5 Reasons Why AMD Processors Are Doomed
Back in late 1999, I built my first computer. It used an AMD Athlon processor clocked at 500 MHz which was quick enough to play most games at the time, and also a better value than what Intel offered. I also liked supporting the underdog.
During the early years of Athlon, it seemed as if AMD has a legitimate chance against Intel. Yet, as the years moved forward, the momentum disappeared. Now it appears to be time to start foretelling the fall of AMD – and here’s why.
The Processors Are Slow
AMD’s downfall began with Phenom. Released in 2007, this microarchitecture was supposed to bring AMD back into the race with Intel, who was starting to win with its Core 2 products. Yet Phenom ended up being a bit of a disappointment, as it was unable to defeat Intel’s best. The architecture was salvaged somewhat by low pricing, but the performance difference only increased in Intel’s favor over time.
Bulldozer, the new architecture, was hoped to be more competitive. Instead, it proved not only be slower than Intel’s Core processors in most benchmarks, but sometimes slower than previous Phenom II processors as well. In other words, it is a failure.
That’s a massive blow. A new architecture takes years to develop. The company was banking on Bulldozer, and now that it has proved uncompetitive, they have no way to quickly follow up.
The Processors Are Power-Hungry
While AMD processors lack performance, there is one thing they excel at – using power. Phenom was less power efficient than Intel’s processors from day one, and the situation has only become worse over time.
The new Bulldozer-based FX-8150, for example, uses slightly less power at idle (about 10-15 watts in most reviews. However, at load, the processor uses about 65 watts more than an Intel Core i7-2600K, which is over 40% more.
Because of this, AMD processor-based systems actually lose much of their value equation over time, as they cost slightly more to run. But the bigger problem for AMD is that this high power draw requires larger power supplies and larger heat sinks. Among OEM manufacturers like Dell and HP, this means an increase in total system cost. It also means high-end AMD processors are unsuitable for small systems.
Discrete Graphics Doesn’t Make Money
Though the company is slipping in the processor space, it is still a strong competitor in the market for discrete graphics cards. AMD and Nvidia are neck-and-neck right now, and there have been several times in recent history that saw AMD’s products offer superior value.
Yet in the second quarter of 2011, though AMD managed a profit, the company specifically stated that the GPU division had lost $7 million. But what is perhaps more shocking is that in the second quarter of 2010, when the GPU division was making money, it only made $33 million. That’s peanuts.
The outlook is grim. Discrete graphics shipments are expected to decrease while integrated graphics become more capable. AMD does have an edge in integrated graphics, but it has no good processors to pair with.
Lack Of Advanced Fab Technology
AMD is a fabless company. In 2009, AMD spun its fabs off into a separate company called Global Foundries in order to decrease overhead.
It is not uncommon for a chip company to be fabless. Both Nvidia and ARM are popular examples. But neither of these companies are direct competitors to Intel, which owns fabs and aggressively pursues new production technology.
As a result, Intel has an edge because it can introduce processors based on a new fabrication technology sooner than AMD processors, which is part of the reason why Intel so often has an edge in performance-per-watt.
Intel also has much tighter control over its production. It’s easier to ramp up production to meet demand when you own the facilities. Production constraints are plaguing Bulldozer, which has been out of stock frequently . That’s not a good thing – it means AMD could be selling more, but can’t because there aren’t enough available.
The Company Is Changing Direction
There’s been a few times in AMD’s life where the company was massively profitable. This seems to be taking a toll on the company. Earlier in 2011, the company’s board of directors forced CEO Dirk Meyer to step down. His replacement, Rory Read, wasted no time – he recently cut 1,400 employees from AMD, which translates to 12% of the company’s entire workforce.
Layoffs have hit the company before, but what’s interesting about this one is that almost all of AMD’s marketing staff was axed. This suggests that the company no longer is interested in marketing itself to consumers or enticing enthusiasts. In addition, AMD’s new CEO has stated that he sees the company’s focus to be in “low power, cloud, [and] emerging markets” rather than the desktop or even server realm.
I see this as an admission of defeat. The new CEO does not believe the company can compete with Intel, so the new plan may be to not compete at all. As the saying goes, desperate times call for desperate measures, and changing AMD’s focus from the market it’s competed in for over a decade is desperate indeed.
Conclusion: Being The Underdog Sucks
Writing this does not make me happy. I like AMD and its processors as a matter of principle. There was a time during which AMD was competitive with Intel, but that was squashed by the use of business practices that were found illegal by both the European Union and South Korea (Intel settled in the United States).
None of this changes the fact that AMD is in a terrible position. It can’t compete with Intel, and it can’t rely on discrete graphics for profits. In this sense, a sudden change in focus is logical – but it’s a difficult task, and I doubt the company will survive as an independent entity.