Investors following Intel's (INTC -2.40%) mobile efforts should remember that the chipmaker recently struck a deal with a Chinese company known as Rockchip. The two companies are partnering to build a set of low-cost system-on-chip products based on Intel's low-power Atom cores and communications IP. These are intended for low-cost tablets, but Intel CEO Brian Krzanich suggested Rockchip will likely try to enter the smartphone market using Intel's intellectual property down the line.

Last November, Reuters quoted Krzanich as saying he expects that, over time, Rockchip and Spreadtrum Communications (another Chinese company partnering with Intel for low-cost mobile chips) will completely transition to Intel's X86 architecture and away from ARM Holdings (ARMH) intellectual property.

I'm not so sure, though, that this "plan" is going to work.

So about that ARM Cortex-A72 announcement
On Feb. 3, ARM Holdings announced the Cortex-A72, its next-generation high-end, low-power processor core. In the press release, ARM said that Rockchip is among the first companies to license this processor core. Presumably Rockchip intends to build and sell mobile processors based on this core.

What does this say about the Rockchip-Intel partnership, then?

Intel is being used for low-end, cellular-enabled chips only
The Cortex-A72 is aimed at high-performance mobile parts, but the deal between Rockchip and Intel is for low-end chips with integrated cellular modems. So it seems Rockchip is willing to use Intel's CPU core to gain access to cellular baseband technology (which is quite costly to develop), but at the high end Rockchip is sticking with ARM cores and IP.

Now, it could be that Intel simply isn't licensing out its newer Goldmont Atom core (wanting to keep it for its own high-end chips). However, given that Krzanich claimed Rockchip and Spreadtrum would eventually move completely to X86, I'm not 100% sold on that explanation.

Does Rockchip sticking with ARM IP for its high-end chips suggest Intel's Goldmont CPU core, built on Intel's 14-nanometer process, is less competitive than the ARM Cortex -72 built on Taiwan Semiconductor's (TSM -3.45%) 16-nanometer FinFET+ process? The answer to that isn't clear, either.

Maybe Rockchip is testing the waters at the low end before committing to the high end?
Another potential explanation, which I think could make a lot of sense, is that Intel and Rockchip are testing the waters with the initial low-end chips with integrated 3G before figuring out whether they want to continue collaborating. If SoFIA 3G-R (the part Intel and Rockchip are working on together) goes well, then maybe Rockchip will transition more of its chips to Intel.

Here's what doesn't make sense
Krzanich is on record claiming Rockchip and Spreadtrum are likely to go all-in on Intel Architecture, but the U.S. chipmaker's management has also been quite explicit that it really only plans to pursue these Rockchip/Spreadtrum-like partnerships for the low end of the market. Does Intel expect to collaborate with partners at the low end, but then compete with those same companies with internal designs at the high end?

It's not at all clear exactly what Intel expects from the Rockchip partnership in the long run. Rockchip and Intel are collaborating for low-end chips, but they are separately attacking the high end of the tablet market, meaning they are direct competitors. I suppose we'll find out more in due time.