Technologies may come and go, but technology powerhouses can be much more durable. IBM, the champion of mainframe computing, is still around today and generating lots of profits for its investors. The mainframe also rans mostly did not make successful transitions to new technologies and business models.
The last three years have been rough for investors in the dominant computing technology, the PC. The funny thing is that PCs are more powerful than the mainframes of the 1980's, and many of them are not really personal; they are the servers of datacenters, internet web farms, and local networks.
Companies that dominate the end PC market like Dell and HP are dependent on the makers of the semiconductor chips that are the building blocks of the PC. Some components have multiple sources, but when it comes to the core, the Central Processing Unit (CPU, or just "processor"), there are only two AMD and Intel (symbol: INTC). In the graphics department at the high end we have only AMD (symbol: AMD) and NVIDIA (symbol: NVDA), but Intel competes in the low end of the market where graphics are integrated into motherboards.
Six years ago the picture looked remarkably like it does today, at least superficially. Intel dominated the CPU market. AMD made CPU's, but not profits, largely using technology licensed from Intel and allowed just enough market share by Intel to keep Intel from be accused of monopolizing the market. NVIDIA led the discrete graphics chip market, with ATI sometimes having substantial market share, but basically always in the number 2 position.
AMD is responsible for shaking up that picture two times in the past 6 years. Realizing that Intel's roadmap for future processors was not what PC makers really wanted, AMD came out with its Opteron server chip design, and with the Athlon equivalent for PCs. The Opteron and Athlon introduced 64-bit (as opposed the the prior standard 32 bit) computing to mass markets, integrated the memory controller on the chips, and were designed to save energy (Intel's roadmap was basically use the old design, keep running it faster and hotter). Despite a great deal of resistance for a complacent PC industry, and some hardball tactics Intel used to minimize the erosion of its market share, suddenly AMD looked like it might actually eventually overtake Intel. But Intel had a lot of money, lowered its prices, and even changed its chip design, then used its marketing power to push back at AMD.
AMD also bought ATI. They paid about $5 billion for ATI, more than the value of AMD today. NVIDIA and Intel both executed well at the same time, and AMD looked to be in risk of bankruptcy.
NVIDIA is also shaking things up with its CUDA technology. Let's just say it uses the graphics chip for general computing, and that endangers both CPU makers, Intel and AMD.
The second quarter of 2008 was one of the most interesting of this decade, and Q3 and Q4 are looking to be just as interesting. Both NVIDIA and Intel screwed up, and AMD's plan to sell better designed CPUs, graphics chips, and chip sets (the "glue" chips that go on the computer motherboard) seems to be getting some traction.
You can get a very good idea of the financial states of NVIDIA, AMD, and Intel, as well as hearing what their respective managements have to say about their current and upcoming technologies, by reading my summaries of their Q2 analyst conferences [See NVIDIA Q2 2008 analyst conference summary; AMD Q2 2008 analyst conference summary; Intel Q2 2008 analyst conference summary]. Note that NVIDIA's quarter ended July 27, 2008, and is their fiscal Q2 2009.
NVIDIA's disaster month was July. It was revealed that NVIDIA chips were melting down, particularly in certain Dell and Apple notebooks (Apple, as usual, denies that its products have any problems). NVIDIA took a $196 million charge to deal with the expected costs of replacing or fixing the notebooks. That is bad, but it is the kind of thing that happens from time to time, and should not affect NVIDIA very much unless the problem gets repeated.
Worse for NVIDIA, they had to drastically lower their prices in July. After about 3 years of offering no competitive threat to NVIDIA, ATI (now a division of AMD) had brought out some great graphics chips and prices them very nicely.
Which does not mean AMD is back in the money. Intel's primary weapon against Intel in 2006 was lowering prices to the point that AMD could not match. It is a monopoly tactic that is tried and true. Force out the competition with ruinous pricing, then jack prices back up to highly profitable levels when the competition is done in. The only difference with Intel is that if they had actually forced AMD out of business, they would then clearly be a monopoly. So they just pushed AMD to the brink, gained back some lost market share, and then raised prices again when they introduced their Core Duo technology.
NVIDIA is also far more profitable than AMD, but they have nowhere near the muscle of Intel. They have not yet tried to chase ATI/AMD out of the market with pricing muscle. Instead they have a long tradition of competing (and winning) with better technology. So the price cuts were necessary to their own survival, and will hurt their financial position in the short run more than they hurt AMD.
AMD stock is practially free these days. While AMD has a respite, it is going to have to struggle just to stay alive. It has not made a profit, even on a non-GAAP basis, in years. Its new notebook technology (a combination of CPUs and graphics chips) is very competitive and has had some good design wins, but Intel's (which suffered a delay in introductions) is certainly competitive. AMD is finally producing quad-core Opterons that are being lauded by end users, but they came out over a year late, giving Intel time to up the ante.
I own AMD stock, and it has been the worst-performing stock in my portfolio. Hope springs eternal, so I bought more AMD recently. NVIDIA stock is also dirt cheap; I am keeping an eyeball on it. Intel is a safe bet, but I don't feel it is a great bargain at the current price.