In a Jan. 1 piece published on Fool.com, I discussed the seeming brilliance of Intel's (INTC 2.20%) Core M product line. In a nutshell, by pushing the standard PC processor power envelope even lower (from 15 Watts to between 4.5 and 6 Watts), I believe Intel is effectively "resetting" the performance baseline that customers will see in new notebook PCs. This has the potential to quicken the PC upgrade cycle at the high end of the PC market -- which is great for Intel and its PC partners.
As I have given this further thought, I've come to realize that this "performance reset" theme isn't solely limited to very high-end systems. Indeed, as you'll recall, the vast majority of Intel's budget-oriented Pentium/Celeron processors are based on Intel's lower-performing Atom processor cores rather than its higher-performance Core processors. In fact, in the linked article, I estimate a whopping 22.7% of Intel's PC chip sales in the fourth quarter were Atom-based.
This is a very savvy move on Intel's part. Let me explain.
The "obvious" benefits to this
The most obvious benefit to these processors being based on the low-cost Atom architecture is that these chips are cheaper to build than their Core cousins, according to Intel. This means that Intel can either collect greater margins at the same price points as prior Core-based Pentium/Celeron chips, or it can drive down into lower price points at similar margins as before.
Additionally, the Atom-based Pentium/Celeron chips tend to be lower power than their Core-based counterparts. This means systems based on these chips require less aggressive cooling, which lowers system costs and can improve aesthetics. The trade-off here is that performance is lower.
However, the lower power consumption and performance actually work in Intel's favor, as it does with Core M in that customers will need to upgrade sooner, helping Intel sell more chips.
Why would customers need to upgrade sooner?
Working under the assumption that going forward, most Celeron/Pentium chips will remain Atom-based, it's worth taking a look at the performance step-down seen in moving from a 15 Watt Pentium 3558U (based on the latest Core processor) to a 7.5 Watt Pentium N3540 (based on Atom) in Geekbench 3:
Source: Primate Labs.
Notice that in single-core results, even the significantly cut-down Core-based Pentium delivers 57% more performance than the best low-power Atom core. In multiple core performance, the four weaker cores in the Atom are able to deliver 17.6% more performance than the two stronger cores found in the Core-based Pentium.
Now, given that it is well accepted that most consumer applications tend to make better use of fewer, faster cores than a greater number of weaker cores, I would imagine that the Atom-based N3540 will begin to "feel" slower in everyday tasks sooner than the Core-based 3558U will.
With Intel reporting that 83% of Pentium/Celeron chips sold are based on the Atom, and given that this number is likely much higher for notebook Celeron/Pentium chips than desktop ones, the installed base of Atom-based budget PCs is likely to turn over quicker because of the relatively lower performance for most consumers.
This is good for Intel, and consumers will be happy too
The "brilliance" here isn't that Intel is selling cheaper-to-make, slower processors in lieu of faster, more-expensive-to-make processors. This transition to Atom-based low-end CPUs is made possible through the PC industry's pursuit of (and apparently, consumer preferences for) thinner, lighter, and fan-less systems.
Fan-less systems necessarily have a lower power budget, so performance relative to what can be had with a fan comes down. However, now that low-end notebook and all-in-one PCs essentially feature high-end tablet and smartphone-level performance, but with usage models that are arguably more taxing than those found on smartphones or tablets, the upgrade cycle should be set to accelerate.
