Windows RT, Microsoft's (MSFT +3.06%) attempt to bring Windows tablets and PCs to the ARM Holdings (ARMH +0.00%) architecture was an abject failure. With limited developer support for its walled-garden app ecosystem, binary incompatibilities with the millions of traditional Windows applications, and a very aggressive push from Intel (INTC +7.14%) -- and its faithful PC partners -- to flood the market with performance, power, and price competitive full Windows 8.1 tablets, Windows RT was dead on arrival.
Recent comments from Microsoft suggest that the company is planning to put an end to these multiple flavors of Windows -- Windows Phone, Windows RT, and Windows 8.1 -- to bring one unified platform. In light of this, should Microsoft will simply cease work on ARM-based versions of Windows and go all-in on Intel-based chips?
In tablets, this is the only logical move
When Intel lacked competitive low-power mobile system-on-a-chip, or SoC, products, it made sense for Microsoft to want to hedge its bets. After all, if Intel couldn't or wouldn't deliver, then Microsoft's tablet ambitions would crumble. However, when Intel launched its first low-power SoC for tablets -- known as Clover Trail -- it was actually a pretty decent chip. Graphics performance was terrible, but the general-purpose performance was quite a bit better than the Tegra 3 in the Surface RT.
Intel's next-generation product, Bay Trail, was more competitive, offering leadership CPU performance and fairly decent graphics performance. The graphics performance of the Tegra 4 found inside of the Surface 2 was still better than Bay Trail's, but the delta wasn't so large as to make the Intel chip look laughable.
Going forward, Intel seems very serious about mobile graphics performance, with its upcoming Cherry Trail product looking to offer roughly four times the graphics performance of the Bay Trail chip -- making it very competitive across the board. Add in the fact that only Intel's chips offer competitive performance nd power and run full Windows applications, and it seems that Intel-based Windows tablets drive a value proposition that ARM-based Windows systems will struggle to beat.
In smartphones, the situation is trickier
While Intel inside just about every Windows tablet makes sense, it's not as clear that Intel would be a good fit for Windows Phone. Intel's smartphone roadmap hasn't generally been competitive, and its products have been plagued with delays and uncompetitive specifications. Qualcomm (QCOM +1.26%) -- which develops ARM-based processors -- is Microsoft's partner of choice on Windows Phone.
Ditching Qualcomm (and, by extension, ARM) would be a very risky move, particularly given how proven a supplier Qualcomm is in smartphones. Further, only Qualcomm's modems at this time support the CDMA standard, which means that if Microsoft were to go all-Intel, Windows Phones would not work on Verizon and Sprint -- two major U.S. carriers.
Moreover, Qualcomm has worked pretty closely with Microsoft to develop Qualcomm-based reference platforms, system software, and so on. Would Microsoft really tell Qualcomm to take a hike just because its chips are incompatible with full Windows desktop programs that very few would ever want to run on a phone?
Foolish bottom line
In Windows tablets and PCs, the ARM architecture is probably finished as Intel moves aggressively and leverages its key X86 compatibility advantage. However, for Windows-based smartphones, Intel is just not yet a credible supplier either on the applications processor or modem side of things. Over time, Intel could build that credibility as a smartphone chip provider, but even then, it is very unlikely that Microsoft will boot Qualcomm -- and therefore ARM -- out of Windows entirely anytime soon.
