Monday, November 17, 2008
Analysts and press constantly talk about a serious slowdown in the PC business and hardware spending. Have we come to a point where a major paradigm shift in technology and strategy will be the only way to drive continued growth for a company like Intel?
The PC industry slowdown is really a function of the economy. It’s not because the long-term demand for PCs is waning. If you look at what’s happening with cell phones or even the telecom industry, there is a slowdown across the board. The same analysts who say the PC market is saturated, dead and gone, are the same ones who, six months ago, said “go buy telecom stocks!” How many of those analysts today have sell orders? The same analysts that were clueless about the telecom sector are now clueless about what’s happening in the PC industry.
What market forces will drive growth?
We can look at three underlying trends that are part of the bedrock of the semiconductor business. The first trend that is shaping every company in the world is B2B e-commerce. In 2000, Intel did about $24 billion in sales over the Internet. We are implementing our backend e-commerce infrastructure. In between, we are working to get our many business units to be interrelated. We’re doing all of this because real cost savings can be achieved by going to an e-business infrastructure. And we’re not the only ones doing it.
So why is this important as a bedrock for semiconductors? Because the infrastructure that needs to be deployed for widespread e-business hasn’t been deployed yet. You will need new servers and high performance clients. This is a fundamental trend that will drive technology investment.
The second trend is the move to the universal IP network — from the current combination of voice, IP and PSTN. On the universal IP network, the backbone that serves the traffic will be comprised of standard industry high volume servers, standard industry high volume network routers and switches and the servers will be running soft switches. It becomes a cost play, and a return on investment play.
On the client side, as you go to a universal IP network, there is a tremendous opportunity. Cell phones will evolve, for example, because wireless goes over a similar IP network. There will be a new demand for “smart” mobile devices. I usually carry around a RIM Blackberry pager, and what you will see is that those kinds of pagers will get subsumed into the standard high volume computational devices, like notebooks. It can’t happen today, because given that the wireless network today is fragmented you’d have to build a wireless notebook for every geographic location you were traveling in. Once you get a universal IP network and wireless standards in place it makes more sense to integrate wireless capability into your notebook. And that’s not necessarily bad news for RIM, because they make their money selling their service. This sort of application will drive demand as well.
The third trend that we see is consumer electronics going digital. Five years ago you had digital still cameras that sold for upwards of $800. Today you can get that same capability for about $100, and 50 percent of people using these digital cameras are sharing the images with friends over the Internet. The same sort of thing can be seen with digital wearable music players.
These digital devices extend the usage of the PC, and make the PC a digital hub. We have done studies of people in the home environment. We found that when families went to broadband, their time with the computer was much more natural. Instead of scheduling time with the computer they used it whenever it made sense. We also found that they spent more time with the computer. As you get wireless in the home environment we think that a third thing will happen. People wanted to do their computing anywhere in the house. Those appliances are yet to emerge. You will see wireless tablets that take full advantage of the PC, and will scale with the performance of the PC in the house. All of that will drive different paradigms of usage and will drive demand.
These three factors affect not only Intel, but the industry as a whole, so I’m bullish.
What about the theory that the PC market is becoming stagnated, because not enough consumers really need the higher computing power that each new model brings? What does that mean for the quest to build a faster Pentium?
The PC market today is north of 200 million units per year — new units and replacement units. If you look at replacement units and where the Pentium 4 is targeted, it’s not targeted at the individual who bought a Pentium 3 PC a year ago. It’s targeted at an individual who may have bought a Pentium 2 266 MHz three or four years ago, or even a Pentium 2 450 MHz two years ago. If you look at the performance increase from a Pentium 2 450 MHz to the Pentium 4 1.7 GHz you get something like a 400 percent performance increase just doing Web browsing.
As the market evolves, we will offer powerful IA32 client-based processors for the home digital hub. All of these future appliances will need a digital hub, which will have to be powerful.
Think of it this way: In the corporate environment you had a mainframe. That mainframe then gave way to servers. Those servers and those mainframes are repositories, in a corporate environment, for the digital intelligence that runs the business. In the home environment, the equivalent is the digital hub, which is the PC. And the PC will still be there and continue to evolve in performance, because you will want the performance.
There will be other clients, like tablets, that will be there, but they will be complementary. All of this will drive continued growth and demand for performance in the long term.
