While there has been a tremendous amount of focus on Intel (NASDAQ:INTC) integrating connectivity and communications into its mobile system-on-chip products, (in order to more effectively compete in the smartphone and tablet markets), many seem to miss that such integration could prove to be a massive boon for Intel's PC chip business. It could also serve to distance it from its only real competitor -- Advanced Micro Devices (NASDAQ:AMD).
The real value of integration
When Intel launched Sandy Bridge in 2011, the company saw an immediate ASP uplift as it gained further PC content share, and it saw a mix shift within its product stacks as customers were willing to pay a premium for higher end integrated graphics solutions. Take this snippet from Intel's 2011 fourth quarter conference call:
From here, our CapEx, really, will be a function of 2 things: it's the unit growth that we see going forward, and it's the speed at which we bring capabilities to that leading-edge process technology. And I think graphics has been the classic example of that. As we move that to the process edged to the leading-edge process technology, it did drive an increase in CapEx but it also drove a really capable product line that we got paid for.
The idea is that as more functionality is integrated onto the main processor complex, the more value that particular chip has. So, in the example of integrating graphics onto the die, notebook and desktop vendors could cut out a low end discrete GPU from their designs -- saving money. Intel, realizing that a move to integrated graphics would save OEMs significantly on everything from additional graphics-specific RAM, cooling solutions, board space, etc., could then charge a premium for its integrated part and still end up with a mutually beneficial arrangement.
Bringing this to connectivity
It's hard to imagine a laptop or even a desktop PC that doesn't connect. A laptop without Wi-Fi is near useless, and a desktop without some kind of LAN (either Wi-Fi or Ethernet) is extremely limited. Today, all PCs either come with third-party chips (or are Intel-designed) to handle Ethernet/Wi-Fi. While higher-performance PCs are likely to be better off with discrete networking solutions (as these can be higher performing/higher quality), the majority of low-cost PCs would be just fine with Wi-Fi/Ethernet built into the main processor.
The benefit for Intel is pretty obvious: it can drive up chip ASPs (at constant gross margin levels) and absorb the value that the discrete chip vendor, which was likely to be someone other than Intel, was getting on its products. It would also simplify the mainboard design and reduce the number of components needed on the mainboard. It would, at least for the majority of mainstream systems, be a "win/win".
This would grow competitive advantage over AMD
By integrating connectivity onto its PC platforms, Intel's chips would offer a material integration advantage over AMD. While AMD appears to have Ethernet IP that it could conceivably integrate onto its PC processors, it does not have Wi-Fi/Bluetooth/NFC. Intel would be able to integrate this on its 14-nanometer generation.
Intel's growing process lead, which manifests itself as improved an cost/transistor, means that this integration can be done in an increasingly economical fashion. With the cost/transistor difficulties that Taiwan Semiconductor and the other foundries are facing, AMD may find it difficult to justify doing all of this work in what is a seemingly dwindling position in PCs. That said, if AMD is serious about tablets, it will need to go down this path.
At this point, Intel should be very aggressive in integrating new functionality and absorbing as much of the silicon value into its processors as possible. The company has been doing a good job, relative to AMD, but compared to the level of sophistication that we are seeing from the mobile SoC players, Intel still has a ways to go. However, that necessity to catch up in mobile could pay off handsomely in the PC space.