POST OF THE DAY
Advanced Micro Devices
State of the PC Industry

Format for Printing

Format for printing

Request Reprints

Reuse/Reprint

By duttermohl
June 13, 2002

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!

State of the PC industry- is it dead or alive?

The recent crash in Intel and AMD's share price has me thinking about the future of the PC industry as a whole. Are good times just around the corner or years away. In particular, Moore's law is continuing in full force as is the speed race it enables, but the added transistors are not adding to revenue.

I'd like to offer some (unoriginal) observations for comments by this board's astute members. Clearly, I cannot cover an entire industry in one post, so I'll limit myself to the four commonly used areas of desktop, High performance Desktop, mobile and server/enterprise.

Desktop: Fixed placement, hard-wired.

The trend in desktop computing seems to be toward lower cost, not more power. Performance is a given as a 1 Ghz chip will meet the needs of any desktop user. This is the true commodity consumer segment. At the low end, these machines are single processor, integrated graphics, econo-boxes. The home market has seen penetration of the PC into everyday life. When my mother-in-law prefers e-mail to snail-mail you know there is a sustained market. This is also the most saturated market in the US and Europe as millions of old computers remain "good enough" to browse the web at 56k. The speed race is not moving this market, although good voice recognition could change that.

The big opportunities for growth in this segment is to "wire the world" and sell really cheapo PC's to developing countries. When the base machines are $500 to $800, $50 difference in price makes or breaks the deal. As the prices drop to $300 - $400, many, many more people can afford to buy access to the WWW, and that same $50 difference is huge! Via is looking like competition in this arena.

An opportunity in the developed world is to recognize that the PC is now a home appliance and to design it with d�cor in mind. Most PCs are now accorded way too much space in a home for the number of hours they are used. They sit, occupying real estate looking very "officey" and just plain ugly. Smaller machines that can be brought out of a drawer, pulled down from a wall or left exposed but fitting into a d�cor will be the drivers for this market. Think of the desktop replacement notebooks, SFF, or desk-notes with some fancy shell. Also think of home networking with multiple machines (just as we have multiple TV's and phones. In this market, a change in style will affect sales more than a new processor speed.

Special purpose machines are likely to become more common. The X-box is a good example of an X86 device that could be followed by several other limited use devices. Video and music editing come to mind.

Many business machines fall into this category as well. The typical office worker uses their machines for word processing, e-mail, limited database access and light spreadsheet work. Total Cost of Ownership (TCO) becomes a consideration when you can save on space (make the cubes smaller), lower AC bills, and have throw-away swap-outs for bad hardware. Think phones.

AMD's strengths in this market are their low costs (Small die, and small overhead), and traditional market share. Although Intel has good brand recognition and big Ghz numbers on the P4 (easy to market), they also have inflexibility of cost (big die, big overhead), high power demands (poor SFF/fanless/desknote potential). Bianis will help Intel a lot, but that is at least a year away.

So far, T-bred is a big disappointment power-wise. Still, AMD looks to have better products for the SFF, desk-note market than Intel in the next year or so. The key to success will be to lower the break-even point on the low-end processors as price matters so much. The conversion to .13 and then to .09 will help with cost control and improve margins for whomever does it first, and the switch to Barton should help power consumption. High volumes and low margins will make cost of production key and AMD is the low cost producer.

High performance Desktop

Price/performance matters most to Gamers, workstation users, and DIY guys. This market used to be the entire PC world and now it is a small, but important, segment. Tower cases and upgrade paths to the latest and greatest will continue to dominate this area. This is the market where buyer's read reviews and a few percent improvement in performance is worth a few bucks more.

This has been AMD's best market. AMD's price/performance reputation has allowed AMD to compete against Intel's brand and market dominance. Unfortunately, Intel is back on top with the Northwood and AMD is stuck in the middle of the pack. The Clawhammer will put AMD back on top. Still, this is a market that is eroding. Intel couldn't sell enough high-end Northwoods even with no competition from AMD to sustain their historically steep price structure this quarter, so they have cut the premium pricing the at the high end. As the bottom level gets faster, that premium will continue to erode and margins will get smaller.

Notebooks:

Most notebooks today are temporally portable. The battery acts like more like a UPC as you scramble from outlet to outlet. You really cannot work on a cross-country flight (2 hours or more) or take notes in a long workshop without an extra battery. You are always monitoring battery charge and desperately seeking to top-off the charge. And once topped off, do you leave it on standby and lose precious charge or do you shut it off and face the two minutes of boot time between stops? The fact is, most notebooks spend most of their lives tethered to the wall and are really a niche in the desktop space. This is why several major manufacturers see using desktop chips in "notebooks" as "no problemo." The PDA, on the other hand, is a great concept. Hours of battery life, very portable, instant on, but useless as teats on a boar for getting real work done.

Despite the lack of good products, this is the fastest growing segment of the industry!

To move processors in this area, AMD and Intel have to focus on long battery life � (8 hours or more), and much smaller form factors. Performance is a given as a .5 to 1 Ghz chip will meet the needs of the vast majority of mobile users, but instant on and small size are key. RAM disks seem like the only option (MRAM or Flash) to make a true portable.

AMD looks good with the purchase of Alchemy. If MIPS works well with a mobile version of Windows XP, then a fully functional, battery operated, truly portable PC device might be just around the corner. AMD looks to have a winner in T-bred in the bigger T&L market if they really can squeeze power down to the 14-16 w range.

The big question in this market is margin. Mobiles are seen as a high margin product, but recent experience has suggested otherwise. Once the PIII-M was no longer a monopoly product, their high prices started to decline. The P4-M offered no particular advantage over the P4 and has been re-priced accordingly. AMD has cut prices to gain share, but has not gained much share that I can see. Still, the T-bred/Barton will be up against the P4-M for the next several quarters and should win all price/performance comparisons. With superior products, AMD will gain unit share, but the revenue gains won't be that great.

Server/Enterprise:

If there is one area where price shouldn't matter it is in the N-way space. You want work done fast, you are willing to pay for productivity improvements.

This is the most exciting area for AMD because Intel is stuck between a rock and a hard place. The rock is the poor multi-processor performance (scalability) of the P4 and the hard place is the (cough) modest acceptance of the Itanium. It seems that AMD should be able to get decent margins for it's products, and sell a lot. Still, it is not a very big market overall. Further, Intel is firmly entrenched in this market and will defend it to the death. Since they cannot defend it with superior products, they will defend it with price. Once again, the issue of margins rears its ugly head. By the time AMD has made any serious inroads into this space, the price/performance curve will be closer to the desktop and mobile segments. I think the real losers will be the few "big-iron" vendors left in the world.

The conclusion

I think the micro-processor market is headed for a long period of intense price pressure. The bottom line is Intel sells their stuff for more than AMD, and has to. Intel cannot cut prices across the board enough to deny AMD a profit without killing themselves. So if AMD has a cost advantage on comparable products it can make a profit, but it won't be much of a profit. I think the margins and ASPs of Intel and AMD are going to converge, but close to AMD levels. While this will be bad for AMD, it will be a disaster for Intel.

The real key is revenue, not margins. AMD's better products will continue to gain unit share and put more price pressure on Intel, to the point that Intel will have to cede share or go out of business. If Barton fixes the T-bred problems, and the Hammers are resounding successes, I think AMD stands a chance to make a good living as they increase market share and revenue.

But everything has to go right. This recent stumble with the T-bred has shaken my confidence somewhat. I am not ready to panic, nor do I blame management, sometimes things just happen. But I am sure that AMD is not going to be a 10 bagger in the next few years.

Dave


Become a Complete Fool
Join the best community on the web! Becoming a full member of the Fool Community is easy, takes just a minute, and is very inexpensive.