AMD's Market Share Gains Shows Deeper Problems

The latest numbers from Mercury Research confirms AMD gaining back 4% of the overall PC market share. The report shows AMD at 22.9% in Q2 2007, up from 18.7% from the previous quarter. Intel moved down to 76.3% from a high of 80.5% last quarter. All 3 segments (mobile, desktop and server) showed double digit growth from the previous quarters for not so clear reasons:

"The market demand in general was pretty strong and to be honest, no one understands why this particular second quarter was this strong," said Dean McCarron, founder and principal analyst for Mercury Research. "All the indications were for a low forecast and this quarter just blew those forecasts away."

McCarron continued to explain AMD’s inventory in Q1 and the unusually lows market share figures:

"Basically, AMD had an overstated share in the fourth quarter and an understated share of the market in the first quarter," McCarron said.

I hope this ends the discussion. AMD initially admitted to the problem and now Mercury Research corroborates with my theory. Between AMD and Mercury, I don’t think there is anyone who is more credible so we’ll leave it at that. (Those who were in vehement opposition need not worry because I promise I will not say “I told you so!” ;)

Going back to the market share report where it proves that AMD did maintain some of its previous gains, on the surface this represent good news for AMD. AMD wanted to increase mindshare and market penetration and this somewhat proves that taking back share is never going to be easy for Intel. But there is a fundamental problem that the 4% gain amplifies. If AMD was shipping at record volumes and at a market share of 22.9% on an exceptional quarter, then what would it take for AMD to return to profitability? 30% market share on a miraculous quarter? This is as good as a quarter gets. The loss of $600M at full capacity only points to a bigger problem and one that cannot be solved by simply producing more. But I seriously think AMD is on a “scorched earth” strategy with complete a disregard for creating investor value. Still, there is no announcement of a business restructure and because of that we can only expect AMD to lose at least half a $Billion every quarter. As for Intel, the amount of market share that AMD continues to hold is very frightening especially if you own shares. It must be very frustrating for Intel to maintain investor interest while at the same time fighting a rival on a destructive kamikaze mission.


Anonymous said...

More share more problems

- Notorious AMD

Anonymous said...

Interesting analysis, but you omit one thing. Even if market share holds constant in Q3, AMD will fare better in terms of profitability because overall x86 market will increase seasonally (Q4 as well) - So I would expect AMD to come in around ~300Mil +/-50Mil in losses. I think their ATI acquisition writeoff costs are now done. Also the overall market continues to shift from desktop to mobile which will help ASP's for both companies.

I do think your analysis is dead on about the scorched earth policy though - Ruiz should be removed for this ridiculous strategy. The fluctuating #'s over the last 3 quarters suggest even if Ruiz had been able to hit 30% share it likely would not have been healthy nor sustainable.

Now AMD because of the cash crunch, due in part to this ridiculous strategy, has sacrificed production capacity when they will need it (K10) as they were forced to pushout 200Mil in H2'07, while not a loton an absolute scale, this is ~20% of their H2'07 CAPEX...

Of course some (the same some who also told everyone that AMD did not stuff the channel and the Q1 dip was due to F36 conversions!) would have you believe AMD is pushing out capacity because their 65nm yield is so good and their process is so healthy.

Anonymous said...

OK, no other way to state this other than Scientia is a complete idiot who understands nothing about Si process technology....

"WRONG, if 90 and 65nm are any indication AMD/IBM process has been inferior to INTEL offerings."

DEMENTIA RESPONSE: Let's see. AMD's 90nm used less power and clocked higher. It also appears that AMD's 65nm will clock higher as well with similar power draw. I guess I'm missing the part where AMD is worse.

Hmmm...yeah let's look at 2 DIFFERENT archtictures and draw conclusions on PROCESS TECHNOLOGY based on clock speed. Is he really that clueless? Why not look at P4 90nm which had higher clockspeed or Core notebook chips which had lower power and use that to draw conclusions?

Apparently Scientia (theoretically an educated man) cannot draw a distinction between a difference in architecture vs a difference in process technology. If he had even a vasic background he would look at the technology node performance metrics to draw conclusions between process technologies.

Here's a case study - my car goes 0-60 faster than your so my engine is clearly superior to yours no? (well possibly, but there could bo other factors involved like weight, body design, etc...)

"INTEL seems to have gotten by with tried and true dry lithography."

DEMENTIA RESPONSE: Right, Intel wasn't up to it.

So the ability to EXTEND an existing technology further is a negative / drawback? Let's see we can either do what we've been doing for the last 4 years or move to a new solution which will not yield and performance gain? Yeah let's just move to immersion because it's newer and sounds cooler.

It is amazing how dense people are - AMD kept putting up immersion litho in it's technology foils for analyst day. The problem is (for those who have a clue) the features that AMD will print on 45nm with immersion will be about the same as Intel's dry litho process - so why is immersion a technological advantage? Does it make the transistors faster? NO. Does it make the transistors lower power? NO Does it make yield better? NO

So the idiots will come in and say a double pass dry litho solution is more expensive than a single pass immersion process, no? (after all 1 <2). However:
- immersion tools are ~2X the capital cost.
- operate at slower runrates than dry tools (meaning to produce the same # of wafer outs you actually don't need 2X the dry tools as people like to think).
- Immersion requires all new tooling (Intel can potentially re-use 65nm litho equipment on 45nm depending on the ramp up ramp down rates).
- Many other intangibles - requires new set of process consumables (resists), have new potential failure/yield loss modes, requires new training/operations, spare parts,etc...these all can be relatively minor but they add up.

Newer tech != better tech... someone needs to buy a clue... I wonder why AMD has not provided a SINGLE METRIC (cost, performance, yield) to show immersion is BETTER than dry litho. In the semiconductor world Intel is renown for their ability to extend litho (and other manufacturing tooling) beyond suspected limits and 45nm is another case... In the manufacturing world this lowers risk of new tech node introductions, lowers cost and streamlines operations - new tooling is only introduced when their is a significant process performance benefit or their is a cost advantage.

Giant said...

You just can't make it up:


Look at slide 20.

They updated the presentation; they now admit it's a simulation!

*Estimated performance based on AMD lab
measurements of Quad-Core AMD Opteron™
processor at 2 GHz

AMD: Smarter Smimulated Products!

Anonymous said...

Scientia is so stupid his stupidity kills kittens.

Anonymous said...

giant wrote

You just can't make it up:


Look at slide 20.

They updated the presentation; they now admit it's a simulation!

*Estimated performance based on AMD lab
measurements of Quad-Core AMD Opteron™
processor at 2 GHz

Based on SPEC fair use rules, these are probably real numbers from SPEC runs (they have to use 'Estimated' in this case); otherwise, would they be running simulated SPEC on simulated Linux?

Axel said...

Saving this here for posterity in case Scientia edits it. I would have put it in the latest blog entry but I felt it was more relevant to this one.


Unless Intel will be able to release chips faster than 3.33Ghz in Q1 I can't see why AMD would have a problem with pricing.

Remember that with Intel's new pricing as of July 22, the fastest 1333 FSB dual core and entry level quadcore are both $266 each for trays of 1000 (though currently the actual street prices are significantly higher due to great demand). One more price cut on the Q6600 puts quadcore firmly into mainstream OEM territory.

By the time Q1 2008 rolls around and Phenom is selling in volume (we hope), Intel will likely have already launched Penryn and Kentsfield prices will already be cut once again. Q6600 and all Conroe dual cores will probably be under $200. It is likely that the 3.33 GHz Wolfdale dual cores will not exceed $266, as Intel evidently intends to exert pricing pressure to the extreme to ensure that AMD cannot make a profit on their Kuma dual cores. Kuma will have to be priced lower than Wolfdale at the same clock and here's why:

Kuma may have similar IPC to Wolfdale for non-SSE apps but will almost certainly be slower clock-for-clock in SSE due to Penryn's new Super Shuffle Engine that accelerates SSE2/SSE3 throughput, as can be seen by the gaming improvements in the HKEPC preview linked earlier by 13ringinheat. For apps re-compiled with SSE4, Penryn will simply destroy K10. From what we have seen so far, Wolfdale simply brings more value to the table than Kuma unless K10 has some hidden rabbit in the hat we don't know about.

Since AMD is unlikely to have Kuma in volume at >3 GHz until Q2 08, they are likely to be stuck with <$150 pricing for all volume dual cores for the foreseeable future. We know they cannot make a profit on that with their current expenditures.

AMD's only hope to bring their overall ASP up is with Agena (desktop quadcore K10). I'm guessing that the bulk of these 283 mm2 monsters will have to be sold at $300 to $500 for AMD to have a prayer of making a profit. Unfortunately, I don't think Intel will allow this to happen. As Penryn is likely to be as fast as Agena clock-for-clock in non-SSE, faster with SSE2/SSE3, and easily dominating with apps re-compiled for SSE4 (e.g. DivX), I believe AMD will have to price Agena lower than Penryn at the same clock. Intel will use their die size advantage to their fullest to price AMD down to ASPs that cannot support their current CAPEX / R&D spend profile.

Speaking of die size: Most of AMD's CPUs produced this past quarter have been Brisbane (126 mm2 die) though a substantial percentage were Windsor (183 mm2). AMD lost $600 MM with an overall ASP of $62 (see http://www.overclockers.com/tips01193/) on a mix of CPUs that was primarily Brisbane. What does their overall ASP need to be to have a prayer of breaking even on a 283 mm2 K10 die with an additional metal layer, let alone making a profit?

With Fab 38 pushed out to 2009 (http://www.semiconductor.net/articleXML/LN648033101.html?industryid=47298&nid=3572), AMD are maxed out on 65nm capacity for the foreseeable future. In addition, some line capacity in Fab 36 has to be reserved for those huge K10 dies and also 45nm re-tooling. Therefore AMD will not be realizing much more in the way of cost savings with 300-mm / 65nm as Fab36 is already fully converted.

No, I don't see how AMD can make money anytime in the next three quarters at least. I see bankruptcy as a very real possibility. At their Analyst Day the execs were completely forward-looking for 2009 onwards, almost ignoring the present difficulties for 2007 and 2008. This means they are likely desperately trying to find someone to keep them alive financially in the face of credit rating cuts and continuing massive losses, because they have products in the pipeline that can bring them back in the running if only someone lends them the cash they need to bring these plans to fruition.

Anonymous said...

AMD and Intel's ASP's


Ouch... $49 Desktop AMD AVERAGE, $63 mobile? $62 overall (with server factored in)

Look at the trends over the last 6 quarters while AMD has been making some market share inroads...

Intel ASP: $137 --> $118
AMD ASP: $99 --> $66

So for every $1 Intel cut, AMD cut $1.50, on a percentage basis AMD cut 33% while Intel cut ~%14%. Still think Intel started a "price war"?

Yes a very sound strategy by Ruiz et, al... rumor has it (INQ) that another round of AMD desktop cuts are coming; of course according to AMD they have stopped the price war to stabilize things in the channel (and if you believe that, they'll be a 3GHz Barcy at launch!)

Hard to make money when your and not have your brand be associated with CHEAP (uh, I mean "value", value is the new AMD buzzword right?) when average selling price is nearly 1/2 that of your competitors.

...$49 AVERAGE for a desktop chip... folks that means there is a SIGNIFICANT amount of chips being sold UNDER that! Thank you DELL, can I have another?

I wonder how much Intel gets for it's premium chipsets?

And for those AMD'roids that think Barcy is going to save AMD - take a look at the server ASP's and how much of an impact they have on the overall ASP! Maybe when the desktops come out (in volume, meaning ~Q2'08) there'll be some relief, but if AMD goes 100% K10 on server this year it sill would have minimal impact on AMD profitability (and of course there's no way AMD is getting to 100% server conversion by end of year)

The numbers are quite scary...