Apple's primary goal with the iPhone 3G is to saturate the market with iPhones, to make the device a major development platform alongside Windows and the Mac. That's why it's selling it cheap -- it wants to get as many devices into users' hands as possible, and it's willing to sacrifice some short-term profits to do it.

John Gruber at Daring Fireball nails it as usual:

So, step one: sell a ton of iPhones and grab a huge chunk of worldwide smartphone market share. That's the new $199 iPhone 3G. Step two: introduce features that people and companies love but which tie them to the iPhone. That's the SDK -- games and apps from App Store, and custom in-house apps for the enterprise market.

The physical phone isn't the story. A year from now, the iPhone 3G will be replaced by another new model. The platform is the story. Platforms have staying power, and, once entrenched, are very hard to displace.

He predicts, rightly, I think, that Apple will blow the doors off Steve Jobs' prediction last year that the company will sell 10 million iPhones by the end of this year.

And the formula applies to AT&T. While Apple is looking to entice users to make a long-term commitment to the iPhone platform, buying many iPhones over the years, AT&T wants to encourage use of its data services, which are becoming more important to the company than hardware sales, A&T Mobility CEO Ralph de la Vega told the Wall Street Journal:

It seems like $199 is the right kind of price point to get significant mass-market adoption. It's going to impact earnings in 2008 and 2009 in a negative way, but will turn very profitable in the long term. We generate a lot of revenue from iPhone users, about twice as much as other customers. And I feel very confident that we're going to have very low customer turnover, based on what we've seen from the initial version of the iPhone.

De la Vega elaborated with the New York Times:

$199 is a magic price point, never mind the higher, $30-a-month data fee. "What really prompted us to redo the deal was that we both felt that, in order to drive more volume, and have more penetration of the device, we had to change the model," Mr. de la Vega said.

"The price point of $199 seems to be an almost magical price point where customers buy in big numbers," he said. He pointed to the carrier's experience with the Motorola Razr phone. "It wasn't until we were able to lower the price below $200 that we saw a mass market for the device," he said. "That is the same thing that will happen here. The customers want the device, but some of them have not been able to afford it."

He added that AT&T's experience is the $30-a-month data plan won't scare off consumers. "That $30 is not a new price point," he said. "That is our standard price if you buy a BlackBerry Curve, a Treo, or a BlackJack device."

We've gotten used to talking about a category of devices called "smartphones," of which the current iPhone, Windows Mobile devices, and the Blackberry are three examples. They're cell phones that do more -- they do e-mail and play music and video and give you Web access.

But the iPhone 3G is really an early example of a new category of device: I'll call them "mobile pocket-sized computer-phones" (thus demonstrating one of my weaknesses as a tech journalist -- I'm terrible at inventing buzzwords): They do everything that smartphones do, and they run third-party applications, too. Smartphones run applications, but that's incidental to their purpose -- with this new category of device, running third-party applications is the point, the same way as it's the point for desktop computers. The iPhone 3G will be an early leader in this category, but Google Android phones will soon follow -- and Windows Mobile devices and the BlackBerry might also.