Friday, January 26, 2007

AMD suffers GAAP loss

Thought I'd point out that my several month old prediction that AMD would suffer a GAAP loss by Q1/Q2 2007 has come true earlier than I thought. Meanwhile, Intel is yet to suffer the GAAP loss the good Doctor Sharikou had predicted.

Where I will admit I might have to recant is on my prediction on Intel starting to re-gain share. While they've re-gained revenue share in Q4, it seems like they have not yet re-gained unit share. I have till Q1 07 and we shall see how it goes or else I will recant.

But meanwhile I'll enjoy the fact that I called the GAAP loss accurately and will graciously accept any public apology the Doctor wishes to make for mis-calling Intel's GAAP loss (among other things he's mis-called).

7 comments:

Anonymous said...

Unit share as both companies anounced is a bit vague to understand. Yes AMD gained share in mobile. But is that really that hard to do coming from the minimal amounts they had? No. Mobile is also the fastest growing segment, so did they actually take share from Intel or just a larger piece of the growing segment?

For server the fact that Intel seems able to stop the loss or even reverse the loss in this space significant to revenues on both sides.

Deep Thought 86 said...

Well, given they just started to put the ATI acquisition on the books starting this quarter that's to be expected. The CPU biz was profitable, although by a 20% smaller margin (40% Q4 vs 50% Q3). Market share was up again. ATI suffered a loss. Overall there was a big loss, and next Q that'll continue because of paying for the ATI merger, regardless of how CPU/GPU sales are.

Is it better to earn 40% on sales of $1bn or 50% on sales of $250mil? Financial analysts tend to be focused on margins and very short-term outlooks. Did you hear the audio cast? I was amazed to hear some questions asked which anybody who's been following AMD even casually on CNet or Dailytech type sites would know. And these people are responsible for making buy/sell recommendations that can result in hundreds of millions of dollars of stock activity!

Intel's market share is down, margins are down, profit is down. Neither company has done well this quarter, and both probably will do worse in the next 2Q's. After Q3 things get very interesting, 65nm transition will be mostly done at AMD hence a big leap in capacity, Barcelona out, Penryn out.

By the end of 2007 there will basically be a glut of capacity, and I think that's when the real pain will hit both of them, particularly Intel. The erosion of marketshare is a ticking time-bomb.

Inel BK Q208? Not a chance, but the possibility of a shrunken long-term outlook is there.

180 Sharikou said...

Well - yes, capacity is an issue right now. But Intel has definitely demonstrated a better ability to balance revenues vs margins vs ASP this quarter. Revenue above forecast, margin flat to Q3 (on under-loading charges) and ASPs up (probably on server growth). This was not a bad quarter for Intel considering where they were coming from. It wasn't great - but it wasn't bad.

As we head toward quad core, capacity may not be as much an issue as it is today.

As for why the analysts ask seemingly senseless questions. Because once a CEO is forced to provide an answer in a public forum, he becomes accountable and that's what they want to see...accountability for the things they are led to believe the company is going to make happen.

Roborat, Ph.D said...

The CPU biz was profitable, although by a 20% smaller margin (40% Q4 vs 50% Q3). Market share was up again.
I think the $500M is a one time charge… Market share being up wasn’t totally appreciated by the investors because it wasn’t aligned with the revenue increase. To investors this means AMD has to make more CPUs while earning less compared to Intel.

Is it better to earn 40% on sales of $1bn or 50% on sales of $250mil? Financial analysts tend to be focused on margins and very short-term outlooks.
Actually, profit margin is a long-term metric. It’s more long term than cash flow or revenue. The fact that AMDs ASPs and margins drop substantially rings alarm bells to investors suggesting AMDs direction is going rapidly towards unprofitability.

Did you hear the audio cast? I was amazed to hear some questions asked which anybody who's been following AMD even casually on CNet or Dailytech type sites would know. And these people are responsible for making buy/sell recommendations that can result in hundreds of millions of dollars of stock activity!
I felt sorry for AMD. They had to be very defensive. A lot of these investors where in the December analyst meeting and they all felt they were lied to. 4 weeks ago AMD never mentioned anything about a massive loss and a very huge drop in margins.


Intel's market share is down, margins are down, profit is down. Neither company has done well this quarter, and both probably will do worse in the next 2Q's. After Q3 things get very interesting, 65nm transition will be mostly done at AMD hence a big leap in capacity, Barcelona out, Penryn out.
The market grows 6-8% in Q3-Q4. This is the benchmark. AMD’s revenue grew 3% = half the market. Intel grew 11%.. Intel’s margins went up while AMD’s drop significantly. Just by this metric, we know that AMD lost money to Intel. You should ignore the YoY financial statement as investors are more concerned going forward and most of AMDs gains are in the front end of 2006. I’d pay more attention to financial trends.

By the end of 2007 there will basically be a glut of capacity, and I think that's when the real pain will hit both of them, particularly Intel. The erosion of marketshare is a ticking time-bomb.
Why particularly Intel? Q406 showed Intel to be trending slightly upward while AMD dropped significantly? The scary part of it for AMD is they really have nothing to offer until Q307 and even AMD’s VP said that it will only have a financial effect in Q407. I don’t think Intel will have any problems with AMD taking up the budget segments.

Intel BK Q208? Not a chance, but the possibility of a shrunken long-term outlook is there.
I agree that very long term wise profit margins for both companies will become similar to ATI-NVIDIA= 20-30%.

180 Sharikou said...

Dr. Yield - less capital intensity does not necessarily mean lower costs to develop technology and nor does it mean lower costs to manufacture. In fact, because you are beholden to someone else's manufacturing advances you must design products to meet the capabilities of their manufacturing roadmap. Which means if they lag then you also lag as you can see even Chartered is still behind Intel on 45nm in spite of having nothing to distract them from their manufacturing.

Being fabless does mean lower capital intensity and the benefit of that is if your business model was to monetize your intellectual property then you could enter markets where the entry barriers are high capital costs (like semiconductors). However, the instant you decide to build a product, you face the same issues. The advantage AMD does have is capacity flexibility which works to their advantage when their demand shrinks as they can fill their own fab and reduce their supply from Chartered which means no or less under-loading charges. Also, AMD must share their profits with Chartered which means they cannot afford to work at the same margin structure as Intel could. As you can see this quarter, this is the game Intel has started to roll the dice on. They are trying to squeeze AMD's profits and hence cash situation which reduces their ability to invest in their future - R&D, capacity, people and marketing.

The reason it works in the graphics guys is that both ATI & Nvidia are fabless which means the only differentiator is their IP. However, AMD cannot compete with Intel without fabs as the strategic advantages to managing your own technology roadmap & more importantly capacity give you tools to lever the market.

As for 20-30% margins. Is it possible...YES. Is it probably - NO. Neither Intel nor AMD are structured to get there and neither wants to. AMD will cede volumes and share moving forward if it means reviving margins. As you can see, they cannot afford to keep lowering ASPs on the client market as Intel continues to re-gain share in the server market thereby limiting their high margin and most profitable side of the business. Long term - 60% margins may not be sustainable but to get to 30% will take a couple of really stupid CEO's at Intel and AMD. The other thing is AMD cannot afford to only fight Intel on CPUs...they must get into platforms and hence higher margins as well as share of revenue inside the client. Which means if they want to be paid a premium there are two competencies they must invest in. Software and branding. If they use software available freely to all from the market then they cannot charge a premium and all they're selling is a bundle with some future technical roadmap commitment to the corporate market but no value add for the consumer biz. Branding allows them to command end user desire and hence command the premium. Both these need investments and Intel is starting to choke the cash flow.

Anonymous said...

"65nm transition will be mostly done at AMD hence a big leap in capacity"

65nm is giving a 33% die area reduction - why is everyone expected a "big leap" when this conversion is done and by the way people keep calling the conversion done when F36 is complete, but there is still a huge amount of 90nm production in F30 (yes it will scale down on H2'07 but it will still be substantial)

Finally factor please factor in the impact of AMD's sales only being ~30% dual core into your "big leap" - as their mix goes >50% dual core, and quad core starts coming up, the larger aggregate die size will offset a lot of the scaling achieved by going from 90nm to 65nm.

Anonymous said...

"I agree that very long term wise profit margins for both companies will become similar to ATI-NVIDIA= 20-30%."

Are you guys talking about GROSS MARGIN or PROFIT MARGIN or OPERATING MARGIN?

Nvidia's operating margin and proft margin is in the 13-15% range. I think you are meaning to refer to gross margin?