Intel's Mojo Returns

Intel has lost a lot of momentum with the ongoing weakness of the PC industry. To be fair, much of this lies on the back of Microsoft, which failed to aim Windows 8 correctly, causing the PC sales momentum to stumble badly at a time when a recession and cheap Android tablets were eroding the market.

Intel’s foray into smartphones also suffered from software issues, primarily by not supporting Android from the start. These problems resulted in the rise of the ARM chip and the decline of the desktop x86.

Still, Intel has huge resources and a pretty good sense of the issues. More to the point, Light Blue is finally putting together a sensible set of solutions that should help it bring back marketshare in emerging markets while staving off challenges in its very profitable and market-dominating server business.

Now, what do we have as serious contenders? Let’s take a look.

Quark will enable Intel to enter the embedded and wearable markets. These are really low-power chips, and they look good up against MIPS and ARM. This market is wide open, huge, and just about to take off in concert with the Ethernet of Everything.

New Atom chips solve the mobile power issues in neatly by turning functions on only when needed. ARM does this, too, but Intel claims better overall power, and in the mobile space, that’s fighting talk.

Some of those Atom chips will be built as targeted SoCs for storage and networking gear. This is important for achieving small-footprint gear in the SMB market.

Variations of the Atom are targeting the nascent microserver market, with high core counts delivering much better performance per node, and yet more SoCs.

Nvidia is opening up its intellectual property to allow its GPU cores to be built into SoCs with standard CPU cores. Nvidia plans ARM-GPU hybrids, while AMD is already shipping x86-GPU solutions. These are going to grow up, with high-powered GPUs in the mix, and it’s a fair bet that Intel is adding its own x64-GPUs to next year’s pipeline.

These units are strong contenders for big data analytics business, which is the fastest-growing IT segment. They also fit the high-performance computing market.

CPU+memory modules are expected next year. One consequence of this type of approach is that the CPU can pack in many more cores, so even x64 servers will speed up a lot.

Intel has tackled the OS issue, too. I’m sure Paul Ottelini’s parting advice was "If you can’t beat them, join them." The next generation of notebooks will be 2-in-1 and will offer the ability to be a tablet, with a keyboard attachable, and price points will be very competitive. There are even versions that can double as a monitor, with an HDMI-IN port.

But the real wizardry is these devices will run both Android and Windows, in a take from Apple a few years ago. That will allow vendors to service four markets with a single product family, and that’s a very strong statement.

All in all, Intel is tackling its demons well now. It will get a lot of respect for new ideas and new approaches, and that will get it business in all these areas. But it will take a bit of time to change the bottom line, and a recent delay of a quarter in the Broadwell XEON product isn’t beneficial while the decline of the PC appears to be accelerating.

These are indeed interesting times.