AMD To Shed 10% of Its Workforce

stress_life writes "Recent rumors about AMD firing 5% of its workforce proved to be understated. AMD just announced that the company is going to deliver pink slips to 1600-1700 workers, or around 10% of its employees. AMD needs revenue of $2 billion per quarter, but Q1'08 is expected to come in around $1.5 billion. These firings have to be complete by Q3'08, the quarter by which Hector Ruiz promised to be profitable." We most recently discussed AMD's struggles in February.

Comments

And if...

[Creepy Crawler]

AMD dies, then Intel will jack their rates up about double.

We saw something like this with Blu-Ray when HDDVD was announced to be dead.

And Via.. Well, they're VIA. Leaves a bad taste in my mouth.

Buggy products

[dgym]

I wonder if they will be getting rid of the people who decided to release the Phenom X3s and the energy efficient Phenom X4 with the TLB bug intact? By releasing a lot of new chips at the same time, some with the fix and some without, it seems as though AMD are trying to confuse people into buying buggy chips with awful performance.

Apparently we have to wait even longer before this mess will be cleared up. Is it any surprise that revenue is down?

Re:Lay off 10%?

[NeutronCowboy]

Well, if your revenue is down, you've got to cut costs in order to remain profitable. Since employees are by far the largest expenditure (ignoring for a second the opening of a new fab), it makes sense to cut costs there. Furthermore, in light of the fact that the last set of products were pretty underwhelming, I'm sure that there was some fat there that needed some trimming.

That said - I agree with your feeling that executives never seem to take responsibility for screw-ups. Instead, they take million dollar golden parachutes into semi-retirement. I'd love to see an exec who says: "Wow, we stunk this year. I'm cutting my salary in half to help the company stay profitable." Or a CEO who says "Wow, we stunk these past two years. I'm obviously the wrong person to run this company, and am forfeiting all salary, bonuses and payments that were supposed to come my way." I guess that technically, the Board of Directors is supposed to do this, but that's a whole different issue.

10% layoff is healthy

[snsh]

Jack Welch at GE advocated the 20-70-10 principle which says to periodically purge the lowest-performing 10% of employees to keep a company healthy. First, it gets rid of nonproductive employees. Second, *not* firing the lowest 10% is bad for the morale of the top-performing 20%.

AMD has still a little head start...

[DrYak]

AMD did well until recently because, with the Athlon 64, they managed to bring something new fresh, interesting and with good performance :
- x86-compatible 64bits architecture, integrated memory controller, low power and thermal
exactly at a time when intel was stuck in dead ends :
- on one hand Itanium proved not to be the optimal way to bring 64bit to desktops and was stuck in the scientific cluster market
- on the other hand the Pentium 4, which was the CPU equivalent of a hummer, and was unable to go above 3GHz although the NetBurst architecture was planned to reach 10GHz

Intel had to lose time, going back to an older generation (PentiumIII-based PentiumM) and developing a decent workstation & desktop processor out of it (Core 2 was the first decent answer to Athlon 64).

Now we are back to the statu quo. With AMD having some technologically interesting products (true quad-cores) and interesting perf/price ratio in the mid-range products, but other wise no massive advantage.
And Intel throwing tons of resources and replaying the "Gigahertz race", except this time with the number of cores bolted to the same package, offering expensive but fast processors.

*BUT*

AMD could still get some advantage in the near future.

First, the gain obtained by multiplying the number of cores will soon top (My crystal balls predict somewhere around 6-8 cores). Intel is going to hit a wall soon, just like they got stuck with their Gigahertz race.

Second, integrated design with the memory controller on the CPU and a standard bus between the CPU and the rest of the PC seems to make a lot of sense. At least that's what Intel's engineer are thinking.
Here again AMD has some advantages :
They already have such an architecture since Athlon 64, the hypertransport bus has been adopted already by several other constructor for various (FPGA and other accelerators, or simply communication between multiple chipsets on motherboard with several northbridges), their socket has stabilised (thank to the compatible family AM2 => AM2+ => AM3).

Whereas Intel will probably once again lose some time developing and perfect their Quick-Path based processors, probably changing their connector a couple of time along the way (can't technically reuse LGA775, will have to develop a new one and as usually will probably change it a couple of time before stabilising), will have to convince other constructor to adopt it (they will, of course as they are "the standard x86 cpu that every PC maker use". But it'll take some additional time), etc...

Once again we will see a transition at Intel, during which AMD has a small advantage (smaller than with the Athlon 64, but still present).
If they leverage their advantage well (partnerships around the HyperTransport, perhaps), they can achieve some success.

Of course that advantage won't stay indefinitely, and after that Intel will probably be back again with big beasts. Probably by then the technology will better take advantage of bigger multicores. And they'll also have a good advantage in the GPU / GPGPU markets by then.