These are unexpected results. When writing an operating system from scratch, programmers should be able to build on the knowledge of the past to eliminate inefficiencies and logjams in t! he code, and to optimize software performance. Granted, there is an inevitable overhead associated with increased security features, but some of these features derived from defects in the earlier operating system, which should have been corrected in the new version. For instance, there is no excuse for Vista’s disk defragmenter being apparently slower than XP’s while maxing out a 1.6 GHz cpu, unless the optimization of file placement and consolidation is vastly superior, which is not evident from performance. In short, Vista runs slowly on a machine where XP would run well.
Therefore, people wishing to run Vista effectively will have no choice but to replace their computers, or at least to upgrade their cpus, to the fastest they can afford. This is good news for INTC. I don’t believe the news is as good for Applied Micro Devices (NASDAQ:AMD), because while AMD’s cpu price points may be more attractive, the name of the game for Vista will be raw computing power, followed closely by wattage per cpu, since faster cpus may reduce laptop battery hours. INTC appears to be in a slight lead on both counts, with its 45 nm chips scheduled for production earlier than AMD.
Historically, people have eventually migrated to new Windows operating systems. Few continue to run MSDOS or Windows 95, although those operating systems remain quite acceptable for many applications. In fact, Wordperfect 6 under DOS with special video drivers (with the look and feel of MS Word) on a 386 was, in my view, as fast as Word for Windows running under Windows 95 on a more expensive Pentium machine. But sooner or later, some useful feature comes along that makes people upgrade, albeit grudgingly, to the new operating system. Or else it can be the "herd" mentality or a system manager’s demand for uniformity among network clients.
The fact that Vista makes most of the PCs on the planet obsolete (in terms of hardware capacity) brings the market to a turning point. Many users who are loyal to Microsoft will start to look at Linux as a serious contender. I don’t think the motivation is strong enough yet to swing much business away from Vista, but some users will indeed choose this time to migrate. They will likely get much more "bang for the buck" when computing under Linux. It is only the low cost of hardware that will stop many more from either migrating to Linux or else staying with XP until support for it stops in a few years (at which time Linux may be a more viable choice).
My take on the situation is that Intel (NASDAQ:INTC) will get a huge surge in business in CY2007 Q4, as their faster chips come to market and as people realize that buying the kids a PC for the holidays won’t make them smile if it takes forever to boot up or to run their games. I expect INTC to partner with graphics companies on motherboards and chipsets that will compete with AMD’s graphics capabilities, so I’m not too worried about AMD’s graphics taking much holiday business away from from INTC. I believe that the company with the fastest (mainstream market) processors, even if they are a little more expensive, will win the holiday business this year. Unless AMD pulls a rabbit from its hat, I think the winner will be INTC. In previous years, people could live without the extra cpu power and still compute effectively for many tasks, which benefited AMD. This year, under Vista, they can’t.
It eludes me whether the increasingly ponderous operating systems produced by Microsoft are so slow because of genuine improvements or from some desire to bolster the stock prices of the chipmakers through failure to better optimize their code. It is also true, for those who haven’t thought about this, that Microsoft has created a gazillion jobs worldwide because of its arcane, operator-intensive, Tolkien-esque, property grammar-based software. Response time on a 64 user timesharing PDP-10 big iron mainframe under the TOPS-10 operating system with one meg of magnetic core memory of 36 bit words and an outboard graphics pr! ocessor was faster than VISTA. The comparison is apples and oranges, I agree, but in the modern world there is no reason why users should ever have to wait for the machine. At today’s cpu speeds, response should be nearly instantaneous for most activities. Similarly, it shouldn’t take someone with a half dozen certifications and a photographic memory for details to manage a server farm and networked clients. But Microsoft has succeeded in building software so esoteric that computer specialists, whose jobs were jeopardized in the 1960s through increased standardization on centralized mainframes, will be in demand for decades to come. A true gift to the middle class by Microsoft, whether intentional or not, and we shouldn’t forget to thank them for it. Anyway, back to the discussion.
Let’s look at the government market, both here and abroad. Vista supports strong hard drive encryption, which governments are fond of. As I understand it, this will require a chipset supporting the trusted processor module [TPM] chips available from Infineon (IFX) and others. I don’t think the market for these chips will vastly increase the IFX bottom line, because it’s a big company, but there should be some profit increase, although competition will keep the margins down.
The main thing is that the U.S. government will eventually insist that its computers have the encryption, which means a nearly universal upgrade to Vista. This in turn implies an upgrade in hardware on a grand scale.
Government employees, many of whom have no need for lightning fast machines, will want the latest and greatest as an office status symbol. Even those who are further down on the food chain will get reasonably fast machines, because IT decision makers can’t run the risk of undersizing capacity and later being berated for poor judgment if their colleagues complain that the machines are too slow.
There is always a delay in implementing new operating systems in government offices, because th! e new systems must be vetted by government computer security specialists, who "certify" that these systems are secure. That certification, in my view, is more eyewash than reality, unless the investigators have access to the source code of Vista and several years to do the testing, which doesn’t usually happen. Typically, a certain build of the operating system is taken as a baseline and subjected to testing. I have no idea whether that is just waving a wand over the PC or is actually a methodical attack on security features. Hopefully it would not be done with unpublished criteria by unidentified teams (perhaps temporary hire consultants, who can disappear afterwards and therefore not be accountable if they are wrong). Anyway, after the grumbling and gnashing of teeth are done and enough time passes so that the government agencies are the subject of derisive jeers at their slowness, somebody finally ponies up the money or the courage to implement the ! new operating system (by that time nearly obsolete, but who's counting).
Vista seems unpopular with the government for the moment. I see on the web that several agencies have banned it from their networks. No matter. It was the same way with XP a few years back. It is only a matter of time before they adopt it. For instance, all Microsoft has to do is stop supporting XP, and the government will have to migrate by default, because it can’t justify keeping an unsupported operating system on mission-critical computers.
The question is timing. If we look at the time it took for XP to replace Windows NT workstation or Windows 2000 workstation, it was about two years after release of XP. So we might expect a big surge in the latter part of 2008. Therefore, I believe that Intel has two surges in its future: a small one at the end of this year, and a much bigger one at the end of 2008. After that, things should slow down a big, because Vista upgrades are not likely to degrade their faster host systems as much as Vista degrades an XP machine now. It will probably be a similar situation to the upgrade from Windows 95 to Windows 98.
So Intel should perform well over the next two years, in my view (my own price target is 35 by the end of CY2008). Its margins may not be as large as before due to competition with AMD, but its high-end chip sales should soar. If AMD gets new top management, my prediction would be a return to the previous (fictional, of course) dance where AMD concentrates on laptops and low-end desktops, and Intel takes the server and high-end client space. If that fiction were realized, then both companies could raise prices. As long as cpu prices are in dollars and the dollar continues to weaken, foreign sales should increase rapidly as well.
It is more difficult to predict who will benefit most directly from new computer sales, because there is a great deal of competition. I think the old faithful Hewlett-Packard (NYSE:HPQ) and Dell (NASDAQ:DELL) should do well, but they have competition from many other brands. These and other mainstream American brands should start to do better in late 2008, as the government jumps on the bandwagon with contracts requiring U.S. suppliers (although many of the parts may be made overseas anyway).
We should not forget to thank Microsoft for the additional GDP all this will generate. In addition to increased revenues from hardware sales, a smoke and mirrors element of GDP is computing power. The faster the cpus which are sold with PCs, the greater the productivity factor of GDP, even if the same old programs are run on the faster computers under a slower operating system. So the economy will look better, regardless of whether it is indeed better.