11.25.2006

Intel: From Winnie the Pooh to Red Bull

Before I start this article I wanted to clear something up. I have been receiving comments from the same people suggesting that my blog no longer talks about industry trends. This obviously isn’t true, in the last month I have written a few industry articles both on the blog (like here, here, and here) - and in various magazines. I don’t talk about HP anymore than I did Voodoo in the past – although there has been a ton of activity to speak about as of late. ...and you better believe that I’m proud of our accomplishments and I love working for HP.

You’ll no doubt hear more about it as time goes on - so for the people who choose to hide behind an anonymous name and write the same weak comment over and over again, please take the time to actually read the site. :)

Thank you.

Intel from Winnie the Pooh to Red Bull

I wrote my Thanksgiving Day (U.S.) blog over here. It seems like Intel has much to be thankful for as well this year.

If you asked me last year if I thought Intel would turn their product line around in twelve months I would have said it's highly unlikely. To me, Intel’s product line was about as interesting as a cold sore last year.

Granted their upcoming products (Conroe/Kentfield) looked interesting, but I figured they would have a tough time blowing out their existing inventory of Pentium branded processors. I was also very sure (and vocal) that Intel’s most exclusive customer would switch some of their product line to AMD in short order. I was right on that point, as were some others – and Intel continues to lose market share to this day as a result of this sudden unexpected switch.

Last year I was so down on Intel that when the stock hit $30 that I thought there was no upside in sight for at least two years.

Well, it turns out that I underestimated Intel's ability to turn their ship around so quickly. In the past I have commented on our love for their low voltage initiative, their Conroe/Kentfield and mobile processors. I absolutely love the overall platform performance that Intel has been able to deliver to the desktop. I have praised the Intel Israeli team for “carrying the company on their backs” when they were down. …and to this day that same team continues to kick-ass.

I also knew that Intel would need to change the way they did business in order to turn things around. They would have to accept the fact that their market share would drop and they would have to control margins.

What I didn’t expect was their ability to swallow their pride and accept the fact that their competition had beaten them good. I didn’t expect Intel to accept the fact that AMD would win business from their largest customer, but as it turns out it’s probably one of the best things that has happened to Intel and the industry.

It looks like Intel submitted, and that said, I am pretty sure they hit rock bottom. There is no doubt they don’t want to go back there, and it seems as if they are making the transition from Bear to Bull.

The thing that people need to remember is it took AMD over three years of being on top of the performance benchmarks in order for their brand to crack into the mainstream. I don’t think it will take as long for Intel to go from Bear to Bull though – and there is a simple explanation for this: Now that AMD is (almost) widely accepted as a mainstream alternative I think the swings will be quicker, similar to the video card competitions of Nvidia + ATI. AMD doesn’t need to “re-crack” the mainstream. AMD has already accomplished the seemingly impossible, now they just have to continue making solid reliable products.

The only problem I see is the growing price wars and sudden dropping of average selling prices. This could be very unhealthy for both companies. Especially since AMD has a giant red pill to swallow - and Intel may lose enough market share to force them to keep margins strong. Regardless I hope both companies can re-strategize quickly and prepare for the long haul.

It will be interesting to see how the swings play out over the next few years. I don't claim to be a master of finance, but the way things are going I'd say Intel is probably a screaming deal right now. Although AMD is in a very unique position because they can take advantage of swings from the graphics and CPU side...

I wonder if Wall Street is paying attention?

As an aside; I have been reading comments from people screaming about AMD "taking channel allocation away" in favor of tier one OEMs. I would say that's completely false because there are those in the channel who have zero issues procuring AMD CPUs. Although AMD's CPUs are in massive demand - I don't believe it's because of any one customer specifically. I believe it's because their notebook CPU demand is out of control. AMD did not expect their Turion processors to be consumed so aggressively, and as such I believe their notebook production is taking up an unexpected amount of their capacity. I wrote an interesting article on the current situation here.

20 comments:

dino said...

Wow, this is an amazing blog. Well written and insightful. This is going into my Google Reader right now. :)

JP said...

Holy *@#! I had no idea Intel was doing that well. What will Sharikou do? I thought they were going bankrupt according to him!

Anonymous said...

Wow, Rahul actually admitting Intel is back on top. For being so "tech savy", Rahul fails to mention that AMD is at least 6 months behind Intel on their quad core and that Intel is diversifying into other fields. Nor does Rahul admit that Intel has lower operating expenses based on the "core" design and can price more aggressively than AMD. Rahul also fails to mention that Intel has shown off an 80 core demo chip. Sounds like Rahul made the wrong bet. So why did you sell Voodoo to HP?

Rahul Sood said...

"Anonymous", have you ever read this blog and checked the dates?

I think I started talking high on AMD when they were trading at $15... They went to $42 at some point I think didn't they? I am still convinced that AMD has incredible long term potential with ATI.

I was not happy with Intel when they were at $30+ - I think they fell to $18 at some point didn't they?

You should really read the blog before you make assumptions. Your final comment on HP is laughable, I guess you're missing their performance as well.

www.stockhouse.com - read it sometime.

mc said...

Wow, Rahul actually admitting Intel is back on top.

Rahul where do you find guys like this? What an idiot, it's clear this guy have never read a stock chart before. Personally I have been reading your blog for over a year and I have always viewed your comments as well ahead of the curve. Keep up the good work!

Anonymous said...

"Intel is diversifying into other fields."

Oh, I guess whoever wrote that dumb comment missed AMD bought ATI.

Raul I love your blog, but where do you find these dumbass readers?

Rahul Sood said...

Well, this industry is cyclic, and as you'll see there are swings one way or the other in many areas of the business. In this case, AMD was winning in price/performance for years before cracking into the mainstream. Now that they're (almost) widely accepted I think we'll see much broader swings.

Also if you read the blog you'll see that there are many great (positive) articles leaning towards Intel in many areas. A look at the most recent Kentsfield article talks about how awesome we think it is...

The bottom line is there is no way that one company will remain King of the Hill in all areas of the market forever - it's simply not possible.

So this blog does attract people from all walks of life - and there are fanboys on either side of the equation.

I've said it before and I'll say it again, I'm no fanboy, I'm just a fan of great companies and great technology.

Anonymous said...

Does multicore really help that much? From what I've been reading, most commonly used applications (including games) are naturally single-threaded, and an extra core mainly just picks up the occasional extra load and makes everything run smoother.

If this is truly the way it is, I see a longer life to 2 core cpus than many would think. Four cores may end up used primarily for special apps and "ultimate" enthusiast computers. Cpus with just two cores could turn out to be the mainstay for the desktop and laptop for some time to come.

Contrary to what the press has been saying for the last few years, the new "juice" in computing may not be having more than two general purpose cores. More likely better coprocessing is the ticket. This would improve both speed and energy efficiency.

Since the acquisition of ATI, AMD has a leg up on coprocessing with their Torrenza initiative, their stream processor, CTM, Fusion, and now cutting the gpu into uniform linked units so it becomes directly scalable (like the AMD multicore cpus).

Future AMD desktops would be mainly 2 core with coprocessing.

Future AMD budget laptops and desktops would sport 1-2 cores with an embedded graphics unit.

Future AMD scientific users and servers might have 4 core with coprocessing or 4x4.

If this is indeed the future trend of the market, I would say from what I read many fans and investors don't get it yet. People are planning on 4 cores for their desktop when maybe that isn't really going to help them too much, and they'd be better off for now with 2 cores, abundant memory, and a top gpu.

It will be interesting to see where the computer industry is at by next fall.

Arjan Oudkerk said...

To be honest I find this old news.

The whole swing Intel would make was already visible in May, when benchmarks of the first prototypes of the Core 2 duo were far better then AMD's current offer and even these prototypes showed great overclockability: suggesting an enourmous additional performance potential. I bought Intel stocks then for $18 and have a 20% ROI already in pocket now.

Next to that Intel has at least a 1/2 to 1 year lead in production-proces: Intel is working on transforming to 4.5 while AMD still has to go to 6.5. Samething on the quad-core processors front. Over a little more then 1 year (2008) Intel will even release eight-core processor.

If you know the cost increase in server size: 2-way to 4-way to 8-way to 32-way and 64-way. You can imagine how a 2-way quadcore can compete price wise against a 4-way dualcore or a 8-way quadcore against a 32-way single core or a 32-way quadcore against a supercomputer. Intel will revolutionalize the sever markt as well with their multi-core solutions.

DR said...

1) AMD should jump out ahead w/ respect to quad core – INTC has same sort of problem now w/ quad core for them means kluging together 2 dual cores and a lot of performance and efficiency is lost. AMD will have native quad core

2) more importantly w/ respect to Intel’s stock – well that is a function of how much Intel can earn and what is an appropriate multiple. And there is more to do with INTC earnings power than how good Woodcrest and Conroe are. For example, after the introduction of both of those parts, Intel still lost share to AMD last quarter for all sorts of other reasons. And with respect to Conroe, we found out people were not willing to pay for it and instead just valued lowered pricing. The problem is in the MPU space, in general mainstream just not huge demand or willingness to pay a premium for all the processing power. It doesn’s get the avg corporate user or consumer that much. The GPU business is diffferent – there is a ton of value to a better GPU (and in high end machines the better processor matters too along with the GPU)….but for the avg guy – well surfing the web and running quicken, etc the CPU doesn’t move the needle. Also most PC unit growth now is emerging markets (US negative growth now) – all they want is price.

So I agree conroe and woodcrest are great parts. However, I still think Intel is a Very pricey stock realtive to how much it will earn over time now. The game has permanently changed. Intel is no longer the monopoly it once was and will no longer be able to earn those returns – but the stock is still priced like it can earn monopoly economics.

Anyway hope u had a great thx giving and hope you enjoying experience w/ HPQ so far – u def went with the right buyer!!!!!

IT Kitty Cat said...

Rahul,

I have written a blog about Voodoo sell out to HP. you can read it at itkitty.blogspot.com

However, I just want to let you know that you have been off for the last year in predicting the AMD vs Intel fight for market share. You were right on target on many predictions. But I want to point couple of market readings you have done which were way off:

1. You predicted that Intel will not have a stragegy of pricing the core duo and core II duo to be below AMD's offering.. Boy, that turned out to be false

2. Your pridiction that AMD will crack up the Dell nut... Well, it was so obvious, even the self proclaimed Dr. Sharikou claimed that he predicted that, too. You also had an inside information as you were talking to Michael.. SO that was not a reading of market.. It was a reading of what Michael would do with your advice..

3. Voodoo's sell to HP is not a strategy of vision, it is a strategy of fear.. You ran out scared that Dell would put out endless resources behind Alienware that may would have hurt Voodoo very seriously.. But only time will tell if HP had the same enthusiasm you have for the deal. I predict that you may have the same fate of Steve Jobs in the 80's..

4. You were off on your guessing AMD's open platform strategy is the way to go. AMD proved that when they abandoned that strategy by buying ATI.. They wanted to become a "platform company". While I have at most respect for AMD's strategy and recent gains in the market, I do not think that it was about GPU's only.

On the positive side, I really respect your blog and really do enjoy it more than any other blog. But I could not help noticing you keep talking about HP and the integration.. You can tell by the number of people who commented that you are losing audience. We like the good old Rahul..

I will read you blog as long as you write.. If any article is about HP and the integration, I will read only to the point where it mentions "HP" and then stop..

MEOW

Anonymous said...

An earlier poster said:

You [Rahul] were off on your guessing AMD's open platform strategy is the way to go. AMD proved that when they abandoned that strategy by buying ATI. They wanted to become a "platform company".

Is AMD being any less open with their platform? No.

Now that ATI has been acquired AMD can provide a unified platform to customers who need that. But open platform is still the way to go, and I believe that's where the bulk of AMD is headed.

I think AMD intends to apply ATI's tech in ways that fill holes in the product line and expand it. I think they have shown they are willing to work with anybody, and they aim mainly to complement, rather than supplant, their partners.

When rumors of the ATI buyout emerged, I discounted them because I couldn't see the sense in it. Rahul's blog helped me understand there might be something about it that I was missing.

AMD didn't need to buy ATI to provide products that were already out there, or to cut loose from former allies. They got ATI so they could open up new ground and better compete with Intel.

Greg said...

Kitty, I don't see how offering platforms as a way of packaging your products makes your platform as a whole less open. AMD and ATI can now be sold together under a certifiable platform, so that the problems any specific machine built on that platform will have are limited to those between AMD and ATI products.

On the pricing issue, Intel did price core products for more than AMD products, so that the performance/price of the two companies offerings was pretty much equal. The fx processors are, and will always be an anomoly. Also, we saw a time when the channel was selling AMD chips for as much as Intel's, because they had yet to implement AMD's pricing scheme (probably to get rid of all the product that Intel was flooding the channel with).

I think multi-core will eventually level out with quad or 8 core, though maybe 16 core designs will make it through. Intel yet again shows its inability to see the temporary trends in the market by showing off their 80 core processor without showing anything of CSI or their embedded memory controller. While neither of these is absolutely essential to Intel now, it's pretty obvious that these will save them lots when they do become a problem as coprocessors become the next big thing, even in home-systems.

Intel has always had the better process in terms of individual transistor performance, however, due to heating and yield concerns, they are unable to capitalize as well on this performance as AMD is able to capitalize on its individual transistor performance, even though it's fairly lackluster. This is due to SOI and their consistent transistor upgrading and implementation. Intel is still on the path of "lets fix it when it's broken," as their consistent use of marketeers as CEOs instead of engineers shows.

Another thing to note is Intel's recent threatening of Via with licensing closure on all of their Intel chipsets if they do not exit the CPU market immediately. This is another example of Intel not realizing the problems they have reaching certain markets are due to the actual products they sell, and not the marketing there of. Via has embedded and extremely lightweight htpc systems all to itself due to its extremely low power consumption and awesome chipset features (for that line of products anyway). Intel can't compete in markets that product addresses, because all they've given those interested in those products are pentium ms and 4s, which are either expensive or hot, use too much power, and have far more performance than any of that audience needs. Geode would be the only acceptable solution left, and AMD's offering of this to the general consumer probably shows they're assuming they'll have the monopoly when it comes to these products soon.

IT Kitty Cat said...

"Kitty, I don't see how offering platforms as a way of packaging your products makes your platform as a whole less open. AMD and ATI can now be sold together under a certifiable platform, so that the problems any specific machine built on that platform will have are limited to those between AMD and ATI products."

When you sell a sticker to certify that your package works best and you guarantee that, that is closing your platform.. Because in a sense, you are saying that we only guarantee our platform to work best but not other platforms.. This is the whole thing behind platformization and branding..

One more point is, AMD critisized Intel for doing that couple of years ago, yet it is following the same path. Rahul and others pointed out that this was one difference between AMD and Intel (open vs close platform). I am pointing out that this is not true anymore.

One more predicion: I expect that future competition is going to be about Platforms and not about CPUs.. CPU competition is dead!

I like this discussion becuase it is pretty mature and insightful. Bloggers on this site are very intelligent. That is what I always expect from this blog..

Anonymous said...

it kitty cat wrote:

"One more predicion: I expect that future competition is going to be about Platforms and not about CPUs. CPU competition is dead!"

While I can kind of agree with that, I'm reminded of the knight in the Monty Python movie who takes serious wounds yet says, "I'm not dead yet!" :o)

AMD's new cpu tech is due out sometime late this summer, isn't it? Unless Intel does something to move ahead, the tech will be at parity again, but with AMD's traditional advantages on memory and multicore. After that, all I can see is die shrinks and AMD using their new cache memory tech. Eventually Intel will implement AMD's tech and the chips may look very much alike.

Of course, there will probably be some new wrinkles by then, like some new processes developed by Intel and IBM that improve performance.

Still, I see use of the central processor stalling out at approximately 2 cores, maybe 3 if such a beast is ever made. After that, coprocessing would be key.

Perhaps some people would refer to coprocessing as "platform" over processor. If so, I would be inclined to say AMD will stay competitive on the high end because its platform is open and can incorporate specialty chips from the likes of Sun, IBM, and possibly Nvidia. They will be competitive on the low end because with ATI they can provide a complete plaform that is tuned to customers' needs (budget, low energy, etc.).

Does Intel have any technology coming that is going to give them an obvious platform advantage?

Zora said...

rahul, what do i need to start working at voodoo. i'm indian. my family consists of a bunch of materialistic doctors averaging half-a-mill a year salaries... and i've been sent to india to become one..(help... -.-).. any input whatsoever will be appreciated.

eznelix@yahoo.com and i wanna buy a voodoo

TheKhalif said...

When you sell a sticker to certify that your package works best and you guarantee that, that is closing your platform.. Because in a sense, you are saying that we only guarantee our platform to work best but not other platforms.. This is the whole thing behind platformization and branding..



I think you're looking at this the wrong way. To understand AMD/ATi you should look at ATi before the purchase or nVidia at anytime.

ANy compnay can provide a more solid "reference" platform if they make all of the parts. it doesn't mean their touting the platform as the be all and end all but as a general guide to how the parts should work together.

nVidia releases reference boards for every GPU. Then the Sapphire's and Asus' can start with a finished platform and tweak it to ad or subtract what isn't needed.

I am convinced that the deal with Asus and the 4x4 is the begining of AMD releasing a reference platform for every new chip.

They are also doing this with the recently un-cancelled PIC. rather than providing a static setup they provide specs and companies decide where on the "perf totem pole" they want their product to be.

It is much easier to receive a reference mobo and duplicate than to start with a chipset.

AMDs strategy is really impressive. They understand how to work with people instead of competing with them.

I even said months ago that Intel Core 2 pricing would affect the industry negatively and sure enough certain analysts are reporting lower ASPs through the industry.


And another HUGE advantage MAD has is that the price "war" started before 65nm was released. Now when they crank up Fab 36 they will get almost 4 times the chips from a wafer.

Thsi means that Intel will NEVER be able to undercut AMD again and if they plan their releases and inventory correctly by Q2, they can undercut Intel. Becuase of the low price Core 2 was intro'd at they cannot raise revenue from it until 45nm when the cost goes down. And because NetBurst will probably last into 2008, their ASPs will be minimalized by the Celeron and D-Series

Anyway, great post. keep up the good work.

IT Kitty Cat said...

Great discussion..
If you watch the recent trends lately, you will notice AMD is talking more about platforms than processors. They are talking about Torrenza, platform for the living room, digital health platforms.. etc. I think the move to buy ATI was a move to compete with centrino as ATI provided the best Mobo for Intel mobile (cut throwt strategy).

AMD will always make good products. But when I look at what Intel is doing, I think its strategy is getting sharper! Take WiMAX for example, Intel is pushing it so hard and it is preparing itself to introduce products that take advantage of this technology first. WiMAX technology brings you the next billion customer.
Take vPro as another ex; IT shops who saw the potential of this platform were wow_ed by it.

From Rahul's recent blogs, Intel seems to have learned a big lesson about customer/vendor service..

Now step back and look at the big picture. It seems Intel is plotting a big coupe!

AMD will always make a great product. They are realizing what Intel is doing.. The change of their tone from last year to this year and their recent moves are evidence that Intel strategy is working.

And as the inquirer has been reporting, Intel is not standing still when it comes to the future of GPU!

Anonymous said...

Rahul,

Why are you filtering out posts (in the appropriate places) that may not have the best opinion about HP's products?

Is this not an open blog anymore?

Just let us know so that we don't waste our time coming here any more if that's the case.

Thanks.

Anonymous said...

What a farce 4x4 was! Zdnet calls it frankenstein in their article "AMD Quad FX slaughtered by single Intel CPU". Its power consumption is off the charts (double!), coupled by the fact that it did poorly against Kentsfield - especially in gaming. Now we know why you chose Kentsfield. Good decision!