Intel has made some rather counter-intuitive pricing moves for high end graphics capabilities on their upcoming CPUs, and now we know why. If you were wondering how much the Intel branding schemes cost consumers, you are on the right track.
You are probably aware that the iSomethingmeaningless marketing scheme is just that, a way to make consumers pay more for having things not artificially taken away from them. Intel used to justify fusing off capabilities such as cache, cores, and units on a purely technical basis. Long past are those days, but the concept still continues, take a chip and remove features. Unfortunately, it is only an anti-consumer strategy now, buyers get no benefit.
If you don’t think this is true, look at three examples; turbo, hyperthreading (HT) and AES instructions (AES-NI). The i7 has all three, the i5 lacks HT, and the i3 lacks HT, turbo, and AES among other things. Turbo is one that has a possible technical explanation, it takes power. Chips that are marginal on clock speed or power draw/TDP can either be sold at a lower bin, or have that feature turned off. If you turn it off, you can sell the part at the same clock, but do it as an i3 instead of an i5 or i7. Consumers are dumb enough to shop only on simple metrics, so this probably brings in more money. If it didn’t, you would see the lines segmented by clock speed faster than you can say, “Oh no, they caught on”.
Update: 12-13-13@1pm It looks like i3s have HT, that feature isn’t turned off until you get to the middle of the Pentium line.
The other two are just technically indefensible, they are solely there to milk the buyer. Why? HT is an integral part of the core pipeline, it is not a unit that can be turned off or on without preventing the CPU from working. If there is an error in the unit, the whole core is scrap. It has to be turned off artificially, basically by being told to not work programatically, the circuitry must continue to function for the core to at all. Intel uses this fake methodology to make a buyer think they are getting more when they buy a ‘premium’ product, but it is just the same exact silicon with one less blown fuse.
Worse yet is AES-NI, it is there to milk the corporate buyers. AES-NI is an instruction set that speeds up cryptography operations. Things like whole disk encryption, SSH, and other such security features are many times faster with these instructions than without. The chip consumes less power with AES-NI than without because it has to brute force security without it. Like HT, there is no way that the chip can work if there is an error in AES-NI, it would break all of the integer math too. So Intel turns it off in the i3s. Why?
If you are a corporate buyer, an i3 or lower end brand is likely overkill for most of your workforce. Call center drones don’t need 4 threaded cores, they don’t even need two cores, and neither do typists, in fact almost noone really does. What most of these people do need is security, for many SARBOX and HIPPA covered companies, it is legally mandated too. And so Intel fused it off on all but the two most expensive lines. In no uncertain terms, if you want an Intel CPU that doesn’t actively cripple cryptography, you are going to pay a lot more for it.
Going back to the original point of this story, we have the rather curious tale of Intel’s GPU pricing. How curious? So expensive that even in light of the performance gains and power saving you get from Crystalwell, it’s not moving. But Intel has a good reason for holding their ground, or at least good reasons in their mind.
The reason for this is something SemiAccurate has been hearing rumors of for a while now, a new brand for the ULX/Ultrabook/GT3 + Crystalwell parts. Intel is taking the stance that if you want power savings, tolerable performance, slim form factors, and the Ultrabook brand, you are going to pay for it. And pay. And pay. And pay.
Remember that $300 million that Intel threw at OEMs called the Ultrabook Fund? It wasn’t for charity. The idea was to make a premium category for notebooks that only Intel could provide silicon for. They didn’t copyright the Ultrabook name, trademark it, and spend money promoting it for no reason, they expect to use it to extract money from people down the road. If you take money from the fund, the number one handcuff is not how you use it, but that you can’t make things with the resultant technology that has competitor’s silicon in it.
So far, this is standard practice. Between MDF (Market Development Funding), basically a kickback scheme for CPU buyers that OEMs and retailers are addicted to, and jacking prices up, Intel is basically charging a sane price for the CPUs to the OEMs and retailers. Doing so through a back door may have crazy numbers on paper, but it all comes out if the channel plays ball. If they don’t, they pay a net price that the market has difficulty bearing.
For consumers, the idea is to get the buyer accustomed to the higher priced product while using the channel to play along instead of fighting back. When Intel feels they can do so, the MDF funding is removed, the higher prices stay, and viola, profit. Think those “Dell recommends i5/Windows/Office Premium” banners all over sales checkout pages are there by chance? Nope, MDF, and you, the buyer, are paying for it.
That brings us back to the Crystalwell saga again and the price premium Intel is trying to get. This premium is there to pay for the fund and marketing activities around Ultrabooks in the short term by subsidizing the MDF funding for them to reach a magical, and if this retail market is any guide, mythical, price points. When eyes are turned, the funding gets pulled, and Intel makes premium profits.
But this time, the OEMs are balking. PC sales are in the toilet and Windows 8 uptake is slower than even the most pessimistic early projections. On top of this, to the surprise of no one, the last time we looked at Ultrabooks in print sales were projected to struggle to reach 25% of Intel’s 2012 goals. Newer numbers out since then from Taiwan indicate that the final tally is going to be half of that tally, if sales pick up. From there the outlook has no bright points that anyone credible has forecast.
So, Ultrabooks are selling at 1/8th the hoped-for rates, Windows 8 machines in general are down 20% or so during a projected up time, not to mention Christmas and Chinese New Year, and Intel is trying to convince OEMs to pay more than the cost of a discrete GPU and the RAM it needs to add Crystalwell to a Haswell GT3. They need to cover the marketing costs short term, and train buyers that the normal course of silicon evolution is somehow worth an unconscionable premium.
This brilliant scheme is rapidly unravelling though. The sales aren’t there, there are options to go around Crystalwell’s premium with a discrete GPU, and no real reason to tempt an OEM to pony up. Unfortunately for Intel, consumers rightly see discrete GPUs as far more valuable than RAM on package options to an uprated CPU. Anyone wonder why? Luckily, Intel has a new brand in the works to ‘educate’ consumers as to why they should happily pay more for less. But they haven’t told OEMs what it is yet, only given them the price sheets. They haven’t told consumers yet either, so everyone is saying no en masse at the moment. Is anyone surprised?
Where did Intel go wrong? Part of the reason was because Ultrabook sales simply didn’t take off, making the fixed costs, both engineering and marketing, higher on a per unit sales basis. There was too much money going out the door and too little coming in so someone probably panicked. Losing hundreds of millions of dollars is a career limiting move anywhere, more so in marketing than other disciplines.
They didn’t establish whatever Intel ends up calling the Ultrabook brand as a premium before they jacked the prices up. In fact, they didn’t even introduce the new brand at all, they just preemptively charged a premium for it without any explanation. In short Intel got greedy, and quite possibly killed what they see as a golden goose before it ever lived, much less started ovulating. This brand, said to be separate from the Core iX line, is going to need a serious rethink if it isn’t going to launch on tainted products in a dying market. That may or may not happen, but one thing you can be sure of, the price isn’t going down.S|A
Latest posts by Charlie Demerjian (see all)
- Thing go bump(gate) in the night for Nvidia’s GP100 Pascal GPU - May 3, 2016
- Cavium’s Octeon TX blends compute and packet moving - May 2, 2016
- Qualcomm releases Zeroth API to developers - May 2, 2016
- Another detail about Qualcomm server SoCs revealed - Apr 27, 2016
- AMD finally really honestly launches the dual Fury - Apr 26, 2016