I was having dinner with Michael Dell a couple of years ago, and was trying to remember how many PC companies there were. 'There has to be at least 50,' I said. And Michael replied: 'No, there are 500.'

The way our industry began is like a little company such as Ariba saying: 'Gee, I'm going to work on this purchasing problem, so I'm going to automate the purchasing department.'

That's like someone saying: 'I'm going to build a railroad system in Vermont.' That's great. Vermont needs

a railroad system. And then New Hampshire should have another railroad system, and that's what's been going on in our industry.

The problem is that people really don't want a separate system for purchasing, human resources (HR) or accounting.

So how do you join these pieces together? You hire a company such as Accenture or IBM. But it's expensive.

No industry remains in this fragmented form. The railroad industry didn't, the automobile industry didn't, nor will the computer industry.

There'll be far fewer companies. For example, there aren't that many auto companies in the world. There doesn't need to be, nor will there be thousands of software companies.

The good news is, this is just the dawn of the information age. This is just the beginning.

People who have the skills to install these systems, and help businesses achieve a good return on their IT investment, where they get real information and real complete automation, will always be in demand.

What's fascinating about the railroads is that after the bubble burst and it consolidated, prices started to fall, and users received huge benefits because they could travel from coast to coast.

I made a comment recently, and Netscape founder Marc Andreessen responded to it.

I like Marc a lot, and I tease him, too, because he's young and I'm not, but Andreessen said: 'Larry has to be wrong because innovation comes from companies that are two years old, populated by 19-year-olds.'

Of course, Marc believes that because he's young, and because his experience in the industry was a few years in Silicon Valley and Netscape.

But, in fact, the greatest contribution or invention came from a little entrepreneurial company called IBM 40 years ago.

It's a long time ago, it was the beginning of my career. I was in college. I'm embarrassed just to admit that. So what's happened in enterprise systems over the last 40 years since IBM invented the mainframe?

Well, for the last 40 years, it's been a quest to build bigger and bigger mainframe computers, and bigger and bigger server computers.

Forty years of chasing the same dream - having the fastest computer in the world.

The first job I had in Silicon Valley was working for a company called Amdahl, and Gene Amdahl happened to be the guy, when he was at IBM, who invented that 360 mainframe. He headed the group, the IBM researchers, who built that first 360 mainframe.

Then, he went off on his own and created a startup, which was pretty unusual back in the 1960s. He created a startup in Silicon Valley, leaving his secure job in IBM, and his dream was to build faster mainframes than IBM. And so the race began. Building faster and bigger machines.

For 40 years my industry has been involved in this race... and it's very difficult for Silicon Valley to adjust to the notion that they won't be forever young.