We’re currently in Austin, TX for an event at the Texas Advanced Computing Center (TACC), which is currently completing work on Stampede, a 9+ PFLOPS system which carries significant amount of Xeon E5 and Xeon Phi processors.
Intel’s timing and location isn’t accidental – according to sources in the know, the event was originally supposed to take place in Oregon last week, but the company pushed the event to the date of AMD’s announcement of its "no bueno" third quarter results, restructuring and culling down 25-30% of its workforce, which brings us to the subject of this story.
Over the past few days, we’ve meet with quite a few AMD insiders and talked the future of the company, which is looking quite brighter than the "doom and gloom" which former and current employees are murmuring about. There’s no doubt that AMD’s biggest problem is execution, which goes from companies representative playing games in the advertising space with media and financial analysts, and then cries wolf when the said media or analyst houses turn their backs on them. In fact, we received information that AMD "helped" the launch of a certain tech website with high single digit support on a monthly basis, which in turn ended with that media attacking its competitors, especially on the graphics front.
This was the strategy on the investment front as well, with the former senior level executive allegedly paying off certain financial analysts significant sums of money, and when the executive left – certain houses downgraded the company since there won’t be lofty six to low-seven figure "special projects" for 6-12 months of work. Bear in mind, it’s not just AMD – this is how pretty much any public company work.
Furthermore, the departure of ex-CFO and the ex-interim CEO was greeted with the warm welcome, both from current and former management. One of people we talked with compared the former CFO with the movie Dictator, in which Sascha Baron Cohen portrays the figure of a dictator which is completely clueless and whoever disagrees with him, gets his head chopped off. In reality, those people are being shipped off to America, which is a "not so bad death sentence". According to sources in the know, the ex-CFO had quite the same attitude, trying to put down financial value to any of the technologies involved, firing off people which disagreed with the individual.
The number of people that were let go in the last round of firing only confirms the statements said above, but the upcoming cut is of a different kind. For too long, AMD has had multi-million and even billion dollar opportunities slip through the cracks, since projects that would be approved by C-level management would get killed off by jealous VP-level executives, or it would dwindle down to the directors level, neatly sent to the dustbin. Or to Intel, Asus, Nvidia and plenty of others, who would proceed to complete those projects and score big time.
Good examples in missed opportunities were the netbook category or the Apple MacBook Air. The new executives we’ve talked with have quite a different view of where the company needs to go, and they’re currently in the shopping mode for open minded engineers and executives, as well as the companies which affect the future of the company.
Unlike the vast majority of media that continuously puts doom and gloom over the future of the company, AMD is downsizing to the point where new and old staff will have the opportunity to actually change the process. Currently, the company already completely reshaped its roadmap for next two years, killing products and is hedging all the bets on the new low power CPU architecture which will combine with ARM processors, complementing the future high-performance, high-power GPU parts, which are gaining more and more capabilities.
We’re of course, talking about future GPU families, such as Sea Islands (SI2 – 2013), Volcanic Islands (VI – 2014), Pirate Islands (PI – 2015) and Fantasy Islands (FI – 2016). This strategy includes re-hiring some of very bright minds that were pushed out by the former management, and poaching talent from other companies.
The biggest question is of course – are the changes enough to turn the tide? More importantly, can AMD learn from its mistakes and stop hedging the bets on wrong projects and companies?