Monday, July 23, 2007

AMD Q2 - what the...

Ok - I did not see it coming the way it did finally. I did mention earlier that AMD might start giving stuff away at close to cost to fill their factories and win share - every unit they sell is a unit Intel doesn't sell. However, this is the weirdest result I've every seen.

Significant revenue and gross margin growth in spite of which losses that were similar to Q1. Get the details here:

Q2 Results

Q2 Earnings Call Transcript


So here's what I think really happened.

1. AMD sold as much as they could at the low end to maximize their factory utilization to improve GM's.
2. Their transition to 65 nm coupled with the fact that they are not incurring 45nm start up costs improved margins.
3. I had mentioned mobile earlier - this did help somewhat. More importantly, they were able to ride the overall growth in mobile through Toshiba.
4. They also seem to have made some recovery in servers.

So what went wrong - how did they lose almost the same amount of money as Q1:

1. The ATI acquisition. Graphics revenues down YoY. Losses continue. And...the acquisition costs hit their profits.
2. They are selling at the low end. So even though they seem to have improved margins QoQ, the reality is those are not due to improving ASPs but more a function of higher units to fill the factories and some mobile/server uplift.
3. The had 30 million additional MGA expenses. Now how much of that do you think are sales/marketing kickbacks to some of their customers to off load the additional volume.

This was a weird quarter for AMD. If they have stuffed inventory into their customers again, then Q3 might be a tough quarter for them again.



7 comments:

Roborat, Ph.D said...

180 said: This was a weird quarter for AMD. If they have stuffed inventory into their customers again, then Q3 might be a tough quarter for them again.

i don't think AMD shipped more than demand. remember that there was a slight increase in market demand in Q2. Also, AMD's inventory numbers went down by $100M. I think AMD's volume was within seasonal trends, its just that ASP's were absolutely terrible.

Anonymous said...

Weird... why, cuz the all seeing, all knowing sharikou 180 didn't see it coming.

Have you picked up a flier or ad for any of the large chains recently (or the past year for that matter?) There's usually 3 or 4 AMD systems listed for every Intel one. And nearly every one beats its Intel counterpart on price/perf. Its easy to see for most impartial observers that AMD leads in price by the same margin as Intel leads in performance. The second part of the stategy is beginning with a slight gain in servers, but will only come to bear after Barcelona. You get the low end and you get the very high end and then squeeze.

Anonymous said...

"Have you picked up a flier or ad for any of the large chains recently (or the past year for that matter?) There's usually 3 or 4 AMD systems listed for every Intel one."

Ahhh.. we have a budding financial genius...

Out of curiosity, do you have any clue about just how much these large chains account for in terms of total CPU sales? Oh and if you're talking anout just US chains (don't know if this is what you are referring to)... that initially small % will get divided by ~3.

"You get the low end and you get the very high end and then squeeze"

Yeah, you get the low volume niche products (4P+) and then you get the low margin low end parts and THEN worry about the sweet spot of the market (which by the way is MOBILE and 1P/2P servers!).

Yeah let's grow units by 38% and incur the same loss...that will help fund future capacity expansion... er.... what we need to cut capex to help fund this market share at all cost strategy?

Anonymous said...

Silly anonymous (no wonder he ticked the anonymous option) said:
"Ahhh.. we have a budding financial genius..." "that initially small % will get divided by ~3."

What we do indeed have in you is a statistical and mathematical genius...

You can put your curiosity to rest, U.S. retail sales account for a lousy 8-10%... of the entire global pc market. But only someone such as yourself would even use this useless stat for any meaningful comparison. That's right, U.S. retail should be veiwed in terms of the U.S. market, first and foremost. And the U.S. market accounts for 25-30% of global sales. Not divide genius, but multiply ~10% x ~4 = ~40% of U.S. pc sales are retail. And this number shouldn't be that surprising, most people still covet what they see and want to see what they buy. Here you thought 90% of pc's were bought online, what century you living in? I'd provide links but since you confuse x and /, then letters and numbers are really gonna mess you up.

Unknown said...

If AMD doesn't bring 2.6GHZ Barcelona to bear this year, Penryn will squeeeze AMD. Until they get the ASPs back up, how can anyone forsee profits of any kind?

Anonymous said...

H8, don't you think that if Intel wanted to price their products so that they lost $600,000,000 a quarter they could capture any segment of the market they wanted and own the weekly Best Buy adds? Until AMD finds a way to compete profitably with a company that brings so much more to the table, they are playing a losing
hand. I'm convinced that Intel intends to bleed them dry and then we will all lose. Until then...great bargains.

Anonymous said...

'What we do indeed have in you is a statistical and mathematical genius...'

OK...smart one... let's get back to the original point...

EXACTLY how much does US retail sale account for in terms of overall X86 CPU market?

Thanks for "creatively clipping" my words... When I asked about US only, I was asking if you were talking about US B&M stores only.

So thanks for coming in here and spewing your genius about how 3 out of every 4 computers in US fliers are AMD and US retail sales account for 8-10% of global sales...

Few problems

1) you assume the ratio of computers seen in fliers is equal to what is actually sold in the store (it is not 75% if you look at AMD's data in their analyst day presentation)

2) Who care what's in the flier? If it's been like this for a year....AMD surely has parlayed this nicely into profits... look at how well the "flier ratio" mimics market share and/or company performance and/or actual chip performance.

Did you ever notice how Intel systems were usually on page 4 and now they are sometimes on page 3 and sometimes on page 7.... they must be having production supply problems...

Ridiculous? Yes? But so is taking an observation like # og systems in a flier and trying to turn into market commentary.

I see huge high cost plasma TV's advertised more often - is that the best price/performance?