One of my readers left a very interesting comment on one of my recent pieces, "Intel Does Something Completely Unexpected" (thanks, steinlaus!) that I believe may be running through the minds of many of my readers,
Bay Trail is fast enough for almost all kind of office work and for playing HD videos. What is left? Only a very small market for users, needing high power workstations, assumingly less than 1 % of the PC market. I am afraid that Bay Trail is a real game changer, marking the end of high price PC processors. That is why I sold my Intel shares and call options
While this may initially seem like a pretty major and game-changing concern, this isn't actually as big of a concern as one may intuitively think. Indeed, there is ample evidence to suggest that Bay Trail/future Atom processors will serve mostly to bolster Intel's (INTC) share in the low end (which it has been losing to AMD (AMD) for quite some time) rather than fundamentally obviate the need for higher performance PC processors.
Intel's High End Chips: Record Shipments
At the recent Credit Suisse Technology Conference, Intel's Kirk Skaugen (head of the PC client group) noted that even during the relatively weak 2013 for PCs in general, Intel saw record volume shipments of its Core i5 and Core i7 processors. Note that Mr. Skaugen did not say record mix (although this would be a natural fallout of record volumes), but record volumes. This is a much higher margin business for Intel (since Core i3/Pentium/Celeron are all based on disabled Core i5/i7 dies just sold for cheaper) and it is an encouraging datapoint indeed.
Why is that? Well, it's encouraging first and foremost as the sale of higher margin/higher ASP products hits the bottom line in a nicer way than the sale of the cheaper chips, but it also tends to indicate that the portion of the market that is most vulnerable to tablets/smartphones aren't the higher performance PCs but instead the lower end super cheap systems. This has a number of pretty compelling implications.
Intel's Low End: An Opportunity To Gain Share
According to Mr. Skaugen at the aforementioned conference, Intel's share in the sub-$399 PC space is under 50%. Part of this is that the company hasn't focused on this space all that much, but the main reason is that Intel really hasn't had a product that could actually compete profitably in this market. While AMD has been building high quality, low end/low power platforms aimed at low cost, Intel had been grudgingly shipping disabled Core processors into this market.
It doesn't take a financial wizard or a financial guru to understand that selling about 150mm^2 worth of silicon (between the CPU/SoC complex and the chipset) in a market where your competitor - content to live with lower margins - is shipping 102mm^2 worth of lower power silicon is a problem. Until Intel beefed up its low power/low cost cores to the point where they could successfully compete with AMD in this market, AMD would likely have continued to gain share with its Kabini/Temash solutions.
However, with Bay Trail and an aggressive roadmap going forward, Intel can at the very least defend its share and in a more optimistic scenario take meaningful amounts of share from AMD (I suspect reality will be somewhere in-between). At any rate, even if Bay Trail and its follow-ons are "good enough", it's better that Intel gain share in the low end than to leave it to AMD and the ARMH (ARMH) players to have it all to themselves.
Another Major Point: Businesses Ain't Gonna Buy Cheap
Roughly 50% of Intel's PC chip business is actually sale of chips into business/enterprise environments, and the other half is consumer. Interestingly enough, while consumers are probably more likely to buy "cheap" PCs, businesses don't really mess around with lower end products. Why? Well, first off, Intel tries to push businesses to adopt Core i5/i7 with a suite of security and management features known as "vPro." Since Intel doesn't offer this technology in Core i3 and below, this is one major incentive for businesses to keep buying high end.
But more importantly, businesses tend to buy high end PCs simply because they want to get as much life out of the asset as possible. Why risk buying a machine that is likely to be too slow in a few years when it just makes more sense to pack said machine with the latest Core i5/i7, a ton of RAM and fast storage? This means the machine lasts longer and, for many use cases where time equals money, offers a lower total cost of ownership. So, while consumers may be more likely to buy cheaper products (and consumers typically buy to their budgets anyway), enterprises/businesses are not as likely to do so.
Bay Trail And Successors Offer Margin Improvement
Quite possibly the biggest "win" from the development of the low power Atom processors is the fact that the Celeron/Pentium processors at the low end can now have a cost-reduced (and lower power) product rather than the unsuitable, feature-crippled higher end stuff that's bad for Intel's margins. According to the presentation from Intel CFO Stacy Smith, the platform cost reduction from going from "Ivy Bridge Celeron" to "Bay Trail Celeron" comes in at a pretty substantial 30%:
Further, as Intel goes out to 2015 with its "Broxton" platform (where's Cherry Trail?), Intel sees another 37% platform cost reduction. This is pretty unreal given that Broxton is likely to offer substantial performance and feature improvement over Bay Trail (30-50% improvement in CPU performance and over 3x the graphics performance, by my estimation).
So, even if there is a "mix shift down," the cost structure of these products is much better. Further, since there have always been low cost solutions in the market, and given that Intel has less than half of the share there, I do view this as a pretty compelling share-gain opportunity rather than anything that is likely to destroy Intel's high end business.
The great thing about Intel's mobile efforts, even despite my criticisms, is that they have driven Intel to focus on lowering the cost of its lower end platforms. This will allow Intel to gain share in the low end of the PC market profitably while at the same time keeping gross margin percent (and perhaps raw gross margin dollars) per unit at "good" levels. I can't wait to see what Bay Trail powered systems hit the market this year and, more importantly, I can't wait to see what kind of performance level "Broxton" brings to the PC market. The only hope is that Intel can advance its "Core" processors fast enough so that the performance gap between the best Atom and the worst Core processor stays at roughly 2x.