The last year or so has shown that AMD is beaten and it knows it. Intel stole a march on the outfit with its Sandy Bridge range and while AMD was supposed to counter-punch with Bulldozer its technology really was not there.
Now AMD is left with a choice, it can go on chasing Intel for a slice of an ever shrinking x86 market or it can do something new. This week it indicated that the Intel versus AMD war is over and Chipzilla is victorious and it is going to try something different
The buzzword in the industry is mobile. If a chip can’t help you to access your email while you are in the bog then the tech press claims your company is doomed.
Intel does not have to worry too much about that. In fact, when it comes to mobile its Atom chip has shown that it is as mobile as a T-Rex fossil found at the bottom of a primordial tar pit. Chipzilla can always be confident that there will always be a large number of people who want x86 chips and even if that market shrinks, Intel will still be quids in.
AMD, on the other hand, has a smaller slice of the market and does not make nearly as much cash as Chipzilla – and its missed beat with Bulldozer has practically killed it.
AMD could do well in mobile. After all, with its graphics arm it could do wonders making low powered, graphic intensive chips. This would be history repeating itself.
The mobile market already has a major player in the form of ARM. AMD would be coming in late and trying to take on an outfit which has a huge customer base. Where have we heard that before?
The only thing that AMD has over ARM is its extended knowledge of graphics, thanks to its ATI purchase. While it has some expertise that ARM lacks, like selling and marketing chips, it would also have to compete with Qualcomm. Qualcomm is putting the spooks up Intel in a way that AMD never did.
AMD would also have to mimic the same trick it did when it started using x86 technology so that it was not peddling its own OS, and that would mean becoming an ARM licensee or trying to do what Intel did and adapting x86 to work with lower spec, lower powered chips.
This would mean that it could still run Windows and might even play nice with Apple gear.
But it is all rather sad and particularly dicey. The question is what went wrong for AMD. Historians who will look at the outfit – which is still Intel’s only rival – will say that it all went pear shaped when it spun off GlobalFoundries.
Former AMD boss Jerry (Colonel) Sanders III once said that “real men have fabs”. AMD’s handing over its production to GlobalFoundries will probably be seen as the point where the outfit took a hatchet to its own testicles.
The advantage of having your own fab is that you can be more adaptable, change product lines quickly and not have to worry about whether or not your small expensive fab gear can be used on other customer’s projects.
GlobalFoundries quickly turned from being a company that AMD could depend on to being a millstone about its neck. Too late, AMD twigged that it needed someone else to make its gear and only recently started talking to TSMC. Such overtures have been as rushed as the one about William Tell and could be just as tricky for AMD in the long run.
Sanders had a few things right about AMD, which unfortunately it is forgetting.
He led it through recessions and out the other side without making any lay-offs. AMD is now fixing its problems by making more people redundant. Sanders did not give up either, which AMD now seems keen to do. He would probably have had someone’s head in a vice until they made Bulldozer work and he would certainly not have created the GloFo mess in the first place.
That was the work of one Hector de Jesus Ruiz who made the right move to buy ATI, even though it turned out that AMD could not afford it, but a wrong one getting rid of the company’s fabs to GloFo.
At the time, analysts claimed that Ruiz took AMD off death watch by “relieving AMD of the cost of running chip plants and allowing more focus on chip design.” Ruiz was behind the deal in which an Abu Dhabi government arm funded the new venture.
But if the deal was such a good idea for AMD, then why did Ruiz quit AMD and go to GloFlo?
It was a question that Dirk Meyer had to wrestle with when he took over. Meyer is seen as the bloke who failed to move AMD into mobile. However, he was completely right not to go down that route. He saw that for AMD to survive it needed its war with Intel and it needed to keep fighting.
In that war, mobile should be a distraction, something to evolve a side strategy to deal with. After all, mobile may not even be the “game changer” that marketers have been pushing. Certainly, betting the farm on the world moving to mobile, in the middle of a recession, would have been suicide for Meyer’s AMD and might prove to be so for his successors.
The way the plan is playing out at the moment, AMD seems doomed to become a poor man’s ARM.