IHS expects shipments of low-end smartphones to rise at a compound annual growth rate (CAGR) of 51 percent from 2011 to 2016. The firm expects shipments of high-end smart smartphoens to grow at a CAGR of only 12 percent during the same period.
"In the emerging markets, optimizing the cost/performance balance will be critical for success," Sideco said. Intel will also need to leverage its 2011 acquisition of Infineon's wireless chip unit to ensure that it offers the best solution for low-end smarphones, incorporating both the applications processor and the modem on the same chip, Sideco said.
To date, Intel has made minimal progress in cracking the smartphone applications processor market, which is dominated by Qualcomm. In basebands, Intel ranks third in market share, according to IHS, generating just 8 percent of baseband revenue in the third quarter of 2012, well behind both Qualcomm and Taiwan's MediaTek, according to IHS.
"While Intel dominates the PC microprocessor market, in the smartphone semiconductor business the company has no place to go but up," Sideco said. “And while Intel certainly faces major challenges in achieving the kind of leadership position in mobile handsets it now has in PC semiconductors, the company appears to be serious about building its competitive positioning in the smartphone chip market."
For what it's worth, IHS expects China to remain the fastest growing market for all cell phones, with shipments expected to grow at an 8 percent CAGR from 2011 to 2016. The rest of the Asia-Pacific region is predicted to have a 6 percent CAGR in cell phone shipment growth over the same period, with the EMEA region coming in third at 5 percent, and North America at 4 percent. Clearly, future growth is coming from the emerging regions.
With Atom, the question used to be whether it was low power enough. Assuming the answer is now yes, then for success at the low end of the smartphone market, one would expect the next question to be whether it is lower cost or no costlier than an ARM-based apps processor.
If so, that would also be decidedly un-Intel.
Intel should reduce the amount of instructions that an Atom processor supports w.r.t x86 ISA. Otherwise the core will have issues like :-
1. Bigger area for implementation
2. More power consumption (means more/better battery tech)
3. More core / processor(or SOC) on chip will have the above 1 & 2 issues in a much bigger scale
4. Cost will always be higher because of 1 & 2 points
5. Going to leading edge process will reduce issues of points 1, 2 & 3, but how to take care of point 4 !
So, Intel should support the required subset of x86 ISA very much like what ARM, Qualcomm and others have.
But having said that, many x86 (PC) software will not run by default which is Intel's biggest selling point in the name of Atom processor which ARM processor clearly lags.
Striking a balance for Atom between x86 ISA & x86 code base for mobile is Intel's greatest challenge.
Intel's culture of over-designing might be hard to change,however i can see they're getting on the path on being lean. i agree with your take on the instruction sets, but that's just what their architecture based on and ARM architectures are designed on RISC which is historically meant for low power. i would be interested on how ARM is playing in the server market on being more high-performance. it's Jack vs the big giant.
Intel is in third place, but are they poised to take advantage of the market growth and will they grow share? Qualcomm has a big lead. Will they continue to dominate the market? Arm is a big advantage and I don't see Intel moving to it.
Interesting. I really don't see the low end cell phone market growing significantly... Unless low end of tomorrow is high end of today...
On another subject, the skin shaded microphone in the picture above looks like a huge mole....
Perhaps a low end cell phone strategy is an admission that existing software ecologies (iOS, Android) are too much to compete with. Like how the Windows ecology stymied any different PC processors - 68000, Alpha and ARM.
Intel can't even compete at the Cortex A15 level, and they want to compete at the Cortex A7 level? Even if it's possible to do it, do they really want to sell $5 chips?
And that's Intel's biggest problem in the mobile market, especially at the low-end: