An excellent article on The Street argues that while Intel are still dominant players in both the traditional PC/Notebook space and the netbook space, they are being challenged in both the notebook and netbook space. According to Robert Castellano at The Street:
....the Atom is propping up Intel's unit shipments in the mobile PC sector. It's (Intel) making little or no money on the Atom anyway. A more important issue is that it (Intel) may be losing market share in the notebook market. Why? Because Intel had to fill orders for netbooks in the fourth quarter of 2008 and the first quarter of 2009 and made Atoms instead of Penryns, resulting in lower margins on a $29 CPU. Once PC OEMs migrated to the AMD CPU, they stayed with it.
I've also forecast previously that while Intel's Atom will hold more than an 80% share of the 23.5-million netbooks sold in 2009, a movement is underway that will enable the processor from ARM Holdings (ARMH Quote) to gain a 55% market share in 2012.
What I find astonishing (not unbelievable mind you) is Castenallo's prediction that ARM processors will have a 55% market share of the netbook space in 2012. This would be a drastic inversion of the market share for both companies. I am assuming Castenello is making this assumption based on his knowledge of both the Intel and ARM roadmap (as far as he can see it anyway) and so this inversion must be troubling for Intel. Certainly ARM has been pressing them in the mobile phone space for many years and this does indeed seem to be the natural progression for them. It would be interesting to see if ARM in the future plans to move further along the product train and try (at some point) to compete with Intel in the notebook space.
In related news Electronics Weekly reports that ARM-based netbook chip orders are set to max out the 65 nm and 55 nm processing plants at both TSMC and UMC. According to Electronics Weekly:
According to the Commercial Times of Taiwan, both TSMC and UMC will be at 100% capacity utilisation for 65nm and 55nm processes by November, because of a flood of orders placed on them for ARM-based Netbook chips.
The orders are coming from Qualcomm, Texas Instruments, Freescale Semiconductor Via Technologies and Nvidia.
Freescale says it has three Netbook design-ins expected to go into production before the end of this year, while Qualcomm says it has half a dozen design-ins. If TI, Via and Nvidia have three or four each, then there could be 20 ARM-based Netbooks on sale before Christmas.
Both Paul Jacobs, CEO of Qualcomm and Rich Beyer, CEO of Freescale, point out that the new metric for measuring computing performance is going to be power efficiency rather than CPU speed. This massively favours ARM which has always designed for power efficiency as against Intel which has always designed for speed.
The Electronics Weekly article sheds further light on the growth of ARM chips on netbooks. Of course there is less visibility into the production of Atom chips as (although some of it apparently is being done at TSMC) most of it would still be in house anyway. ARM chips do seem to be making a big bang on the netbook stage and one wonders just how Intel are going to compete in the long term. Remember Intel are used to competing as the dominant market leader and not as one among equals so it will be interesting to see if they have the ability to alter their strategic approach to remain dominant and also to see if they adjust their business model to compete against ARM. One thing is for sure, many people around the world would prefer a weaker Intel.
Electronics Weekly: Netbook Chip-Set Orders To Max Out Capacity At TSMC & UMC
The Street: Intel's Grip on Notebooks, Netbooks Slips
2 comments:
Here are some longer, rambling thoughts:
Very interesting prediction about ARM. I think he is looking at things on a purely spec basis, but I agree, while, as you say astonishing, not unbelievable.
When Mghz was all that mattered, Intel could always develop faster than everyone else and stay in front. When laptops and power efficiency became popular, Intel stumbled at first, but then came back with a roar to get back firmly in front again. They certainly have the resources to out-develop and out-manufacture anyone else on x86.
But now, when we get to mobile device levels of efficiency, it looks like x86's architecture has some inherent limits that make it very difficult for it to compete with ARM. And Intel does not seem to have any inclination to invest in another other architecture either.
Interesting to me is one year ago, there were still netbooks with Linux installed widely available. But since Atom has come out, everything has switched to pretty much XP.
So there are two forces going in opposite directions here:
1) Power efficiency (esp with 3G, 3.5G), is a serious problem. ARM can go a long way in helping solve that.
2) When consumers have a choice, they still prefer the familiarity of Windows XP over Linux. Windows 7 may yet be even faster than XP. But Windows isn't available for ARM. Ubuntu on ARM? I'm not sure battery life by itself (which is steadily improving with 6-8 hours already achievable) will be a convincing enough argument to your average consumer. I admit that with the trend of applications all going to the web, I didn't expect this reversion to WinXP, but the trend is now a historical fact.
Something that really caught my eye also: I noticed netbooks being offered with up to 10GB of free online storage. That is a small, but fundamentally revolutionary step.
When Acer or Asus begin offering online storage, they are fundamentally altering their relationship with the customer--they are providing a service, not a product. With online syncing, the consumer needs to create an account with you, and there become opportunities to provide all sorts of additional services. Done wrong, you just simply spam the customer with email "special offers". Done right, and you have a platform like the Apple iTunes App and Music Store.
The small baby step is flawed--the online storage doesn't solve a ubiquitous problem of netbooks, which is syncing them with your other computer(s). And I question if the Taiwanese PC makers have the ambition and courage to take the additional steps and get into the app store business. That, along with ARM, appears to be a remarkable opportunity that isn't easily replicated. But it requires working against the habituation of consumers to Windows, which is certainly no small task.
Wow, even Nokia thinks enough of the netbook trend to start making netbooks: http://www.nytimes.com/2009/08/25/technology/companies/25nokia.html
Post a Comment