It takes a hell of a lot of capital to build a utility. A hundred years ago, when Samuel Insull began making the investments that would lead to the first large-scale electric utility, Commonwealth Edison, a lot of people thought he was crazy. He first bought up all the little mom-and-pop generating stations around Chicago, then he closed them all down and replaced them with a giant central station operating mammoth new generators. He literally had to beg suppliers to build the huge turbines he needed. They couldn’t imagine anyone needing that kind of capacity. They were wrong. Within a few years, Insull was one of the richest men in America.

Google’s announcement yesterday that it would raise $4 billion in capital with a second stock offering underscores its ambition to become the world’s first great computing utility. With $3 billion already in the bank and tons of dollars being generated through its current operations, it’s going to have a massive amount of cash to invest in expanding its data centers, buying up fiber-optic network capacity, pioneering new technologies, and acquiring other companies. As leading tech analyst Mark Stahlman put it, “”There’s a lot of infrastructure that they need to build … a couple hundred thousand more servers, a bunch of dark fiber, maybe a few data centers.” As someone who believes that most computing services will ultimately be provided, in one form or another, by utilities, I think the kind of investments in centralized infrastructure that Google is making will serve customers well.

But will those vast capital outlays serve Google’s investors well? I seriously doubt it. Given Google’s stratospheric stock price, it will have to earn enormous returns on those billions of dollars of capital if investors are to make out well in the long run. It’s doubtful that Google’s current business model – selling advertising – will generate those kinds of returns. When Insull built his electric utility, he was paid directly for the electricity he produced – and because companies and consumers always needed power, the revenue streams were protected from the vagaries of the business cycle. Advertising is a different beast. It’s highly cyclical, with big highs and low lows. The revenues are unsteady. And unlike electric utilities, which had monopolies over local supply, advertising is a highly competitive business.

I certainly don’t mind having starstruck investors pay for the buildout of infrastructure that will provide many benefits to many, many people, including myself. But if I were one of those investors, I’d be thinking hard about how exactly Google is going to be making money in the future. Stahlman says that “the mistake here is people think that Google is somehow an advertising company and therefore in the media business. They’re not. They’re in the infrastructure business.” Well, they’re making investments like an infrastructure company, i.e., a utility, but they’re making money like a media company. The question is: Is that sustainable?