"AMD Moves Away From PCs Amid Steep Losses"

[ShintaiDK]

Excavator is most likely cancelled completely. If we see steamroller products will be questionable. But the chance should be in favour that we do.

But that slide essentially means AMD withdrew from the PC market.

http://seekingalpha.com/article/934...-results-earnings-call-transcript?part=single
http://www.tomshardware.com/news/amd-ultramobile-tablet-apu-cpu,18546.html

Seems AMD is almost on a do or die within the next year.

 

[KingFatty]

I apologize if this belongs in a different thread, but has the discussion come to any conclusions as to whether AMD just represents a past that will go away, and be replaced by more powerful tablets/phones?

I mean, I can't help but think that I'd be fine with a more powerful tablet, so long as it runs what I need and I can hook up my monitors, mouse, and keyboard.

AMD missed that boat, but is there agreement about this? Or does this line of speculation belong in another thread?

 

[Idontcare]

http://seekingalpha.com/article/934...-results-earnings-call-transcript?part=single

When I read the transcript that specifically captures what Rory said, it reads to me as "we are betting the farm on Jaguar (Kabini)".

Which is pretty much exactly what we were told is where the BoD wanted AMD to go but that Dirk refused to give up on the future being based on bulldozer's microarchitecture.

 

[pelov]

I apologize if this belongs in a different thread, but has the discussion come to any conclusions as to whether AMD just represents a past that will go away, and be replaced by more powerful tablets/phones?

I mean, I can't help but think that I'd be fine with a more powerful tablet, so long as it runs what I need and I can hook up my monitors, mouse, and keyboard.

AMD missed that boat, but is there agreement about this? Or does this line of speculation belong in another thread?

No, that's quite pertinent to the discussion. If you look at Intel, with Haswell being focused on 17W and a yearly cadence on Atom, it's obvious that both players are looking at mobile rather than desktop as their bread and butter.

Intel was also late to the game, offering only their single Medfield SoC which is in 2-3 products (none of which are doing well), while AMD has no SoC to speak of and their Bobcat successor is a year late.

Which is pretty much exactly what we were told is where the BoD wanted AMD to go but that Dirk refused to give up on the future being based on bulldozer's microarchitecture.

BD was a Ruiz/Meyer baby. And to think their best selling processor was Bobcat, and it's still one of their better selling chips even to this day.

 

[ShintaiDK]

When I read the transcript that specifically captures what Rory said, it reads to me as "we are betting the farm on Jaguar (Kabini)".

Which is pretty much exactly what we were told is where the BoD wanted AMD to go but that Dirk refused to give up on the future being based on bulldozer's microarchitecture.

Thats how I read it too. If Kabini aint a massive success aka K8 style. Then its over.

 

[nismotigerwvu]

Right. Scrapped. Ancient history. Now, lets talk about actual reality and what is in use today. 😉

Yes, Knights Corner, which is completely dominating the graphics industry now right? Regardless of Knights Corner being derived from it, Larabee itself was never taken to product stage. How is this even an argument?

 

[Keysplayr]

Yes, Knights Corner, which is completely dominating the graphics industry now right? Regardless of Knights Corner being derived from it, Larabee itself was never taken to product stage. How is this even an argument?

Wow. You guys just will not look at Intel's HD series. Wonder why.

 

[exar333]

At $2-ish the upside of shorting is fairly limited. The downside is not. There's still a chance some miracle might occur. There are likely reasons the board of directors hasn't voted to just end operations and liquidate, they think there's still a possibility of yielding a return on current assets.

I use options to avoid falling into the already mentioned value trap. If I think 1000 shares are worth buying for $2,180 as a gamble on a return to viability that forces me to step back analyze, set targets and at least mull over a strategy.

In this case the my reasoning would go like this: "Ok, if they're not viable in 2 years they're definitely gone." I put a time limit on my gamble, and instead plonking down for the whole 100k shares I buy 500 bucks worth of January 2014 calls with a strike price not far from where it's trading today. If AMD pulls a rabbit out of its hat and price spikes, I still get to participate in the move. I can even short calls along the way. If AMD goes chapter 11 Ive lost at most 1/4 of what I'd have lost if I bought the stock outright.

For a risky "investment" like AMD looks today it's a much better way to optimize risk/return, IMO.

Still room to fall, and the price to enter is pretty low. I do agree though, there isn't a lot of upside anywhere.

I put in a short position in April/May when the stock slid to ~7. I cashed-out when the stock fell to 3.5 last month. I felt pretty good about cashing-in, buit definitely could have held on longer. At this point, I think the stock will have some serious ups and downs until we finally where the company is headed. For good or for ill.

 

[mrmt]

When I read the transcript that specifically captures what Rory said, it reads to me as "we are betting the farm on Jaguar (Kabini)".

At Dirk time this could have worked as they would have arrived earlier to the party, they could have lead the Win8 marketing push. Now is too late.

In 2013 I see Kabini sandwiched between 8-10W Haswell on top and ARM15/Atom 22nm on bottom. They can't compete in performance with the former and in power consumption with the later.

 

[VirtualLarry]

At Dirk time this could have worked as they would have arrived earlier to the party, they could have lead the Win8 marketing push. Now is too late.

In 2013 I see Kabini sandwiched between 8-10W Haswell on top and ARM15/Atom 22nm on bottom. They can't compete in performance with the former and in power consumption with the later.

But they have better power-consumption than the former, and better performance than the latter. Which is not a bad place to be, IMHO.

 

[ShintaiDK]

But they have better power-consumption than the former, and better performance than the latter. Which is not a bad place to be, IMHO.

That sounds more like an excuse than an actual benefit.

The segment that covers that area is very niche at best. Abit like the rerun of APUs and the 100mio$ writedown later on.

 

[Idontcare]

But they have better power-consumption than the former, and better performance than the latter. Which is not a bad place to be, IMHO.

It is a good place to be, sure, question is will it be enough (chip volume) to generate revenue necessary to sustain AMD's development pipeline for post-Kabini products. The TAM for that market segment is rather small.

 

[mrmt]

But they have better power-consumption than the former, and better performance than the latter. Which is not a bad place to be, IMHO.

For a product you are right, but for a company? Not at all. As IDC correctly pointed out there isn't such thing as niche semicon company.

It may work for a generation or two but after that you simply cannot match the money your rivals are going to put on R&D and then your product is hopelessly outdated.

R&D costs for each new generations will go up but your revenues and/or margins will be under pressure because of your competitors. There will be a point in the future that you will either make a mistake or you simply won't be able to muster enough resources to develop a new architecture. The moment this happens, you are done for.

This is more or less what happened with AMD since ATI acquisition. The moment AMD shelled out cash for ATI AMD signed its death writ. After that AMD simply didn't have the financial resources to upgrade fabs in the same pace of Intel and neither keep pace with the engineering improvements coming from Santa Clara. Instead they had to resort to milking the Phenom architecture for years and going lower and lower in the price ladder. They were cramped in a smaller market bracket and margins got under pressure by competition with Intel.

Then, they made the mistake, Bulldozer, and we all know what is happening with AMD market share wherever we look at them. But even if Bulldozer didn't backfire, or at least if they had come with something more along Phenom lines, they would simply fare a little better than with Bulldozer, but the fundamentals, that AMD simply could not match Intel schedule and qualitative improvements, wouldn't be changed.

By betting the farm in Jaguar, they are simply entrenching themselves in a niche that is too small to sustain the company in the long term. AMD stock is getting pounded because the market sees AMD having less cash flows and competing in a market with even worse margins than those they are in now. And more important, market cannot see AMD viability in the medium term.

 

[pelov]

By betting the farm in Jaguar, they are simply entrenching themselves in a niche that is too small to sustain the company in the long term. AMD stock is getting pounded because the market sees AMD having less cash flows and competing in a market with even worse margins than those they are in now. And more important, market cannot see AMD viability in the medium term.

You're assuming this will be their only product, which it won't.

I think what you're saying depends heavily on whether AMD can have a semi competitive Steamroller core in their APU and server lines. If either one of those is delayed -- which might be the case -- then, yea, I agree with you. Competing against a 14nm Broadwell with a 28nm product just isn't going to work. Keep in mind AMD is on a yearly cadence, so Steamroller is supposed to be the direct competitor to Haswell, with Kaveri coming out roughly at the same time as Haswell (assuming neither are delayed). The desktop variants aren't the money makers and haven't been for a while now.

That's just for 2013, though. Who knows what node Excavator will be using or if it will even tape out.

 

[mrmt]

I think what you're saying depends heavily on whether AMD can have a semi competitive Steamroller core in their APU and server lines. If either one of those is delayed -- which might be the case -- then, yea, I agree with you.

Yes, I wrote off Steamroller for this analysis. Excavator too.

 

[pelov]

Yes, I wrote off Steamroller for this analysis. Excavator too.

They'll matter, but less.

If AMD have a very good Jaguar core, then with the current market trends they should benefit more than they did during the netbook phase. Considering their E-series Bobcat chips are still selling well, they might make a pretty penny with these.

But IDC is 100% right. If GloFo hasn't ramped up production of 28nm by early 2013, then it doesn't matter how good their Kabini/Temash product is. With the low margins, they'll get choked on volume and the chip won't have the room to stretch its legs, so to speak.

The demand for desktop chips is diminshing, and oddly enough that benefits AMD. The biggest issue they face is that they still make products primarily around desktop/laptop with a very tiny server market share. GPU sales will progressively decrease as well, so they stand to lose money on the GPU end regardless of how good their GCN follow-up is.

 

[Agent11]

I apologize. I wasn't aware of how dire AMD's situation was. Why did they fire Dirk Meyer???

 

[pelov]

I apologize. I wasn't aware of how dire AMD's situation was. Why did they fire Dirk Meyer???

Have you not seen Bulldozer?

The question should be 'why didn't they fire him sooner?'

Meyer replaced Ruiz in 2008 and 3 years later we got bulldozer. So while Bobcat was great, he didn't pull the plug on BD when clearly he should have. Wasted far too many resources on a product that he knew was going to underperform instead of diverting them elsewhere (Llano/Bobcat)

 

[Agent11]

The nature of the business is like that though. You don't really know what yields are going to be like until you go into production.

 

[mrmt]

They'll matter, but less.

That's clear by now. If PC and conventional servers aren't a priority, then Steamroller and Excavator will matter less.

If AMD have a very good Jaguar core, then with the current market trends they should benefit more than they did during the netbook phase. Considering their E-series Bobcat chips are still selling well, they might make a pretty penny with these.

By the time Jaguar arrives, A15 ARM will be a mature reality, Haswell 10W should have arrived too and Silvermont should be around the corner. It is not like Brazos, where nobody had anything to put against it. Jaguar is simply too late to the party to make a big splash.

The demand for desktop chips is diminshing, and oddly enough that benefits AMD

Excuse me? AMD margins are in the toilet and part of the problem is that AMD has not any clout with OEMs and there is no such thing as direct retail channel on notebooks as they have in the desktops. If there is someone losing badly with the desktop demise it is AMD.

 

[mrmt]

The nature of the business is like that though. You don't really know what yields are going to be like until you go into production.

Yields are not the only Bulldozer problem. Die size is too big, which means less candidates per waffer, raising base cost for the thing, and performance per watt is really, really bad, which is a non-starter for servers. This also bites Trinity on notebooks and desktops too.

Bulldozer was the industry equivalent to Harakiri. They conceded performance per watt, raw performance and die size advantage to Intel on every single market segment Bulldozer was in.

 

[Zor Prime]

AMD isn't moving away from the desktop completely. It won't be competitive in IPC compared to Intel for the most part but instead they'll continue to try and provide value.

Even from the horses mouth they are not abandoning servers so there's always the trickle-down effect to the desktop.

Too much sensationalism regarding AMD dying. When was the last time AMD died? It's a case of "prove it." People have been saying AMD was going to die since the 1990's, which makes me chuckle when I see someone write AMD off.

The most entertaining part about that is Intel fans clamoring for it when it's definitely not in their best interest if they actually stopped and thought about the repercussions of a x86 arena without AMD.

Anyone who buys x86 machines need AMD around, unless you happen to have disposable cash.

Is AMD in trouble? Sure. AMD tends to historically be in trouble. Wake me up when something definitive actually happens.

 

[Agent11]

Without AMD there wouldnt be a brand new $140 Intel hyperthreaded dual core, that I guarantee.

 

[ShintaiDK]

Without AMD there wouldnt be a $140 Intel hyperthreaded dual core, that I guarantee.

And why not? Besides it will be 130$.

 

[mrmt]

Too much sensationalism regarding AMD dying. When was the last time AMD died? It's a case of "prove it." People have been saying AMD was going to die since the 1990's, which makes me chuckle when I see someone write AMD off

Could you please tell us when AMD was so near the bankruptcy than when they had to spin off their fabs and in a so shattered market position they are in now?

Could you tell us when AMD stock was quoted so low as it is now?