AMD has long been weak on the non-technical leadership side, the technical leadership was not in question. When Dirk Meyer was summarily ousted, SemiAccurate saw several paths ahead for the company. On the technical side, there was absolutely no up side, the firing would hurt long term and replacing Dirk Meyer with a similar talent was not really possible.
On the non-technical management side, especially surrounding the retail channel, things had room for improvement. If the board picked a candidate who understood the retail channel, it would be a serious short term gain. If strong technical leadership was also put in place, that could minimize the damage. The less said about the board, the better, but the retail aware CEO seems to be in place, as is a strong technical team.
Part of the problem was in the transition. It took forever, far too long to be a tolerable thing, and the person put in place to hold the fort, Thomas Seifert was not a technical nor a retail person. He came to AMD from Qimonda, you know the one that went bankrupt, and to be fair was put in the CEO spot somewhat accidentally. He was also well liked by the financial community, but was singled out for scorn by many current and former technically aware AMDers.
Why? He didn’t understand the semiconductor business cycle and how things have to be done on a schedule. Intel has their “tick-tock” model, but the same general idea has held true for semiconductor companies for a long time. Moore’s Law dictates a pricing and cost curve, and if you don’t stay on top of it, you fall behind with frightening rapidity. If you miss a single point, your company is at a serious disadvantage, two is likely fatal. You can’t tweak an existing plan much and have the result not be woefully out of sync, it is do or die.
One thing that several people credit Mr Seifert with is something called “Project Win”. This is described by some AMDers as basically putting the development pipeline on hold to balance the books. The exact details on how it was implemented were not shared with SemiAccurate, but many many sources all say the same thing, it more or less stopped development. It also likely balanced the books, much to the joy of the financial set. Short term it worked, long term it was the beginning and a good portion of the end. That said, it is exactly what a financially oriented CEO that doesn’t understand semiconductor development would be expected to do.
The technically aware among the management and employees were somewhat aghast. They saw the problems with Moore’s Law and the consequences of falling off the curve. Many fled, including the best and the brightest. A new non-technical CEO was put in place after a massive delay, and questions of whether he understood the semiconductor development cycle were common. A raft of senior technical hires followed, and those were given quite high grades by many observers, but the questions still lingered. You just can’t pause development at a semiconductor company without massive consequences, but AMD did.
Then came the most ham-handed mass firing SemiAccurate has ever seen. To make matters worse, like the upcoming axe swinging, it seems to have been done more to carry out internal vendettas and consolidate power among the survivors. It didn’t just destroy morale internally, it also caused chaos for most product plans. This lead to further delays and slowdowns in the product cycle.
Project Win, coupled with unprecedented management churn, led to unacceptable delays. Llano was late, but that was not the fault of the new order. The successor, Trinity, missed its intended Christmas 2011 debut, and came out in a destructive dribble that sapped any remaining confidence in the company. Why this was done, to protect inventory for both AMD and partners, had roots in sanity, but it simply was not explained to any onlookers, full stop. There was no attempt to get the AMD side of the story out, partially explained by the firing of 90% of the people who would do that job.
Products slipped for technical and non-technical reasons, but almost all had a solid and well thought out reason behind them. Damned if that message ever got out though. Instead, AMD management seemed to revel in watching these self-inflicted wounds fester. Confidence in the company plummeted, and this compounded the actual problems. No one believed that the new crew could draw a roadmap with crayons and drool much less carry out the technical side of execution.
Instead of straight talk we got the same damn chirpy messages about teamwork, execution, and sports, but nothing other than analogies. It got so unbearable that it was mocked by the press, (Editors note: We mean press other than SemiAccurate that don’t normally mock executives as a matter of course.) but still nothing was messaged. Wall Street analysts repeatedly stated that AMD needed to put forth a plan that was clear, comprehensive, and achievable. Cue more peppy speeches filled with the finest fluff but devoid of anything that the company needed to say. The worst part is that there is a plan, it is a good one, and it probably would have worked. Unfortunately management never stopped digging.
Instead of explaining that plan internally, externally, and to anyone who would listen, AMD is about to cut the only remaining people who can carry out that plan. The engineers, the only group that could possibly save the company, are said to be on the chopping block. Even if there is only another 10% cut in workforce, the churn that it will cause will delay the roadmap yet more. SemiAccurate moles say that the cuts are far deeper, especially in the places that will disproportionately hurt the roadmap. The plans to patch the holes will only exacerbate the problem. No, they are flat out insane.
AMD senior management, or (mis)management, as we are now calling them, have delayed the roadmap past the critical point. Project Win was survivable, barely. The churn of technical talent made things worse, far worse, and put the company at the breaking point. Layoffs sapped confidence, and senior management was negligent in not messaging a damn thing to those who mattered internally and externally. The cuts that will follow ensure that the plans in place are not achievable, and SemiAccurate can not see AMD surviving at this point.
To make matters worse, the original firing of Dirk Meyer was done for technically illiterate reasons that have not been stated publicly. The AMD board of directors had a plan and it is both technically naive and counter to the plans in place at the time. Instead of listening to the people who disagreed, they fired dissenters and put in to place a team that would first and foremost not question their new disastrous direction.
Any of these things individually were survivable. Instead, AMD has done its best to do the wrong thing at every step of the way. From direction to leadership, from institutional knowledge to morale, from internal to external messaging, everything was done wrong and never corrected, only compounded. Few outsiders know AMD’s direction and because of (mis)management ineptitude, no one cares. If AMD comes out later this week and hits a home run messaging their plan, there is noone still listening, they lost their audience long ago. By October 25th, there will be no one left to carry it out, it’s too late now. Game over.S|A
Author’s note: There are other reasons for AMD melting down both internal and external. As we receive more information in the coming days and weeks, we will put out what we can.
Latest posts by Charlie Demerjian (see all)
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018
- Intel can’t supply 14nm Xeons, HPE directly recommends AMD Epyc - Sep 7, 2018
- AMD reintroduces the Athlon name with two CPUs - Sep 6, 2018