Late last year, Intel (INTC 1.74%) launched families of low power, low cost notebook and desktop processors based on its Bay Trail Atom architecture. These products were intended to allow Intel to compete profitably in lower-cost segments of the PC market, which it had traditionally served with disabled variants of its premium Core processors (sold under the Pentium and Celeron brands).

Over the course of 2014, Intel management had given various data points describing the mix profile of the company's PC chips. During the company's second quarter earnings call, CEO Brian Krzanich revealed that "over 60%" of the company's Pentium/Celeron processors sold were based on this architecture. He also noted that Bay Trail based systems made up "nearly 20%" of the company's notebook chip shipments.

At the company's investor meeting, management gave a further update on just how much of the company's Pentium/Celeron processors were made up of these Bay Trail based products. This is worth exploring in greater detail.

What do the numbers look like today?
According to the following slide, Intel's Bay Trail-based PC processors now make up approximately 83% of its Pentium/Celeron product mix.

Source: Intel.

Now, if you'll notice, Intel gave the mix of Bay Trail as a percentage of Celeron/Pentium, and in other slides, Intel even breaks out the proportion of Core i3/i5/i7 within total Core processor sales. What it didn't give detailed numbers for are the proportions of Core versus Celeron/Pentium. That said, we can use the information we have to try to come up with a reasonable guess.

Doing the math
In order to come up with a reasonable estimate of the total proportion of Bay Trail of Intel's product mix, we need to make some educated guesses on a number of things:

  1. The relative proportion of desktop chip sales to notebook chip sales
  2. The mix of Celeron/Pentium as part of Intel's desktop sales

For (1), let's use research firm IDC's statistics. According to the market research firm, worldwide desktop PC sales will come in at 129.1 million while worldwide notebook sales are expected to come in a 166.8 million. Given Intel's dominant position in the market, I assume that its mix is more or less the same as the markets. This implies 43.6% of Intel's PC chip sales go into desktops and 56.4% go into notebooks.

For (2), note that Intel's CEO made a point to say that Bay Trail made up nearly 20% of the company's notebook mix. This would suggest that the mix of Celeron/Pentium in notebooks is higher than in desktops. This is corroborated by the fact that Intel, over the first nine months of 2014, saw a 9% decline in notebook platform average selling prices but a 2% increase in desktop platform average selling prices.

I assume that approximately 33% (derived by dividing 20% by 0.6) of Intel's notebook mix is Celeron/Pentium and about 20% of Intel's desktop mix is Celeron/Pentium. Further, I assume similar Bay Trail as a proportion of Celeron/Pentium within each category. This implies that about 16.5% of Intel's total PC chip sales were Bay Trail in the second quarter of 2014.

As that proportion moves from 60% in the second quarter of 2014 to a projected 83% in the fourth quarter of 2014, the proportion of Bay Trail of Intel's total PC processor unit shipments moves up to approximately 22.7%.

What does this mean for investors?
The bottom line here is that Atom-based processors are now a very significant portion of Intel's very profitable PC Client Group. As a result of this, Intel should have every reason in the world to continue to develop competitive low-power cores and system-on-chip products.

This will, of course, help Intel stay hyper-competitive in the PC market. However, perhaps more interestingly, the R&D effort for Intel's low-power cores and other IP is now justified given that they contribute a substantial amount to Intel's PC chip sales and profitability. A virtuous circle, indeed.