Per the title, no, that’s premature and panicky, but the latest AMD financial news can hardly be called good.
AMD announced its revenue will be down another $100 million for the third quarter.
The flash memory business actually showed some improvement, so all that drop comes from the processor side. That means AMD’s CPU revenues dropped from $380 million to probably about $270 million.
Prices didn’t drop all that much for most of the quarter, and when AMD’s PR starts talking a lot about marketshare, you know sales took a big hit. 🙂 Sorry, AMD PR people, you’re too predictable.
So the breakdown we’ll hear in two weeks is probably a bit over five million processors at an average selling price of a bit over $50.
We’re probably looking at a loss at or a bit above $250 million for the quarter.
From my own personal perspective, what’s scary about all thiz is that for the last couple quarters, I’ve been pretty gloomy about AMD’s short-term prospects, and then they do ever worse.
Sorry, AMDroids, but when your company goes from about $700 million in CPU revenue to less than $300 million in six months, that’s just not good.
Any Rays of Hope?
There was no chance these numbers were going to be any good for the third quarter, but, provided AMD can get 2400+ and 2600+ out in bulk this quarter, the ASPs should be able to crawl out from the sewer into at least the gutter.
How much PC sales will pick up for Christmas is anybody’s guess, but since AMD has lost disproportionate sales, a decent pickup will likely mean disproportionate gains.
We don’t know about Intel sales yet, but if these Digitimes numbers are right, Intel’s October 25 price cuts about as gentle as AMD could have hoped for. That’s good news for AMD is that Intel isn’t going in for the kill; it’s good news for Intel in that the stable pricing for most lower-end models is probably an indicator that inventories are probably under control.
What To Watch For
1) The AMD Conference Call in two weeks should be very interesting. If you listen in yourself, see how many questions are variations on the “Are you going to die?” theme. If there’s only a few, that’s good. If a large proportion of the questions are like that, that’s bad, because it means the analyst rats are looking to jump off the sinking ship. If a large proportion of the questions are more like, “How can you expect not to die?” that’s even worse, because the rats are just trying to make themselves look good before they jump.
2) Look to see how many places are selling 2400/2600s as the next few weeks progress. If by, say Halloween, the Pricewatch CPU list is 5+ pages long, that’s good. If it’s more like 2 pages, that’s bad.
3) If Intel’s price cuts end up being what Digitimes says they’ll be, that’s good. If they’re more aggressive, that’s bad.
As said in previous articles, the Christmas quarter is critical. If AMD can get $500 or million in more in sales, that’s good (well not good, but at least not too bad). If they don’t do appreciably better than this quarter’s $270 million; that’s not only would be bad, AlMost Dead would start looking appropriate. Fortunately, unless those TBred2s stay largely vaporous, that’s unlikely.