrate IoT, you might have to move downmarket more. Have you given this any sort of consideration whether you guys are going to move downmarket, stay there, or perhaps penetrate it a little bit more?
When you say downmarket, I just want to make sure you're talking about lower cost products?
Exactly.
Okay. First, if you take a look at where FPGAs are going to be applied, you take a look at the industrial automation, I used that as an example. I used ADAS as an example, the automotive assisted driving. Those are not necessarily downmarket. They actually have very good margins and are using a lot of compute capability. Those will be clearly our first targets. If you take a look at our IoT strategy overall, we have a set of products like Quark, Atom, and then Core that allow us to, we think, move across the market ranges. I believe we can go, using your terms, fairly downmarket. If those downmarkets want an FPGA that's scaled appropriately to that size, we would absolutely provide it at that level as well.
I don't think we've completely investigated how far we can implement on this FPGA roadmap. Certainly if a Quark plus FPGA makes sense, we'll now have the ability, once this acquisition closes, to provide a product like that. That is what I would consider then a more downmarket type of application that you're talking about.
Yeah, I would just add, Chris, one of the things we're excited about, when we hear from customers, the reason they want this is because of the performance it gives, the total cost of ownership from their perspective, it lowers their total cost of ownership, and it reduces their time to market and improves their flexibility. Those tend to be things that we can get paid for. I think we can address large segments of the market, and do it in a very accretive fashion.
Great. Thanks, guys, congrats again.
Thanks.
Next question is from the line of Ambrish Srivastava. Go ahead, please.
Hi, Stacy and Brian. I'm really struggling with some of the assumptions that you're laying out. Altera has had a negative CAGR for the last three years. Okay, we can take a longer-term view, and they have barely outgrown the industry. You're saying a 7% target is, if I heard you correctly, Brian. Then you layer on the fact, I'm going with what Altera has publicly said, a billion-dollar opportunity for the co-processor market. It's very small compared to the already $14 billion in change for the DCG. How do you get to that 7% CAGR? How do you make the math work, Stacy? I don't think you answered Chris Danely's question on what are the metrics that you used for return on the capital that you're deploying to this deal. Thank you, sir.
Look, I'll start with a little bit of the strategic side and a little bit of the view of their business, and then I'll let Stacy get into more of the financial detail. I think when you take a look at it, as you said, they said, well, gosh, the co-processor segment's worth $1 billion. When you probably talk about a standalone part that sits on a motherboard, isolated, I'm not going to absolutely give credit to that number, but it's probably not nearly as big a number as what we're talking about. It's probably not an unreasonable number value of $1 billion. What we're talking about is significantly improving the performance beyond that. Going to the fully integrated part and providing something that's 2x the performance for that. We think that opens up the market quite a bit more.
That's where we said approximately 30% of the workloads in the cloud, we think will have this capability or require this capability as we exit this decade. That's going to be quite a bit bigger than that billion-dollar number, when you take a look at it. How do we get to the 7%? We believe that when you look at the FPGA market, it's all about providing leading-edge silicon. If you take a look at the historical, the person who's brought the products on the leading-edge silicon technology first and have brought the right products to that marketplace, they were designed correctly, right performance and power and cost along with that, tended to win share and tended to grow.
We believe by integrating, as we said, we can move even quicker to the leading node, and help them with their design coming quicker because their architects are now working hand-in-hand with our architects. A lot of this is about how good you are with the tools and how good you are working with the silicon. That's a place we can bring the strength over to them on our leading-edge silicon, right? They're very good- Using other silicon at all. We think that combination provides the right growth, that 7% number, and the numbers we've used for data center and IoT growth. For the rest of the financials, I'll let Stacy answer.
Actually, I thought you did a brilliant job. The only thing I'd add to that is to put the value driver in perspective that we're looking at on the product side of this is, as Brian said, you get this two plus X increase in performance. Through integration, you get a massive decrease in cost. We think that opens up a really broad swath of the market that will want to take advantage of this capability. We said earlier on the call that we think that it will grow to be a third of the cloud customers. Think about the size of that market, how fast-growing it is. That's a very significant value driver for us.
Thanks, Ambrish.
Thanks. The next question comes from Ian Ng. Please go ahead.
Yes. Sounds like you're placing a lot of hope on this monolithic CPU and FPGA being a much better solution to go after ASICs and ASSPs. You've had Stellarton announced since 2010 as the FPGA co-package. If the integration doesn't work, would you actually consider going after ASICs and ASSPs using Altera IP, actually doing those kind of parts?
Wow, okay. That's a little bit of convoluted question. Let's step back for a second. The parts we talked about in the data center aren't really replacing necessarily ASICs and ASSPs. What we're actually doing, if you take a look at something like facial search, what happens right now is those algorithms sit out in memory and storage, and they're coming back on and off the CPU on some regular basis as the algorithm is applied. What we're allowed to do now is actually put that algorithm directly in the silicon with the FPGA. Now they're only going out to get the data, the massive number of faces they're looking at, pulling those down in and applying the algorithm real time on the silicon piece. That's not something an ASIC or an ASSP is doing today.
It's something that's basically moving in and out of the CPU today. Where we said ASICs and ASSPs were being applied is something like assisted driving, where there's a specific ASIC now that somebody's built that has all the algorithms that watch all of the sensors on a car and look at something like lane changes and the sensors that sense whether somebody's driving up alongside of you, and put all that together and shake your steering wheel or flash lights that says, "Don't change lanes or you're going to hurt somebody." That is done by an ASIC or an ASSP today. What we can do is move that workload off those ASICs and ASSPs at we believe an equal or lower cost, and onto the FPGA.
What that also does is allows people to update those algorithms real time, not have to go out to their car line and do adjustments to the hardware, but actually be able to do that as the car's shipping out the door in software. They can continue to make improvements in the safety and the quality of that vehicle as a result. Those are two different workloads and two different scenarios.
The cost piece of this becomes really important because we're now getting to the point on the process technology, with our process technology leadership, that the cost difference at the silicon level is being more than offset by the lower development cost and mass cost and everything else. It's unlocking more and more of the ASIC and ASSP market where these integrated FPGA solutions are cost competitive. We think that on the IoT side, that opens up pretty significant opportunity for us, a big chunk of SAM.
Did Stellarton go after any of those opportunities that you listed, or it just wasn't the right product at the time?
Well, Stellarton prior generation, relatively expensive. Think of it as on this evolutionary path, but what we're planning to do with Altera will be significantly more evolved.
Okay. Thank you very much.
Thanks, Ian. Operator, we have time for two more questions.
Thank you, sir. The next one comes from Matthew Ramsay. Please go ahead.
Yes, thank you for taking my question. Brian, maybe you could talk a little bit about how this acquisition might be an extension of some of the customizable or the customized or customer specific Xeon SKUs that your company's already doing in DCG and might open up some opportunities for customers that might not be so keen to give you guys access to the algorithms to do customized stuff and give them a little bit of a sandbox to play with on that.
Yeah. Wow. I should have had you help me write my marketing foil. You almost gave me the answer there in the question. I mean, you're absolutely right. What this does, it not only allows people to have a customized part without having to hand me over their algorithms. To be honest, they can do that in multiple ways, and they don't necessarily have to give me those algorithms per se, in order for me to build those into a customized silicon. They can either design that segment of the silicon themselves or give it to a third party that can do the design for them that hands us simply the circuit layout, so to speak. This does give them, as you described it, a playground. You can think of an FPGA as a large sea of gates that they can program now.
If they think that that algorithm may change over time as they learn and get smarter, or they want to be more efficient, they don't have enough volume to have a single workload on a single piece of silicon, they can use an FPGA to have accelerators of multiple segments, facial search at the same time as doing encryption. We can basically on the fly reprogram this FPGA literally within microseconds that it's occurring on the CPU itself. That gives them a much lower cost and a much greater level of flexibility than a single customized part. You'd have to have quite a bit of scale and need for that singular part that only is going to do one workload. That's where we find people are really interested in this versus just a customized piece of silicon.
They want to be able to do multiple workloads, or they know that workload's going to grow and shift and change over time. Those are really ideal for this kind of application.
Thank you very much. Congrats again.
Thank you.
Operator, can you please go ahead and introduce our last question?
The last question comes from Harlan Sur.
Hi, good morning. Congratulations to both sides on the deal announcement. Lots of discussion on data center. IoT and embedded has been a key focus area for both the Intel and Altera teams. Can you just talk about a timeline you can walk us through to help us understand when the team can bring an integrated processor plus FPGA solution to the IoT market, integrated all into a single piece of silicon? Then on the manufacturing side, obviously the team has the 14-nanometer partnership. I assume that on a go-forward basis for new manufacturing technologies, these will all be primarily Intel manufacturing. Can you just tell us how long it's going to take or what the plans are to migrate the existing product base over to Intel internal manufacturing, just given the long product life cycles of the FPGA platforms? Thank you.
Sure. Okay. I think there was a couple of questions in there, but let me try. I don't see a real need or desire to move existing products into the Intel silicon. They're yielding well, running fine. Most of them, I believe, are at TSMC, good partner of ours. There's not really a need to go do that. We're going to focus on the future, and the future is both driving their future products at a faster rate to the better node or to the newer node. That'll be those existing products or those next-generation products, and then these integrated products for both the data center and the IoT. We said that the first of the data center as a co-package would be in the second half of 2016, ramping in 2017.
We said, I think, shortly after that, products that are monolithic or on a single piece of die. That's slightly after that. IoT, we're still trying to understand whether co-packaged parts make sense, whether there's enough cost savings, whether the workloads are there, or whether we have to move all the way to that integrated part to hit the cost envelope and footprint that they need. We just haven't finished that analysis. Part of it's just so workload-specific, and you have to design a part relative to the workload and all of that. We just haven't laid that out yet. That's one of the projects we'll really be working on. We know what it can do and how it can improve the performance of those devices.
We know that an integrated part will provide significant, and you'd see that integrated part probably around the same time the data center part's fully integrated would be. The question is, will a co-packaged part be in a good interim? We just don't know those answers right now.
Great. Thanks for the insights, and again, congratulations.
Thanks a lot.
Thanks, Harlan. All right. Thank you all for joining us today. Dave, please go ahead and wrap up the call.
Thank you for your participation in today's conference. This concludes the presentation. You may now disconnect. Good day.