[Beyond3D Article] AMD Q1 2007 Warning Analysis

[Arun]

Article link
Please now use this thread for discussion of AMD's current financial problems rather than the "AMD needs money?" one, thank you!

As AMD announces that their Q1 revenue will fall significantly short of expectations, the biggest question that poses itself is why things are so much worse than in Q4, but also how difficult it will be to fix, and what other profitability numbers should be expected in Q1 and Q2 2007 for both CPUs and GPUs. Read on to see what we think, and what industry dynamics are at play...

 

[Razor1]

heh you are on a roll, great article!

 

[Geo]

I'll tell you this, they are either awesome bluffers (always possible), incompetent to the point of insanity (unlikely), . . . .or. . . .they're really not too worried about it and think they have it under control.

 

[trinibwoy]

Interesting analysis Arun. It was a good read. Any thoughts on how lowered pricing on R6xx products would fit into AMD's strategy in light of their current situation?

 

[Periander]

I registered just to say what an outstanding, in-depth article this is. It's good to see that not every enthusiast site has dumbed themselves down.

There is I believe a typo I'd like to point out on page 5:

Based on that, we can estimate gross profits of about $415M. The problem obivously is that their operating expenses are in the $670M+ range, which implies an operating loss of more than $225M.

Shouldn't that make the implied operating loss then be more than $255M?

 

[Arun]

heh you are on a roll, great article!

Thanks! Just in case that wasn't clear though, remember the Intel piece was a team effort, I only contributed part of the analysis. Not sure if you meant 'you' as in Beyond3D or not, so I thought I might as well point that out now.

I'll tell you this, they are either awesome bluffers (always possible), incompetent to the point of insanity (unlikely), . . . .or. . . .they're really not too worried about it and think they have it under control.

My honest opinion is a little bit of A, a little bit of B, and a little bit of C. I mean, it'd make sense for them to forecast higher than they expect them to manage, but I'd be quite surprised if they expected such rapid decline.

At the same time, as for 'having it under control', I'd expect management to be quite confident in the Barcelona. Whether that trust is misplaced or not remains to be seen, of course.

Another thing to consider is that *real* CapEx might be $250M+ lower than what the claimed numbers are, because of grants and subsidies. Fabtech's Mark Osborne has recently claimed that he believes much of Fab30's conversion will be paid by the Saxony government, although I couldn't find any other source of information implying subsidies would be that massive. Either way, that *is* $200-300M that they could get away from having to pay.
http://www.fabtech.org/content/view/2482/73/
http://businessweek.brand.edgar-onl...ingHTML1?SessionID=oaQIjAWGR6h_1Kr&ID=5002366

Interesting analysis Arun. It was a good read. Any thoughts on how lowered pricing on R6xx products would fit into AMD's strategy in light of their current situation?

In Q4, former-ATI operations lost $13M (-$33M from Graphics & Chipsets, and +$20M for CE/Handhelds)...

I don't think the pricing of R600 is very significant to them, as long as they still make a small profit. Even with a rather high profit margin, it would be unlikely to drive much gross profit at this point, so that's probably not their biggest worry. I don't think they can afford pricing RV610/RV630 more competitively than needed, though, obviously...

Either way, I'm sure AMD realizes that if they don't undercut NVIDIA's current pricing for a given performance level, their launch can be completely ridiculized by a combination of new drivers and minimal price cuts. I doubt they have much of a choice there, but they are very unlikely to make them ridiculously cheaper than they have to be in order to be competitive, either.

I registered just to say what an outstanding, in-depth article this is. It's good to see that not every enthusiast site has dumbed themselves down.

Thanks!

As for the typo, well, I guess you caught me there! It's not really a typo. It's more that I decided to make the figures look less pessimistic prior to publishing, because I was accurately pointed out that operating expenses very well might be lower than $670M. So I forgot to change the expenses number correctly to match the lower loss - oops! Thanks, fixed now.

 

[Geo]

My honest opinion is a little bit of A, a little bit of B, and a little bit of C. I mean, it'd make sense for them to forecast higher than they expect them to manage, but I'd be quite surprised if they expected such rapid decline.

At the same time, as for 'having it under control', I'd expect management to be quite confident in the Barcelona. Whether that trust is misplaced or not remains to be seen, of course.

I think they are heavily fixated on OEM and marketshare. Which, given how long and hard they had to fight (unfairly, they believe, and have sued over) to break out of being a "channel only" player, is understandable. I just don't think they are suicidal over it, and pursuing it this aggressively and with nearly cavalier disregard for the short term consequences does bespeak a great deal of confidence in the midterm.

But then miscalculations have been made before in the tech industry, so who knows.

 

[3dilettante]

AMD's really getting hammered.

http://www.tgdaily.com/index.php?option=com_content&task=view&id=31666&Itemid=118

Cripes, I thought it was bad now, but would likely get worse or stay bad for the rest of the year.

I'm not an expert, so when I say I didn't think it was that bad, it doesn't mean much.

But man, I didn't think it was that bad.

 

[Geeforcer]

I expected AMD share price to go up today because Intel had a solid quarter and some people seem to operate on the bizarre principle: "If Intel is doing well, then the industry is doing well and since AMD is in the industry, let's buy their stock as well". I wonder at what point the fact that Intel is doing well at the expense of AMD will register. Probably Thursday.

 

[Deusp]

I expected AMD share price to go up today because Intel had a solid quarter and some people seem to operate on the bizarre principle: "If Intel is doing well, then the industry is doing well and since AMD is in the industry, let's buy their stock as well". I wonder at what point the fact that Intel is doing well at the expense of AMD will register. Probably Thursday.

You're right, AMD's stock price has gone up so far today.

 

[Arun]

Cripes, I thought it was bad now, but would likely get worse or stay bad for the rest of the year.

I agree with that statement, but I don't really agree with the source/link, let alone because I think iSuppli are some of the most incompetent fools in the industry imo! No offense to them, but some of their stuff is just really bad. There are exceptions of course, but...

Anyway, in this instance, I wouldn't be surprised if they just concluded that since revenue was reduced by one third, so was marketshare. That would obviously be completely incorrect, since ASPs are down too and Q1 is always down slightly compared to Q4. But yeah, things are worse than I was expecting them to be myself. We'll see what the conference call tommorow reveals...

 

[Razor1]

Thanks! Just in case that wasn't clear though, remember the Intel piece was a team effort, I only contributed part of the analysis. Not sure if you meant 'you' as in Beyond3D or not, so I thought I might as well point that out now.

Ah cool, in any case, that was one of the best analysis from top to bottom I have seen on AMD's situation thus far, and I've read quite a few papers on the matter :smile:

Isuppli is good for numbers like units sold and BOM basic stuff (a good starting point), the analysis part well, I don't think I need to explain any further lol.

 

[jimmyjames123]

I agree with all the others, great article and great work Arun!

 

[Geeforcer]

Number to watch tomorrow: over/under 40% margins.

 

[_xxx_]

So what's up with that conference call? Wasn't that today?

 

[overclocked_enthusiasm]

Here is the big problem for AMD:
http://www.tgdaily.com/images/stories/article_images/processor_performance/20070413/data/data3.jpg

They have lost all pricing power and ASPs/gross margins are in the toilet. There is a real risk that GM's dropped below 40% and if they did AMD's loss could be $500 million or more this quarter. I expect the same or slightly worse in Q2 so AMD will be down to less than $500 million in cash which is one quarter of operating cash. News out last night was that Spansion results were double the loss previously expected and AMD still owns like 40% of Spansion as well.

The cash crunch is what AMD must face up to this CC. AMD MUST get more cash to operate and fund their CapEx plans or Intel will simply vault ahead with 45 nm etc. To do this AMD must also pay back the vast majority of their financing used in the ATI deal FIRST as part of any new financing...thus complicating the deal. AMD authorized the issuance of another 750 million shares (not dollars but units) to help fund their ongoing operations and cash requirements. Any such issuance will send AMD stock tumbling with the dilution associated with new stock.

If AMD cannot recapture the performance and technology lead with Barcelona, Agena and Agena FX then AMD is truly screwed as they will be forced to slash prices (see above chart) as their only tool to slow market share loss to Intel. If 45 nm Perynyn continues the domination of Core2Duo then AMD is likely in very serious trouble and I am not so sure they can survive as is. What does that mean? That means they will have to sell off assets such as Spansion, raise capital and otherwise restructure their business model.

Here are the big questions for AMD today:

1. Gross margins
2. Sequential unit shipment contraction
3. ASP contraction
4. Market share loss
5. INVENTORY
6. Possible writedown of inventory
7. Reduction in full year 2007 quidance given at ANALyst day back in December including 50% GMs
8. Q2 and full year 2007 guidance

Listen to the Intel CC from Tuesday and you will see how Intel set up AMD on most of these issues for today. the questions will be asked and the answers will be ugly. The pop in stock price yesterday made no sense to me as I think the results today and guidance will be horrific. The big issues that AMD will be asked about over and over today is GM's, inventory, ASPs and market share...with a healthy does of raising capital thrown in.

p.s. The AMD executive who came out last week and said AMD did not and has not lost any market share was lying through his teeth unless AMD was simply giving stuff away.

 

[3dilettante]

Trouble is, Barcelona isn't expected to impact AMD's fortunes too much in 2007.
Given the high margins of the server part, and the die size, I would imagine a lot of capacity used for the new cores will be going there.

AMD's mix is still chock full of 90nm parts, a lot of it in the high-end because the 65nm chips so far have failed to clock as high.

I'm not sure when it is AMD will be fully 65nm, and it may still have 90nm product when Penryn comes around.

Even if Barcelona does compete well with Core2, it won't matter too much financially before the likely introduction of a very good 45nm competitor from a company that is far faster at making product transitions.

 

[Arun]

I don't like quoting myself, and maybe I'm being a tad overly pessimistic (we'll see about that in a few hours), but:

First, let's just look at operating income excluding one-time items, which thus corresponds to gross profit minus operating expenses. As can be seen in the tables on Page 1, non-GAAP margins were 40% in Q406. Given that CPUs had higher margins than GPUs, and that GPUs will now represent a higher proportion of revenue while CPU ASPs also went down, it seems likely that non-GAAP gross margins of roughly 34% are an optimistic estimate at this point.

In the unlikely event that margins are a fair bit higher than that, then chances are they are building inventory (remember margins are lower if your capacity is underutilized, because fab running costs etc. are counted in there!) and given that they are also losing marketshare, you'd likely see an inventory write-down in Q2 or Q3, and especially so since they keep cutting prices! Either way, my expectations for Q1 are non-GAAP gross margins between 33% and 37%.

Also, I don't want to sound like I'm trying to make news worse than they are already... but I don't believe Barcelona/Agena/Kuma will change anything. Penryn will boost their server FSB from 1066 to 1600MHz, reducing the advantage AMD has from the IMC, and AMD will likely be competing with a 2.4-2.7GHz part against a 3.33GHz Intel Quadcore as early as Q1 2008. By the time their volumes ramp up, they'll have to cut prices! Ouch.

As for the desktop market, Penryn is a very small and power-efficient chip. In a price war, that's exactly what you need to crush your rival. Another very important point there is that whoever has the performance crown dictates the market's pricing. If you're 20% faster, you can release your highest-end chip at $999 and one equal with the competition's best at $300 or less to hurt their margins. Surprise, surprise... That's exactly what Intel seems to be doing in September with the E6850 and the Q6600: http://www.hkepc.com/bbs/hwdb.php?tid=753250&tp=Intel-c2d-e6050&rid=753256 (if Agena FX isn't much faster than the Q6700, they'll be competing with a 280mm2+ die in the sub-$300 space; ouch!)

We'll see what happens. The problem is that as far as I can see, all of the former-ATI never had an operating profit anywhere as high as the operating loss AMD's CPU business will suffer from in Q1. If things are doing nice there, it might help a bit, but it won't do miracles either. It's really up to their CPU business to fix itself, somehow, and the 12 months will be decisive. Either they succeed, or they'll crash and burn, with all the crap that implies for the industry as a whole.

 

[3dilettante]

Agena FX doesn't strike me as a product that will have the volumes necessary to significantly impact AMD's gross margin. They could probably afford to go somewhat cut-rate on the upper end of the enthusiast segment, for all that it does for them.

If some of the tables indicating Agena FX or a variant may be targeted at the quad FX platform, then the volumes will likely be miniscule.

Barcelona's the important one, and AMD's made the server segment a high priority. The market pays the good margins and in volume, and it knows how to take advantage of multiple cores. It's also the one segment (4-8 (16?) sockets) that AMD stands a chance of competing on performance with Intel, since it's there that the platform can really matter.

The fact that AMD stated it doesn't expect its savior chip to have significant effects for all of 2007 is extremely worrisome, especially given the heavy emphasis AMD is putting into Barcelona with a quarter's lead time over all other variants.

 

[epicstruggle]

OUCH!

Advanced Micro Devices Inc. (AMD.N: Quote, Profile , Research) posted a bigger-than-expected quarterly loss on Thursday as the No. 2 maker of computer processors was hit by falling prices and unit shipments.

AMD said its net loss for the fiscal first quarter was $611 million, or $1.11 per share, compared to a profit of $184.5 million, or 38 cents per share, a year earlier. Revenue fell 7.4 percent to $1.23 billion.