- AMD shares have tucked on impressive gains on news about a possible split of its server business from its core graphics card business
- While this at first appears to be counterintuitive since it would remove the synergies that its graphic cards and server business share, the pros outweigh the cons
- AMD can still benefit from its much-anticipated Zen chips even after a split
Shares of semiconductor chip manufacturer Advanced Micro Devices (NASDAQ:AMD) have gained over the last few days following a series of positive developments surrounding the company. The first was AMD unveiling its much-anticipated Fury GPUs that will help it compete on a more even keel with GPU market leader Nvidia (NASDAQ:NVDA). AMD unveiled three graphics cards a few days after Nvidia’s launch. However, R9 Fury X will retail at $649, same as Nvidia’s new GTX 980 Ti, is what garnered the most accolades. The card features 4096 stream processors, a big improvement over the previous-generation RX 290X’s 2816 stream processors. This has led to the card being tapped to become a market favorite.
Also See: AMD Stock analysis video
But what really caught the fancy of the investing world were rumors that AMD is looking to split/breakoff parts of its business. According to the speculations, AMD is planning to spinoff its server business from its graphics cards business. Though AMD is yet to officially confirm the rumors, this would not be the first time that it would be exploring a spinoff. The company has considered such a move in the past but decided against it. After all, such a move would void the synergies that comes from integrating the two businesses. So what incentives does AMD now have to breakoff parts of its businesses? How will this move help AMD to compete with its chief server chip rival Intel (NASDAQ:INTC), and perhaps wrestle some of the server chip market share it has steadily lost to the chip giant over the years? AMD introduced 64-bit x86 server chips in 2003 that became very popular and helped the company to increase its server chip market to an all-time high of 26% by 2006. But Intel soon responded by introducing 64-bit capabilities into its Xeon processors and managed to wrestle back what AMD had managed to grab. Intel’s current server chip market share is reputed to be an overwhelming 98% while AMD’s has fallen to a sorry 1.5%.
Server business has failed expectations
There could be any of a number of reasons why AMD would like to split its x86 server business from its graphic card business. Graphics cards and APUs, or Accelerated Processing Units, have long been at the core of AMDs business. Its ARM server business has, however, failed to live up to expectations. AMD and a host of other server chip businesses were banking on microservers really taking off, but this has unfortunately failed to materialize. AMD had predictions that its ARM chips would account for 25% of the server market by 2017. Current ARM market share, however, is around 2%. AMD recently announced that it is ditching SeaMicro, a microserver business that it purchased for 334 million.
Even if the company’s Zen architecture that it plans to launch in 2015 performs beyond all expectations, it would be a stretch to expect it to grab 25% market share in the space of a single year. Zen is a new architecture that will be used in AMD’s Opteron chips that the company has touted as offering a huge 40% performance improvement over current Opteron chips that use the older Bulldozer architecture. Unlike older Opteron chips that AMD positioned to compete with Intel’s low-end Xeon E3 chips, the new Opteron chips will be aimed at Intel’s mid-range Xeon E5 chips.
AMD is taking a different approach when competing with much bigger and resource-rich Intel. AMD has seen mounting losses, and simply lacks the R&D wherewithal to compete directly with Intel. Instead of trying to compete head-on with Intel in similar market segments, AMD is now focusing more on trying to discover niches where Intel is either non-existent or not-so-dominant. This appears to be a rehash of its glory days when it managed to pip Intel by being the first to the market with 64-bit x86 chips. But, while hoping for the best, AMD is also being cautious, and rightly so. Despite the promise for superior performance that Zen brings to the table, it does not automatically translate to market success. Zen will simply position AMD for success but cannot do much to guarantee it. It therefore pays to have a contingency plan that the company can fall back upon in case Zen fails to take off as expected or takes much longer to gain traction. The option here is for AMD to split the server business but retain the x86 licenses. This way the company can spin off the server business but still retain control over it. AMD can therefore still benefit big-time if Zen really takes off even after a split.
Server chip R&D tends to be quite heavy, and can easily drain off a company’s resources if it goes through a patch of several launches that fail to take off. By splitting its server business, AMD can concentrate its resources on its core graphics cards which have been more promising. AMD’s GPU cards such as Radeon are reputed as offering superior performance to some of Nvidia’s leading cards, which has helped AMD to prevent Nvidia from crowding it out of the market a la Intel.
AMD can still benefit from its upcoming Zen chips even after it spins off its server business by retaining the licensing rights of x86. At the same time, the company can direct the resources it would otherwise have to continue channeling into the server business into its graphic card business which has been doing quite well.
Note: You might also be interested in our Nvidia Stock Analysis video