Advanced Micro Devices (AMD 3.46%) showcased its next-generation Beema and Mullins APUs at CES 2014. Now, unlike Intel (NASDAQ: INTC), which failed to show any interesting silicon at the event, it was encouraging to see how Beema/Mullins offered substantial improvements in graphics performance and power efficiency over previous generation APUs. However, unlike Intel, AMD has a rather sizable gap in its IP portfolio that it will need to fill to compete in the tablet segment.
Connectivity and communications
It's clear that the vast majority of tablets that will sell going forward are in the $100-$200 range. To adequately service these markets, chip vendors will need to integrate connectivity (Wi-Fi, Bluetooth, GPS, NFC, etc.) as well as communications (3G/4G LTE technologies) onto their chips to create cost-competitive products.
This is an area in which Intel, the leader in PC chips, has struggled for quite some time. However, with its upcoming SoFIA product, the company claims that it will have the "most highly integrated chip on the market." While Intel, at the very least, bought Infineon (as well as a number of smaller companies) simply to acquire this connectivity IP, AMD does not appear to have much in the way of in-house connectivity.
Today it's a problem for tablets, tomorrow for cheap PCs
The trend in computing has been toward higher levels of integration and lower cost, and given AMD's position as the leading vendor of PC processors in the sub-$399 PC space with over 50% market segment share, this is going to become more important long-term. In Android tablets, leaders such as Qualcomm already ship products with integrated connectivity and communications.
However, in Windows tablets, neither Intel nor AMD has integrated connectivity/communications, so there's much less pressure. However, given that the PC market is in decline, both Intel and AMD must look for ways to capture silicon content share in PCs. An obvious route that Intel will likely be taking with its next-generation "Broxton" part (which will span high-end phones to low-end PCs) is the integration of connectivity.
AMD, unfortunately, doesn't seem to have a robust connectivity portfolio. On a longer-term basis, the company will need to acquire such IP to integrate into future system-on-chip products. It is unclear from whom AMD could buy in order to acquire the IP and, if there existed such a company for sale, if AMD could afford it, given its current financial situation.
Foolish bottom line
The good news, though, is that integrated connectivity is unlikely to be necessary across the gamut of the PC market -- just the very low-end/cost-sensitive markets. However, connectivity and communications are a key part of the smartphone and tablet world, which is part of why Intel has failed to gain traction in phones thus far. The integration of that IP, particularly in lower-cost models, will prove a necessity over time. Can AMD secure some connectivity IP going forward? Only time will tell, but this is something to keep in mind.
