There's a lot of gold out on the internet for those who are willing to look. When I cover a company, I cover the company. It's not enough to look at the raw numbers, nor is it enough to follow the news. It's important to watch the competitive landscape, and to watch what the companies themselves are doing. This is what professional buy-side analysts do, and this is why they seem to "know" how things are going well before the public does -- it's not insider information, it's simply a better understanding of the market and competitive dynamics by virtue of "going out there" and learning what they can.
I'm not a professional buy-side analyst, but I make do with what I have. One of my most invaluable tools is actually watching the job listings at the companies I follow to see what they're actually up to. In doing my weekly scan of the Intel (INTC) job postings for something noteworthy, I saw a very interesting open position:
The position is for an ARM (ARMH) Software Analysis Engineer. As you can see from the description, the purpose of the job is to figure out Intel's competitive position against the ARM-based players and then to educate the sales guys on how Intel's stuff stacks up (read: what Intel's stuff is better at). These sales folks then make sure that the customer is made keenly aware of why they should buy Intel's stuff over, say, Applied Micro's (AMCC).
The importance of this job listing isn't that they're doing it (it is expected), but it underscores that Intel is actually taking competition here very seriously. Very vocal ARM bulls/Intel bears generally pounded the table on the fact that Intel would not take competition seriously in the low power server space for fear of cannibalizing higher margin Xeon sales. This is silly for one very simple reason: if Intel doesn't do it, somebody else will.
With Qualcomm (QCOM), AMD (AMD), Applied Micro, Broadcom (BRCM), and others gunning for this space, Intel is in the unenviable position of having to make sure that none of these players makes significant inroads. There isn't enough room for more than perhaps 2-3 players (the nature of the chip business has generally favored few suppliers), and I believe that Intel's absolute economic advantage in the CPU space (thanks to its fabs, no license fees to be paid, designs leverage existing IP/knowledge, experience with server designs and clients, customer trust/confidence in this space) will lead to Intel more-or-less winning the space, as it did in the mainstream server space (and what it is doing in the big iron space with Itanium/Xeon E7). But to do so, it will need to push full-force to make sure no other competitor takes any sort of non-trivial market share.
In short, if you think Intel isn't taking the ARM threat in servers seriously and isn't moving to intercept well before the first player ships silicon, then you're wrong. Dead wrong.
Additional disclosure: I am short AMCC and ARMH via puts.