The company is building an Internet Valhalla in Kansas City. Most observers have scoffed at the effort. But they're wrong.

By Kevin Kelleher, contributor

FORTUNE -- Last month, Google announced its Google Fiber initiative promising 1 gigabit of Internet speed for $70 a month and throwing in dozens of high-definition television channels for another $50 a month. The announcement was greeted with equal parts enthusiasm and skepticism.

The news elicited a lot of excitement and envy from geeks outside of Kansas City, where Google (GOOG) had chosen to do a pilot run of the project. Not only would Google waive the $300 installation fee for early subscribers, it would give them online access that's 500 times faster than the 2-megabits-per-second access most Americans have lived with for much of the past decade.

But that wasn't all. Google Fiber also offered the ability to record eight TV programs at a time. And to download a single high-def movie in seven seconds (versus 22 minutes for a 5mbps connection). And a free Nexus 7 tablet as a remote control. And a free terabyte of data storage on Google Drive. And, on top of all this, no data caps.

MORE: The information economy is reaching maximum overload

A week after the announcement, Google had signed up more than 7,000 homes in Kansas City, 4,900 in Missouri and 2,100 in Kansas. That's about 5% of the cities' total homes. Residents are halfway through the six-week rally to get others in their designated "fiberhoods" to sign up. As of this week, some neighborhoods have as much as 39% of their homes plugged in.

It's a promising start, but the debate is far from settled over whether Google Fiber will be disruptive in the way its search engine or Gmail was, or whether it will be an interesting -- but costly -- experiment like the company's plan to sell the Nexus one through its own online store.

That debate is likely to continue for several years, because it's going to take that long for Google Fiber to have anything more than an incremental impact on the way people access a high-bandwidth Internet. What's clear is that Google Fiber, as conceived at launch, has a lot going for it. But it also has a lot going against it -- notably the entrenched power of the giant ISPs Google is taking on.

MORE: Netflix has an identity problem

Verizon (VZ), for example, has offered its FiOS fiber-optic plan in select cities for years. The company says it has 4.2 million subscribers receiving 530 channels and broadband speeds up to 150 mpbs for the price of $205 a month. Verizon also claims to have spent $23 billion on the network, although some observers have questioned its numbers.

AT&T (T), Time-Warner (TWX) and Comcast (CMCSA) also offer Internet/TV plans at lower speeds. Unlike Google, they have longstanding relationships with their ISP customers and a stronger brand recognition in the broadband game. If Google expands its Fiber program to other cities, it will need to spend more on customer support -- which can be a quarter of an ISP's costs -- and compete for a broad array of cable channels.

In its favor, however, Google has launched Fiber with a thought-out plan aimed at the biggest frustration surrounding current ISP plans: relatively slow speeds at high prices. As video chat, streaming media and photo sharing have increased, people are finding their broadband isn't broad enough anymore. The U.S. is 15th on Akamai's list of the countries with fast broadband. Google Fiber offers faster speeds at lower prices than most ISP's offer today.

MORE: The trend that terrifies Big Tech

Google is also using the power of old-fashioned social networking to draw subscribers. The fiberhoods in Kansas City have incentives to get neighbors to sign up -- higher subscription rates bump you to the front of the line, and ensure that schools and emergency responders have access to fiber connections.

There is also a risky cost-versus-return bet that Google is taking. The Kansas City project will cost $500 million. Scaling that up to many cities would quickly burn through Google's $43 billion in cash. But the surge in broadband content could bring Google new revenue from web and TV advertising. And by becoming an ISP, Google could win a strong presence in many U.S. living rooms.

Google Fiber is one of the most ambitious undertakings in the company's history, up there with the push to redesign all Google sites around the social structure of Google+ and the surprising move to pay $12.5 billion for Motorola.

MORE: Hulu's network drama

Right now, Google's chances at disrupting the major ISPs looks like a long shot. But this is a gambit that will take years to play out, and it's way too early to count Google out if it's serious about offering one-gigabit access to millions of people at a rate competitive with much slower connections today. It's dramatically improved the business case for high-bandwidth access, in a way that incumbents will have to work hard to avoid.

As Ben Schacter, an analyst at Macquarie Securities, said in a recent report, "this initiative is less about a long-term revenue opportunity for Google and more about pushing current Internet providers to increase speeds and innovate (which could benefit Google in the long run)."

And that may be the real goal for Google. Not to take business from ISPs, but to get them to get their act together – improving service for consumers, and increasing revenue for many companies.