AMD Responds to the Sandy Bridge Bug: Marketing ramps up... but new processors do not
Timing is Everything
The timing of the announcement is also of great import. Sandy Bridge based parts had only been released some three weeks before the recall. Estimates are that some 8 million chipsets have been shipped to retailers, OEMs, and consumers alike. This would obviously have been a much more significant problem if that bug had not been found until some six months after launch. By then tens of millions of parts would have been shipped, and the impact on Intel’s bottom line would have been disastrous (for a quarter or two). As such, a potential $1 billion set back is manageable for a company of Intel’s size and financials.
Finally the issue is with the supporting chipset, and not the actual Sandy Bridge series of processors. This does not require Intel to stop all production of these relatively expensive CPUs (as compared to relatively inexpensive chipsets) and do a major redesign. The issue with the chipset was determined in around three day’s time, and the fix could be implemented in the metal layers. This means that the change can be applied to wafers that are in production, and fixed silicon is only a few weeks out rather than a few months.
This was very much tongue in cheek, and I loved it. I think Ryan only "liked" it.
The potential financial windfall for AMD, even if they had reacted faster than what they have shown, would still have been relatively small. AMD does not have Bulldozer or Llano parts ready to ship, and no amount of bribes and whips would have made production and introduction of these parts possible in the timeframe that Intel has opened up with the Sandy Bridge issue. There were no plans to move up introduction of these new parts.
The biggest potential win for AMD would have been a successful AM3+ motherboard launch, but they still were not able to move up that particular product. If AMD had been able to come out swinging with this new socket, and new motherboards designed around it, we could very well have seen a much more significant boost to AMD. While there are no Bulldozer nor Llano based chips available, AM3+ boards do support current AM3 processors. Releasing the high end 990FX and budget 970 boards, neither of which utilize the integrated graphics of Llano, and populating them with current Phenom II and Athlon II CPUs could have provided a nice impetus to AMD and given the company a marketplace more prepared for the new CPUs to arrive later this Spring. Admittedly, moving up such a major release is not exactly easy, but at this past CES we saw a number of motherboard manufacturers with nearly final AM3+ designs. Considering that current AM3 chipsets can be utilized with AM3+ CPUs, this should have been something in the pipeline a long time ago for AMD and its partners. From what I gather, the 890FX will be rebranded the 990FX. And just as Intel has done, it is quite feasible that AMD will feature AM3+ motherboards without video outputs, even though AM3+ does support that functionality.
AMD has shown itself that it is not nearly as agile as they perhaps once had thought. It took a couple weeks to put together a marketing campaign to address the issues that Intel is having, and to help promote current AMD products as an alternative when buying a new computer. They were unable to pull forward any launches (something that Intel has had great success with in the past), and the response from marketing when initially asked about this issue was disappointing for those of us hoping to get a little controversy brewing.
As of the date of this article, Sandy Bridge based chips are available and for sale. Getting a decent motherboard is another issue altogether.
Intel on the other hand has been burning the midnight oil to reduce the impact of this issue. When first announced, it was expected that consumers would see fixed boards in a mid-April timeframe. Now we are expecting to see product available in quantity by the end of March. In fact, some notebooks and other systems will be released in the first half of March. Motherboard partners have already started to receive fixed B3 silicon from Intel, and they will be rolling boards out using these chips in a matter of weeks.
Considering that most OEM based product families take months to prepare for launch, it is pretty unlikely that AMD has gained any traction with big names such as Dell and HP. There may have been a small boost with current AMD products that these companies already provide, but certainly these folks are not going to create a new lineup in short order to fill the hole Sandy Bridge left. Consider again that products based on Sandy Bridge will again be produced and sold in a matter of four weeks or so. We also are sitting in Q1, which is the slowest quarter for PC sales. So the potential impact on sales that the Sandy Bridge bug provides is minimal. We might be singing a different tune if this had occurred towards the end of Q3, when PC makers are gearing up for the holiday season.
If anything, AMD should take this experience and learn from it. They looked to be rather flat footed in their response in terms of rearranging product releases, as well as a perceived lack of any kind of marketing thrust. Intel does not make many mistakes, and if AMD hopes to gain any kind of marketshare, they need to be ready to pounce on any stumble that Intel experiences.