America's biggest tech companies are focused today on the problem of creating, not technologies, but platforms.

Not that there's anything wrong with that. Intel and Microsoft and Cisco all rose to prominence with platforms. The first two had "WinTel," a marriage now on the rocks (Windows works fine with AMD, Intel will make Apple chips). Cisco had the Internet platform.

These companies changed the world. But the world is a funny place, a "what have you done for me lately" place. In business, it's a "what are you going to do for me next" sort of place.

Microsoft, like it or not, has defined a platform strategy. To that extent, Windows still works. The problem is seen most visibly at Cisco and Intel. Let's tackle Cisco first.

Cisco, which defined the dumb network, and came to dominate it in the 1990s, is now trying to build smarts into its network gear. The company calls this an "applications network" strategy and the problem is that it contradicts both the company's history and the Internet's true needs.

Today, at the end of 2005, there remain in this country tens of thousands of miles of dark fiber, optical cables that aren't lit because there is no traffic for them. It's estimated that the lit cables are only working at 1% of capacity. But rather than embrace the need for more traffic (which the Internet is doing by itself), Cisco is working with its large telco partners to gain more revenue from the traffic they have, by "shaping" or "managing" (restricting) what people can do.

It's dumb. It's what Cisco's customers want. But it's not what the customers' customers want.

Intel's platform problem is somewhat different. As Intel CTO Justin Rattner told MIT's Technology Review this month, Intel is (finally) committed to a platform strategy. Rather than compete on a chip-by-chip basis against rivals like Broadcom or AMD, where it's getting hammered, Intel wants to define entirely new technology platforms, then deliver systems that can implement the future profitably.

The problem is that, at the present time, defining platforms has become a political struggle. For technologies like WiFi and WiMax to prove themselves we need more unlicensed frequencies, and we need compatibility in the wireless regulatory regimes of many different countries. Trying to win that kind of political struggle is as useless as waiting for a Bell Company to define your opportunities -- it ain't gonna happen in your corporate lifetime.

What Intel should be doing, in my view, is proving its platforms. Work within one country to define a wireless regulatory regime that would really show the world what wireless can do, in terms of delivering bandwidth where it's needed, and services based on that bandwidth. When Ghana (for instance) can have a better telecommunications infrastructure than the U.S. (which it can, with WiFi and WiMax) then the world will be forced to respond, or it will be buried.

Another way to work the problem is through the application space, specifically the AlwaysOn application space. This would start with a PC-Router, a box that can run applications which live "in the air," that don't go down at night. Medical applications, inventory applications, home management applications. There's a huge market there once someone defines the platform, and gives application developers something to work off of.

That's another way of working the same problem Intel has defined for itself. But there are other problems Intel must pursue if it's to grow, like defining new high-power platforms, and new desktop platforms, and new low-cost platforms. They've got the problem down right, it really is all about platforms.

But having a Clue, as I've told you here many times over the last decade, is entirely different from winning the game.