Ballmer on Web 2.0 and the importance of infrastructure

I'm still reading through BusinessWeek's interview with Steve Ballmer, and still fascinated. Love him or hate him, Ballmer is a smart guy.

Ballmer refused to call the YouTube valuation ($1.65B) overvalued. He rightly said that it depends on a lot of factors (factors, incidentally, that I don't personally think add up, but I'm willing to be wrong):

You're clear as a bell on the YouTube valuation…
No. I'm not saying it is overvalued. I'm not trying to say that. It depends on a set of factors. I'm not saying I wouldn't write a check for that amount of money. I might.

Really, even though there's no identifiable business model?
It is one of those things where you have to think. You can't punt either way. If you're asking me if I would offer $1.7 billion if no one else was offering $1.6 billion, no I wouldn't do that. On the other hand, if somebody is really going to offer that amount of money, you cannot reject these things. And I'm not saying that we are.

Take Facebook. Every month and a half, there's a rumor that it's going to get sold. What was the last rumor, a billion bucks? Is it worth a billion bucks? It hasn't proven to be worth a billion bucks. But it also hasn't refuted that it might be worth a billion bucks.

Look, there are only [a few] buyers who can buy anything over a billion dollars in this space. There's us. There's eBay (EBAY). There's Yahoo. And there's Google. I don't think there's anybody else. Even Amazon (AMZN) has a market cap of only $13 billion. I don't think they're going to do too many $1 billion deals. So you've got maybe four guys who will buy in that range. And there's the public market....

What do you think about Web 2.0 valuations?
There are hot companies, and there's everything else. The question is: What's the value of an eyeball? Take Facebook. If you knew for sure that you were going to have the kind of minutes of eyeball time on the percentage of college students that Facebook has today, for the next 15 years, it's an easy billion-dollar check to write, even though it doesn't have a business model that establishes it. If [Facebook founder Mark] Zuckerberg is going to grow older, he's going to lose his hipness. The site is going to be replaced by Facebook Prime, you know, the guy will be old. Pretty soon, he'll be 25. It's not like it's so sticky. College students churn every four years anyway. If it's that, it's not worth anywhere close to a billion....

So, maybe the deal makes sense, but it seems like Ballmer is saying that it's only worth a lot if others are willing to pay a lot. Yes, that's how markets set prices, but it seems like he's following valuations, rather than setting them. Of course, he and everyone else is still trying to figure out how to value eyeballs in the Web 2.0 world. Which makes Web 2.0 sound a lot like Web 1.0 (and its associated Bubble 1.0).

More interesting in the interview is his take on infrastructure:

These are hits-based phenomena. Don't you and others have to monetize very rapidly before people shift onto something else?
If they are hits-based, which is the implication I've made here, and you pay a lot of money, you're going to have to move in, milk, and get out. Or, it may turn out that the money is in the infrastructure for these things. What is Google to MySpace? It's the advertising infrastructure. In some of these things, you'll have to decide, if you're the big guys: Do I want to play at the application level or do I merely want to play at the commerce and other platform level in a way that strengthens the rest of my assets and allows me to make money?

Which calls out a larger question: is infrastructure the big winner in Web 2.0 and in enterprise software, generally? Yes, Oracle has been buying big application companies for big bucks, and is arguably extremely well positioned to compete for every dollar spent by an enterprise, but the core of Oracle's strategy is its database. Infrastructure.

The most interesting open source play right now? Red Hat + JBoss. That platform will wield an increasing amount of clout.

It would be interesting to see how other open source infrastructure players deal with this. I can see a world where MySQL builds a constellation of applications around itself. I can even see (and perhaps am paid to see :-) an ecosystem around Alfresco, to provide an alternative to Microsoft's use of Sharepoint to serve as the hub of a new level of lock-in.

Infrastructure may matter more than applications, both in terms of Web 2.0 and in terms of open source. It is clearly the monetization engine for Web 2.0. Will it be the same for open source?