David Versus Goliath Or ARM Versus Intel
I have just read this article in the Motley Fool about David or ARM Holdings taking on Goliath or Intel.
The article has been given the title The Reason ARM Holdings Is Dominating Intel, but interestingly the web page is called Why ARM Holdings Destined Destroy Intel. This name is chosen by the blogging software, from the first title used for the article taking out the short connecting words, like is and to. Look at the title and web address of this post.
But I suppose someone in management at The Motley Fool felt it was a bit provocative.
I don’t, as I think Intel is doomed.
As we need to process and store more and more data, one of the biggest limitations will be the amount of power needed.
As an example look at the average smart phone. It may have lots of features and processing power, but this often comes at the price of a short period of operation before the phone needs recharging. Now virtually all smart phones are built around ARM chips, as the Intel chips use too much power. This power consumption of ARM chips is one of the main reasons they are dominant in portable devices like smart phones and tablets.
The vast server farms have a similar problem and I’ve read that about ten percent of the world’s electricity consumption could be used to power them in the future. So a low-power route would be to everybody’s advantage. And ARM is the low power processor, although Intel are starting to design processors that are more efficient. Intel might be able to put up a strong fight, but I believe there’s one big reason, why it won’t.
Intel is a conservative company, that sticks to a philosophy that has worked for years. And getting companies with a vast investment in a proven philosophy to change, when the writing appears on the wall is not easy! Especially, when you are selling product to equally conservative companies, who don’t want to change their philosophies.
Let’s say you are a server manufacturer like Hewlett-Packard. You know you have to go with lower powered processors to maintain your market share.
You have two choices; wait for Intel’s product to arrive or design your own low-power processors around the ARM technology and get them manufactured by any one of a number of companies.
The second approach is probably the one with the lowest risk, so you at least have to try it. If it fails, you can still go with Intel’s new low power chips.
If you don’t succeed, then it is absolutely certain that someone else will develop a low power server using ARM technology. So your traditional market will go bang anyway!
It also has the great advantage, that if you do it successfully, Intel will start to worry, that you’ll go down the same route with all the other chips you buy from them. so you may get better prices for the other chips.
If you don’t succeed, then it is absolutely certain that someone else will develop a low power server using ARM technology. So your traditional market will go bang anyway!
The main loser in either scenario is Intel.
You have to remember how dominant IBM was in the 1960s. They had a philosophy that worked well, but where are they now! They’re just a shadow of their former selves where hardware is concerned. Smaller and more innovative companies chipped away at their market.
I would also throw in a view on how hardware designers and programmers think. Basically, ninety percent are anarchists, who believe in destroy and rebuild much better, so they will always decry the architecture of companies like Intel and go for something that appeals to their dark side.
ARM was that forbidden fruit that only a few companies used. But now it is so mainstream, you can satisfy your own preferences and those of management as well.
In five or ten years time, they’ll be writing articles about how the new processor on the block is going to destroy ARM.
The Anonymous Widower 
There is a race on to provide ultra low power processors and all processor manufacturers are involved. Intel are no exception. The problem is that it is not so easy to provide low power consumption with high performance. The processor for a smart phone doesn’t need to be particularly powerful, but the processors in large servers need every bit of processing power they can get. Almost everything we do these days involves using the internet, and soon we will all be getting our TV over it. The volume of data managed by servers is already huge, and it will only increase in the future. Legislation now requires companies to keep massive amounts of data for decades and this is creating a major storage problem.
The market for ultra low power devices will continue to grow, but so will the market for high performance processors. Hopefully, one day someone will find a way of producing high performance processors that are also ultra low power.
Comment by John Wright | January 18, 2013 |
The great thing about the ARM route is that it allows distributed development. If a company has a use, they develop it all themselves using ARM technology. There is no limit to the number of teams that can be working on ARM technology in the world.
Comment by AnonW | January 18, 2013 |