The fine folks over at Anandtech have recently pointed out that the Microsoft (NASDAQ:MSFT) Surface Pro has recently gotten a $100 haircut. Anandtech believes that this price reduction is Microsoft's attempt to clear old inventory to make way for the next generation Surface Pro built with Intel's (NASDAQ:INTC) 4th Generation Core processors (codenamed Haswell). While it is clear that the first Surface Pro devices didn't exactly sell like hotcakes (although thankfully Microsoft didn't have to take a huge write-down as it did with Surface RT), the real question is whether next generation models will do any better going forward.
Is Haswell A Silver Bullet?
A few of the major criticisms of the original Surface Pro was that it was too bulky and had rather poor battery life. These shortcomings were largely the result of the use of Ultrabook-oriented Core i5-3317U. The problem with this chip isn't necessarily the fact that its thermal design point was 17W (and this is actually 20W if you include ~3W from the PCH), but from the fact that the idle power characteristics of the Ivy Bridge processors were not particularly well suited for a tablet. The trick to achieving dramatically better battery life is to have the processor consume next to no power when it is doing next to nothing.
The Ivy Bridge U + PCH platform consumes roughly ~3W at idle, while the Haswell-ULX (4.5W/6W SDP parts for tablets) processors consume from 0.04W - 0.85W in their various idle states according to Intel's processor datasheet:
So, right off the bat, if Microsoft uses one of these 4.5W/6W "SDP" parts, users will likely see a dramatic increase in battery life (given that most tablet usage models involve significant amounts of time in idle and light-workload states). But you don't have to rely on just theoretical data and Intel-supplied measurements. Take a look at the battery life improvements that Apple (NASDAQ:AAPL) saw in moving from Ivy Bridge to Haswell-ULT (15W TDP part with the highest end graphics) in its MacBook Air, in Anandtech's review of Apple's latest ultra-portable notebook:
While the new MacBook Air has a slightly larger battery than its 2012 cousin (8% larger capacity), it is clear that the majority of the gains are thanks to the significantly improved low power capabilities of the Haswell part.
So, it is my view that the next iteration of the Surface Pro (which will very likely use Haswell in its 4.5W or 6W SDP incarnation) should offer similar performance to the previous generation model but feature significantly better battery life, quite probably in a much sleeker form factor. I wouldn't be surprised if the next gen Surface Pro with the new 4.5W SDP Haswell parts is thinner and fanless. In short, the device itself is likely to be significantly more appealing from a portability and aesthetics standpoint. But will this be enough?
What About The Price?
Microsoft's Surface Pro received rather harsh criticism for being expensive; after-all, $999 for a tablet isn't cheap! It is no secret that the explosive growth in computing is coming in at the lower price points, while the higher end portions of compute - from smartphones to notebooks - are seeing slowing growth at best and negative growth at worst. However, as Apple has repeatedly shown, if a company develops a truly compelling high end device, then there will likely be a market for it. It's not farfetched to believe that in a few years, the traditional clamshell PC will give way to convertible devices across the board, and that Surface Pro-like devices will be the norm (albeit not all of them will be so high end). So, at worst, a very well executed Surface Pro (and the numerous devices like it) will help to re-energize the high end market (which has been largely flat in the PC/notebook space).
However, I view the Surface Pro as much more a halo device to help to drive sales of the upcoming swarm of cheaper, Baytrail (new Atom based on Silvermont) based Windows 8.1 tablets and convertibles. These will be devices in the $499 - $599 range, and probably be even thinner and lighter (if less powerful) than the Surface Pro.
Maybe It's Time To Kill Surface RT, Eh?
In light of this, I think it's time for Windows RT to just die. With ASUS, a major partner on previous Windows RT devices, announcing that it's throwing in the towel on this platform, and with Microsoft's only obvious competitive advantage (at the moment) in its full X86 software compatibility, does a next generation Surface RT make sense?
No, not even a little. If Microsoft wants to really sell these things, then Surface Pro with Haswell should be the halo product, while Baytrail-T powered Surface 10.6" and Surface 7-8" products should drive the volume and the revenue. These will be full X86, Windows 8.1 devices that support every legacy peripheral under the sun, and every legacy program that the user wants to run in addition to all of the Modern UI stuff. All of this at comparable prices to the ARM (NASDAQ:ARMH) SoCs.
Until Windows RT and Surface RT die, users will still be confused, and it could hinder the potential sales of Windows 8.1 based tablet devices - both Microsoft-branded as well as from third parties.
I look forward to the next generation Surface Pro. However, more importantly, I look forward to competitive Windows 8.1 based tablets at reasonable prices based on either the new Intel tablet parts or the upcoming AMD (NASDAQ:AMD) Temash tablet part. This should only help Microsoft gain market- and mind-share in the tablet space, particularly as businesses and their IT departments are likely to be much more comfortable integrating Windows 8.1/X86 tablets in their networks than iOS or Android devices. Microsoft has everything in place to make Windows 8.1 and its successors shine in this new era of computing, and I'm hoping to see the company really tap its potential this holiday season.
Disclosure: I am long INTC, AMD. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.