Guy Kewney

We've got a problem with PCs. They're too cheap. Of course, this isn't the sort of problem that you and I get too worried about, but it worries Intel and its rivals dreadfully and they have different ways of trying to fix this.

What would make you spend an extra £50 on a PC? Blue LEDs? More power? Lighter weight? The latest triumph has been - strangely - not the 64bit chip. That's a triumph which is still developing. No, the trump card for 2003 was Intel's notebook design.

By producing the Centrino package of low power, high processing speed and integrated wireless, Intel made a massive increase in its average profit per processor chip. Nice for Intel, but what did it have up its sleeve for the next trick?

The Pentium 4 M processor was what impressed me. Reversing the trend of getting faster processing rates at the price of burning more electricity, it gave a more powerful machine with battery consumption back at Pentium II levels. But Intel knows that there's not a huge amount it can do to improve this.

It thinks that 'Free Internet' will be the next magic trick. Over the next two years, Intel is going to pour a vast amount of money into a technology which, on the face of it, will not be something it can either sell or use.

It's a wireless technology, which was originally designed as a 'last mile' solution to bring broadband to people without needing to dig up the road. It's called WiMax.

The original spec said that people would have WiMax receivers in their homes which would feed high-speed data into a wired or wireless local area network. It would work just like a cable modem or ADSL, but without the cable; you'd have a router in your study and you'd plug things into it.

What Intel has done is to urge the WiMax group to add a new feature to it. Instead of being a box on a mast talking to a lot of antennae on the corner of house roofs, the next generation of WiMax will be built into the motherboards of mobile devices. You'll be able to get internet, fast internet, on the move.

So, why am I left wondering what's going on? Technically, nobody doubts that mobile WiMax is possible. The standard is still (at the time of writing) waiting for final ratification but, again, this isn't in doubt.

The cost of the technology is a bit of a poser. Today's designs would add £280 to the cost of a £330 PC and are not even remotely feasible for a mass market. But within three years, according to Intel, the cost will be down to below £28. And within a year of that, perhaps ... below £8.

Here's my problem: I don't see what's in this for Intel. Intel makes silicon; specifically, it makes processors. WiMax is a standard, which anybody can meet and which Intel won't have an exclusive on.

There's no obvious reason to suppose that if you bought a WiMax computer, it would have an Intel WiMax chip built into it, nor even that Intel's chip would be necessarily better, more powerful or use less electricity than anybody else's.

So why is Intel driving this standard so hard? And trust me, Intel really is driving it. It's been lobbying other people, setting up infrastructure outfits, such as Roampoint, ostensibly in the Wi-Fi market, but which would be able to operate in a WiMax market too and I don't doubt that's what it is intending to do, one day.

It obviously hopes to make money out of it, but how? I think there are two coins which Intel hopes to get paid in. First, the obvious one: it thinks that people will buy more notebook PCs if they have permanent high-speed internet.

Quite correctly, Intel looks at the world's telcos and regards them as dragging their feet. In a truly competitive market such as Tokyo, you can pay $25 (£14) a month for a fibre link providing 10Mbps, while you can try to pay nearly twice that in London for a 500Kbps DSL link and be turned down because you are too far from the exchange.

The technology isn't more expensive in London: BT just sees an opportunity to milk the old copper a bit longer. And Intel wants to stick a pin in the idle backsides of BT and its fellow telcos around the world. 'If you don't get your act together, we'll force the pace,' is the message.

But there's another message too: Intel wants to have some game tokens. The game is 'intellectual property' and it's worth money. All the big processor silicon designers share IP with each other and get access to each other's patents.

And I suspect that Intel wants a big coin to spend, so that it can force AMD to part with some of its 64bit X86 technology. Intel thinks WiMax is going to be a lot of big coins, which AMD isn't collecting.

But one thing I'm pretty sure about: Intel won't be the world's biggest WiMax producer. That's not what is going on, as you'll see over the next two years.

Like this story? Spread the news by clicking below:

 del.icio.us     Digg this     reddit!