Since this topic came to mind after Intel's wimpy re-structure announcement, I thought I would deal with the things I think they should have done. Also, the first comment to come in from "Anonymous" is about their flash business being spun off, it seems like this is a good place to start.
There are 3 things top of mind:
1. Spin off or sell their flash business. Eventually, while flash holds great promise as a replacement for hard disks in the future, it is essentially a lower value add and hence lower margin business = commodity. While revenues in the Flash group are ~2.3 billion in 2005, they still lost 154 million. There are only two reasons I can think of that Intel has not spun off the flash biz. First, Apple has startd buying Nand from the Micron JV between Intel & STM. Since Intel seems to be investing so much into the Apple relationship, they may be keeping the flash business and hence control to ensure Steve doesn't get pissed with them. The second reason is the role of flash in the future of the PC...namely Robson technology. Both these are really not good enough reasons to retain the flash business which is dragging down profits. While 154 million is not large in the scheme of Intel's profits last year, it is this consistent annual drain on their overall profits that does start to add up...450 million from 2003-2005. There are other issues like not allowing them to stay focussed on microprocessors, lowering their return on capital and other investments, etc. All in all, I do agree with "Anonymous" who triggered this with his comment that the divestiture of this biz is coming. Probably the reason it was not announced this week is the deal is not sealed yet. I would expect this to happen w/in the next 6 months.
2. Divest the motherboard business. Intel's board business was a strategic initiative primarily to help drive the industry forward faster than the industry would move. Think of it as a "kick in the ass" to get things moving. 15 years ago this made eminent sense. But today, with a highly developed industry and their boards while recognised for stability aren't at the cutting edge of design, I really question the need to continue this business. At the very least they need to trim this business back and stop making lower end boards with lower end margins. Overall, my thought remainse, sell this baby or at least take it outside Intel so it isn't cushioned and becomes agressive and innovative. If you need to have a strategic foot in, invest in a real motherboard company and influence the industry through that.
3. The last thing I wanted to cover was communications. It's not clear to me how much garbage is still left after the Marvell/Eicon sales as well as the 600 million $ write off last year. My only comment is dump it unless it's Wi-fi or Wimax. Anything else do outside or partner. Intel is crap at growing new businesses.
The only thing I thought about and then re-considered was the graphics side of the house. In spite of the fact they're crap, Intel has no choice but to build this capability and build it well now that AMD has ATI. Even a partnership with Nvidia is not sufficient. They need to own all the principal pieces of the platform and graphics is a big one.
I think all this would have probably taken another 10k employees out of the company and increased profitability further. A good thing even at the cost of shrinking revenues. Of course, some serious axe to at their senior management structure would probably remove a small number of employees but provide yet another significant bump to profitability and hopefully their strategic and executional abilities.
Subscribe to:
Post Comments (Atom)
14 comments:
1. Flash -- The funny thing about the flash business is how quickly the market turns on a dime. Intel has been in and out of flash about 4 times in the last 12 years or so. What really surprised me with the sale to Marvell is that Intel sold the rights and intellectual property to the very successful Strata Flash it had developed. I'm not sure if it as AND, NAND, or NOR, but in the not so distant past it was bringing in a good revenue; especially with handhelds like iPAQ's and Motorola's Q. I don't think the market is taped out there.
In November 2005, Intel and Micron formed IM Flash Technologies, LLC. LINK This is/was actually quite brilliant; 2 leaders and innovators uniting and sharing the cost of 3 fabs so that their other fabs can be used for larger margin products. And let's face it... a $500 million dollar pre-order from any company like Apple really is saying that Intel's Flash isn't crap after all.
2. Motherboards -- I can't really argue one way or the other. Frankly, I don't know much about the motherboard business. What I do know is that having an in-house MoBo division probably aligns nicely to Intel's platform strategy. I don't see it going away. Maybe you can verify this, but I heard Intel is the sole supplier of MoBo's for 3 OEM's. I'm certain one is Sony but I'm unsure of the other two.
3. Communications -- Much the same way I was with MoBo's, I'm not a Communications expert. My only point of discontent is the blanket statement of dumping anything that isn't Wi-Fi or Wimax. You added that Intel is crap at growing new business. Clearly Wi-Fi and Wimax didn't bloom over night; they had to be developed and grown. Both have now or are on the cusp of having worldwide appeal and acceptance. It is with tongue in cheek that I would say Intel can grow new business and if they probably have something up their sleeve that will obsolete Wimax. I'd say cut what clear losers, but don't cut potential winning projects in development.
And finally, ATI. You'll have to correct this, but I thought I read somewhere that ATI supplied less than 1/4 of the GPU's found in Intel systems. My point is simply this; I work for a company that medium to large size. To me, ATI and NVIDIA are high-end graphic chip companies. While I would love to have a sweet video card at home, at the office, I really only need a basic one that that can handle e-mail, Office applications, and casual surfing. At work, I would rather have more memory, more storage space, and more processing capability. Sound and Video are dead last of my office needs.
You said: "They (Intel) needed to own all the principal pieces of the platform and graphics is a big one." Last week, just for fun I tore open my small form factor PC at work and you know what I found?
-- Intel CPU
-- Intel MoBo
-- Intel NIC
-- Intel GPU
There are easily 5o,ooo+ of these same computers within my company. The point being is that Intel is clearly king of business computing and it really outranks home or personal computing in terms of units sold. Does Intel need NVIDIA the way AMD needed ATI? Hell no! The business computing segment is so large that Intel makes money by selling volume. Do businesses need the latest and greatest 256MB GPU to read e-mail and run spreadsheets? Hell no!
The point is finding what your market is and selling in volume what the custom want/needs and offers the best bang for their buck. Having a platform enables a company to have more or their products combined within one case. That being said, Intel makes less money per component but increases profit by having multiple components in the same system.
Azary/Sheepshagger - great posts. My thoughts:
1. On flash, I think the Micron route was absolutely the right way for Intel. And this is something they need to do more of. Moving the flash business out from their Mobile group with a direct report to Otellini and then the Micron JV both give this business a chance of real success. On the handphone thing, what they really sold was the intellectual rights to their application processors (Hermone & Bulverde) going into smartphones. Strata Flash was not part of the deal. I read somewhere the reason they sold it was while they were making good traction, they realised that the effort/investment required to be #1 or 2 against players like Broadcom was too high. This business is different from the PC biz. Your solutions need to be highly customised with software sitting on top to meet the needs of multiple customers like Motorola, Palm, etc. The flash biz with Micron is moving ahead w/ new product announcements and I believe they have started shipping to Apple already.
2. Mobos - here's my take. The actual board is not critical to the platform. What was important to Intel was driving adoption of new technologies which the ODM industry was often reluctant to do. However, today the market is hyper-competitive and many of Intel's board competitors do far more innovation than them. However, Intel boards are known for stability. My hypothesis is they may not dump the business but opt for a JV or something. But the point is as they look for to expand other categories, running design, manufacturing, sales & marketing for this division probably consumers resources they could use better to cover more accounts and go further down-stream in places like emerging markets. At the end of the day, it's a question of strategic focus and whether boards enable that or distract from it. I think the board itself is not important to their platform strategy. It's the brain (CPU), eyes (graphics), hearing (audio) and connectivity (communications) that they care about. You don't need to be in the board business for this. With the route Intel has taken they just need to be in CPU + chipsets. I am assuming here that the margins and return on investment is much lower than CPUs which would be a good reason to trim it.
3. On the comms piece - perhaps my statement was a bit sweeping. I should clarify. From my equally limited knowledge on this aspect of Intel's biz, it appears there really are two visible pieces Intel cares about. Wi-fi and Wimax. Many pieces acquired by Barrett blindly have disapeared. Now Intel does not make money directly from their investments in R&D around wi-fi or Wimax in terms of selling communications equipment to ISPs/telcos. They make money helping bring the tech to market and then place the hardware the end user needs inside their platform and very importantly branding it which allows them to charge a nice premium. What I was suggesting was unless pieces of the comms biz align to a strategic initiative, they should rid themselves of anything that's just running w/o really having a direction.
4. Graphics - I agreed with you even in my post. They need to build this business. However, now that AMD has acquired ATI, they may challenge Intel's platform for biz with a better technology. Hence, Intel needs to significantly improve their graphics and then use it in their platforms to create more opportunities to sell and higher ASPs. Otherwise, AMD/ATI are determined and they will start to grab share here too.
How about flipping this thing on it's head? How many paths to success?
From Otellini's 4 month or so 'review' and what it seems that are going to do with cutting headcount, I propose one path to success - Bill Ford style...
Bring in a Neutron Jack and clean house.
A turn around expert is clearly needed at Mission College Blvd.
This is what I was talking about when I said I hope Intel's senior management get their "reward" in one of my earlier posts.
Check this - Otellini's Google stock is worth 24 million...more than his Intel stock. He has more incentive for Google to succeed than Intel. Isn't that a conflict of interest of some sort. Specially since they buy significant AMD products.
Link
And considering the crappy situation at Intel, shouldn't he be spending every waking moment worrying about how he's going to turn things around and not attending board meetings at Google. I mean WTF - I think I need to do a post on this!
Divest the motherboard business
I don't think that's possible anymore. With ATI effectively out of the picture on supplying chipsets (they have cut the Rx700 series from roadmaps and mobo makers are boycotting the RD600), that only leaves Intel and nVidia in the game. I can't see Intel giving the market over to nVidia, which has always been quite comfy with AMD.
I think Intel's main mistep in chipsets lately was not releasing a new high-end chipset and the delays in the 965 series. However, that was kind of unavoidable since they are transitioning chipset production from 130nm 200mm wafers to 90nm 300mm wafers. This transition should seriously lower their production costs, but it's just unlikely the transition and it's associated ramping delays had to come at Intel's most promising time (with Core 2) in years.
The bigger reason why this is unlikely is that Nehalem in 2008 will likely bring major platform changes which Intel isn't going to leave up to third parties to provide for. It's unclear whether CSI or some other form of FSB will be used, but it'll certainly be a major transition that Intel will need to be in on to assure it goes as smoothly as possible.
In regards to graphics, in theory the GMA X3000 has a great architecture. It uses unified shaders, is light on memory bandwidth, and can do HDR with AA. It definitely has advantages over nVidia type offerings and is comparable to ATI's designs.
The problem with Intel graphics has always been in the drivers though and not the hardware. For example, the reports of poor X3000 performance are not because the hardware is flawed, but because the current drivers don't support hardware T&L and hardware VS so those features are inactive. I believe hardware PS has just been activated, but tweaking a new architecture is going to take some time. Even the GMA 950 wouldn't have been that bad if Intel spent time with the drivers. ATI and nVidia drivers are dual core optimized, but I don't believe Intel's are. Dual core optimization would be very beneficial especially to the mobile GMA 950 which is coupled with Core Duo. Games are generally single threaded anyways and even if they aren't the 2nd core utilization is poor. This means that the game could be run on Core 0 and the software VS and T&L in the GMA950 could run on the core 1.
I can only hope that Intel acquired some good driver developers when they hired those fired 3DLabs employees.
Dude... seriously... Google and Intel are like comparing apples and oranges; Google doesn't produce anything. They provide a service that you and I enjoy for free and make money off of advertising which is an interesting concept in itself when you consider their own advertising is word of mouth more or less. When was the last time you saw a Google commercial or ad in a newspaper?
If you're looking for a person to hang, Otellini is not the person. He inherited the downward spiral of Intel from the Craig Barrett years of purchasing 20-30 companies and being able to glean a profit off of three or four.
During Barrett’s tenure Intel was unable to fill $1Billion worth of orders in the 4th quarter; I believe this was 2002. While AMD had great products, I think it was this error that really gave AMD their big chance because OEM's needed had orders to fill.
Otellini has to make some tough choices and everyone wants to play armchair quarterback on how Intel should be run based on stock price. Being on a board and meeting for a few hours once a quarter is hardly what I call being asleep at the wheel and not minding your own sheep.
It wasn't but 10 years ago that an 'affordable' computer was $2400+. Today you can get entire systems for $399. What your average Joe doesn't realize is how Intel is migrating to a company that enjoyed high margins to one that now makes money on higher volume with lower average selling prices.
There should be a corollary to Moore's Law that states:
"The ASP's will be inversely proportional to the number of transistors on a chip, or rather the number of chips produced from a wafer."
Hmmmmm... The Sheepshagger's Corollary. It has a nice ring, don't you think?
Sheepshagger - a few thoughts:
1. I agree - Barrett would be a big recipient of my "reward" system. But Otellini was COO for the last few years and running the Microprocessor business before that. It would have been his decision not to release a new architecture in 2001-2002 and he has to take accountability for allowing Intel to hire 20K people in less than 18 months.
2. Intel likes to present themselves as the gold standard of corporate responsibility. Andy Grove has talked about it in public interviews as something he drives. Then what kind of message does it send shareholders when the CEO is better compensated from his participation in another company than his own. And, what of the 10k employees who lose their jobs. My belief is corporate responsibility is not a set of rules, it is a set of principles & ethics driven by the board.
I think your corrollary is excellent. In fact, I believe it is exactly what is driving a fundamental shift in Otellini's thinking and hence the efficiency re-structuring.
Anonymous - I'm talking about the motherboard business. The actual business that designs, manufactures and markets Intel's line of motherboards.
What you're describing is the chipset biz which I completely agree Intel needs to retain...and ensure their technology gets better and better here. It is crucial to their platform strategy.
While Intel is very big in the board business (specially in the DIY channel), it is not rendering them any stratgic advantage. It is not enabling their platform strategy and can only be distracting shared sales & marketing resources who have to continue to ensure they hit their targets here too.
thanks...
I think the more you write, the more you think about the situation, your viewpoints will get be closer my mine--there is but one truth after all.
Sharikou, Ph. D said...
I think the more you write, the more you think about the situation, your viewpoints will get be closer my mine--there is but one truth after all.
That almost sounded like: "There can be only one" line from Highlander.
Sharikou, Ph. D said...
I think the more you write, the more you think about the situation, your viewpoints will get be closer my mine--there is but one truth after all.
But he writes so much better than you, and doesn't express himself like a rabid dog on crack, which must pain you greatly.
Sharikou:
Unfortunately (or fortunately) our POV's are very different. I believe you see this as a limited time frame battle with only one possible end - Intel's demise. I see this as a series of long drawn out games of chess with attack and counter-attack. I don't believe either company will die any time soon and my focus is to try and call the moves of both players before they happen.
Wise Investor:
Sorry this comes little late...
The 1 thing INTC mess up big time:
Not making a REAL 64 bit chip.
Apparently M$ disagrees with you - poster who said Intel needs a real 64 bit chip.
That 64 bit garbage is as tired now when Intel is all 64 as it was in 2003 when there was no 64 bit infrastructure - meaning no OS'es (anything under 95% market share doesn't count) and certainly no applications. There have been fleets of AMD64 processors that have come into service, lived their lives and been thrown into the dumpster with their 64-bit extensions never having been fired once.
The "Ready for the Future" stuff that AMD has been pushing is pure unadulterated marketing hype.
Post a Comment