In January I traveled to Kansas City, Kan., to talk to people who have been using Google Fiber, the search company's effort to wire a large American city with 1-gigabit Internet lines. That's 100 times faster than the average American home broadband speed, and it's not even that expensive. Google will sell you a gigabit line for $70 a month, TV service and Internet for $120, and it will provide free Internet service at a respectable 5 Mbps to anyone who pays a one-time $300 installation fee. (By comparison, in Chattanooga, Tenn., where the local power utility recently launched gigabit service, the fastest tier goes for $350 per month.)
Everyone I spoke to in Kansas City -- local officials, businesspeople, aspiring startup entrepreneurs, and ordinary residents -- was thrilled that Google had blessed the area with superfast Internet. As I argued this week, though, there was just one problem: Nobody knew what to do with 1-Gb broadband. Even the most far-reaching ideas didn't begin to utilize all that bandwidth.
This should not have surprised me because providing cheap gigabit Internet to a few hundred thousand people in one city, while awesome for them, isn't very useful on a macro scale. For ultra-high-speed Internet to deliver on its boosters' promises, we'll have to see lots of developers create apps that require huge bandwidth. That will only happen when a critical mass of people -- which likely means a few million, at least -- get access to gigabit broadband.
It's obvious how Google would benefit after we hit that critical mass. The search company's research (and that of other Web companies) proves that when you give people faster Internet service, they use the Internet more often. And when people use the Internet more, Google makes more money. Some of Google's biggest products -- YouTube, Android, Google+ Hangouts -- are only possible because the world moved from dial-up Internet to broadband. Google's future prosperity likely depends on the next big shift, from 5 Mbps broadband (the American average) to gigabit service. Yet widespread gigabit broadband is not inevitable. Compared to most developed countries, American Internet lines are slower, more expensive, and cover less of the country. And unlike the increasing speed of microprocessors and the falling cost of storage, American broadband hasn't been improving at a regular, dependable rate.
Thus, like everything else Google does, its Kansas City project only makes sense if you think about it on the grandest scale -- as the first step in a path toward revolutionizing American broadband. So that's the main question we've got to ask about Google Fiber: How does Google expect its relatively small Kansas City installation to spark a big shift in American broadband?
And the answer is: It won't say. A week after I returned from Kansas City, I drove to Google's headquarters to chat with Kevin Lo, general manager for Google Access, the division that houses Fiber. Lo spoke at length about why, out of more than a thousand contending cities, Google chose Kansas City for Fiber and why it settled on gigabit speeds for the launch. He was also happy to speculate about what people will do with Fiber lines once they're up and running all over the Kansas and Missouri sides of Kansas City. (Google says it will complete its initial rollout by the fall.) But Lo was cagey about the big question: How will Google turn Fiber into a big deal?
But it's easy to speculate about Google's potential path from Kansas City to the rest of the country. The first approach is direct -- if gigabit lines turn out to be popular and profitable for Google in Kansas City, it could move to launch them in other American cities. Because it had loads of applicant cities the first time -- many of which promised to cut through local bureaucracies and give Google the green light to install its own infrastructure -- the company would have an easy time setting up shop in a second city, and then on and on.
Then there's a second, indirect approach. Call it the Chrome model. When Google launched its lightning-fast Web browser in 2008, it did so with two goals. The first was to turn people on to Chrome. Because the browser is deeply integrated with Google's services, people who use Chrome are much more valuable to Google than users of other browsers. But Google didn't just want to create a better browser -- it also wanted to inject competition into the browser market as a way to force other browser makers to improve their products. For years before Chrome was released, the browser business was dominated by Microsoft's Internet Explorer, with Firefox running a distant second (though gaining). But neither IE nor Firefox fit what Google needed from a browser -- speed and constant, breakneck innovation. Google hoped that if it released a faster browser, Microsoft and Firefox would have to speed up their browsers to keep up. And, whichever browser came out on top, speedier browsers would ultimately benefit Google. And that's exactly what's happened. Chrome sparked a new round of innovation in browsers, and today all browser makers gush about their speed.
I asked Lo if this was the model Google has in mind with Fiber. Like the old browser business, the telecom market lacks competition -- most Americans have just one or two companies offering broadband to their homes, meaning that firms have little incentive to increase their speeds or lower their prices. Is Fiber an attempt to correct this -- to force telecom rivals to offer faster, cheaper broadband to compete with Google?
Lo shot me down. "I would say at face value the answer is no," he said. "That is not the objective." He went on to say that while Google wants more competition in broadband, that isn't the primary aim of Fiber. "We're focused on providing great service to our users in Kansas City," Lo said -- and then, as I pressed him in various ways to comment on the bigger picture, he repeated the mantra several times. He was especially reluctant to discuss any possibility that we might see Fiber installations in other cities.
Now, this calculated vagueness could be just a way to just keep Google's aims close to the vest. It could also be a way of saying that Google doesn't yet know what its plans are for Fiber -- that it's too soon to determine whether it should roll out elsewhere and whether Fiber will spur other high-speed broadband installations.
Lo was clear that for now, Google has rather straightforward aims: It expects to make a profit from Fiber in Kansas City. "You have to have sustainable business models if you want to be in business over the long haul, and we do. We have a model that we think makes a lot of sense," he said. He didn't elaborate on how the firm could make a profit by providing screaming broadband at low prices, but here, again, the Chrome model is useful. People who get Internet access through Google will be expected to use a lot of Google services, thus making a lot of money for Google -- and subsidizing the cost of its Fiber lines.
Google also thinks that it can use Kansas City as a test-bed for new services that require speedy Internet connections. "We have product managers at Google that have products they want to ship that they cannot ship because people at home can't use them," he said. Lo couldn't tell me what they are, but he suggested that we'd soon see some of them running as pilot projects among Fiber users. And it won't just be Google services: "After we launched in Kansas City, I had five different CEOs of very large companies call me up to ask about our deployment," he said. "They were all interested in going in there to pilot and try out new services. That points out the underlying point here about critical mass -- everything that we're trying to do in Kansas City is really trying to see what happens when you get a lot of people using these services."
Google has never said how much it's spending to install fiber lines in Kansas City. Because the company built its own fiber optic installation around the metropolitan area -- rather than lease fiber space from existing companies -- it's expected to be a lot, at least $50 million and likely many times more. In Google terms, though, that's not very much money to spend on a big bet -- especially since there's hardly any way Google can lose with Fiber. Even if Fiber doesn't spark a wider improvement in American broadband, and even if Google has a hard time finding a profit in Kansas City, it will still have built a small broadband nirvana where it can perfect the products of tomorrow. For a company whose livelihood depends on anticipating the future, getting a peek at how people with high-speed Internet will react to the service could be an end in itself. "We are in this for the long haul," Lo told me. "It's not an experiment."
• Manjoo is Slate's technology reporter and the author of "True Enough: Learning to Live in a Post-Fact Society." Twitter: @fmanjoo