The Fall of AMD

Note: This is an old post from 2007

AMD’s latest flagship Phenom processor is an utter disappointment and marks the beginning of the end to the once successful company that dominated the enthusiast CPU market. AMD shares are at a 4 year low of $9.67 after another quarter with losses in the hundreds of millions[400 to be specific], with debts nearing 9 BILLION dollars. Its fall was just as surprising and quick as its rise. After hitting a high $40 in Q1 of 2006, it lost half of its value in just a year, only to decrease by 50% again this year now sitting below $10.

There are many personal feelings associated with AMD since it was one of my first and most successful investments in the stock market; I purchased the stock in 2004 and sold it the next year for a nice profit. AMD’s processors were also the CPU of my choice and the choice countless other computer enthusiasts since 2003. I purchased almost a dozen AMD processors, using a FX-62 to construct my own computer and the many others when I assembled computers for friends, family and business. Its processors at the time provided superior performance to Intel processors while having a lower price; a no-brainer.

In 2003, AMD’s Opteron and Athlon 64 succeeded where Intel failed, by starting the revolution for 64-bit computing. The Athlon 64 brought 64-bit processor to the masses with non premium over Intel’s comparative Pentium lineup while providing near perfect 32-bit backwards compatibility, the primary reason Intel’s Itanium line up[first shipped in 2001] failed. This innovation by AMD would be the slingshot that would take down the Goliath.
[Note: 64-bit Processor Timeline from Wikipedia]

AMD’s stock price would double consecutively for the next two years. It would also destroyed Intel’s monopoly of the computer processor market during this time period when major OEM’s such as Dell[Q1 2005] and HP[Q1 2004] prominently offered desktops as well as servers powered by AMD’s Athlon 64 and Opteron processor respectively. The CPU market was starting to stabilize as a duopoly with processor pricing hitting all time lows while performance increases hitting all time highs.

So what went wrong? AMD got complacent and stopped innovating. This is evident in the release of its unpolished and disappointing Phenom quad-core processor a year behind Intel’s first quad-core processor, the QX6700. Intel also beat AMD in the desktop dual-core race by being the first to release its Pentium D Extreme in Q1 of 2005.

But the stake in the coffin for AMD was Intel’s Core architecture released in Q1 2006, the new architecture vastly improved the efficiency and performance of the CPU through various design optimizations, and Intel further improved on Core when it released its Core 2 architecture in Q2 2006, that brought a 40% increase in performance while reducing power consumption by 40% compared to its old Pentium architecture; blowing away AMD’s processor offerings on the market.

AMD’s Phenom processor was suppose to do for AMD what Intel’s Core processors did for Intel, utilizing the multiple innovations including native quad-core to improve performance. As previously mentioned, the Phenom debut is a debacle to put it lightly. The clock speed was a few hundred mhz lower than what promised due to a bug in the chip that was not discovered until late in the production maxing out at only 2.3ghz.

The top of the line Phenom 9600 is easily bested by Intel’s low end quad-core Q6600. Unless AMD can pull a miracle in Q1 with its higher clocked Phenoms which barely measure up to Intel’s slowest quad-core offering, it will effectively loose any hold in the desktop CPU market. Already deep in debt due to high costs and low margins, one wonders how much longer AMD will be able to stay in the market, since Intel’s recent move from 65nm to 45nm manufacturing process allows it to decrease costs while further increasing performance.

I would hate to see Intel regain the monopoly that it once had; the rate of innovation will decrease and relative prices will increase. There is already signs of this since there has been rumors that Intel is holding back the release of its higher performing 45nm Penryn to avoid lowering the prices on its current processor lineup since there is no pressure from AMD to do so.

AMD’s final hope may be in its acquisition of ATI in Q2 of 2006. If AMD is able to integrate the GPU in to its CPU, then it may bring significant power savings and possible performance improvements. AMD may also get helping hand from the courts if there is any validity in AMD’s antitrust complaint. Either way, AMD is going to need some miracle to return its former glory and it is going to have to be soon or never.

Disclosure: I am long INTC

 
29
Kudos
 
29
Kudos

Now read this

Computer Science and Math

I came across an interesting discussion regarding the role of math in computer science education on SIGCSE’s (ACM Special Interest Group on Computer Science Education consisting of CS professors) mailing list. Brad Zanden, a professor in... Continue →