It looks like a large Styrofoam takeout container. The 14-pound box would fit into a backpack were it not for the two antennas, set well apart. It can withstand subfreezing temperatures and 165-mph winds; it's even lightningproof. With the lid bolted down tightly, the box offers no clue as to what's inside. But disassembled, it reveals intricate innards that look like nothing so much as a city viewed from a plane: A million tiny wires crisscross like streets and weave among square parks the size of your thumbnail.

The magic of the box occurs when you mount it on the horizontal arm of a city lamppost, so that its long ears reach up to the sky. Install 30 of them per square mile (which isn't hard, since an installer using a single tool can put up a unit in 15 minutes) and they immediately begin communicating with one another via radio waves. Data, the same information that flows through the wired Internet, begins traveling between them. Establish some hub connections to usher the data back onto the Net and you've created a wireless network that can transmit signals all over real, life-size cities--into parks, schools, juice joints, bars, offices, playgrounds, and homes.

The boxes, known as routers or nodes, are made by Tropos Networks, a Silicon Valley upstart that's landed in the middle of a burgeoning movement among U.S. cities to create municipal wireless networks, or metroscale Wi-Fi--essentially, an effort to deliver wireless bandwidth to the masses. Since Tropos began selling its equipment in 2002, dozens of municipalities have signed up. The Twin Cities suburb of Chaska, Minnesota, built a wireless network to cover its 16 square miles and serve all 18,000 of its residents. Corpus Christi, Texas, bought 300 Tropos nodes to cover 24 square miles and has since decided to expand to 147 square miles. As it rebuilds in the wake of Hurricane Katrina, New Orleans plans to cover the whole town with a Tropos network. This summer, Anaheim, California, will hit the switch, giving 325,000 citizens across 50 square miles ubiquitous broadband Internet access. Tropos-powered networks also are in the offing in Philadelphia and San Francisco.

Launched with what Bill Gurley, a Silicon Valley venture capitalist and early Tropos investor, calls "four guys under 30 and an algorithm," the Sunnyvale-based company spent less than $3 million getting its first product to market. Since then, it has grown into the leading equipment provider in this incipient market, with more than $15 million in revenue in 2005 and a projected $45 million in 2006. It has had roughly 350 customers to date--including some in far-flung locales such as Bangkok, Kuala Lumpur, and Doha, Qatar--and partnerships with EarthLink, Google, Motorola, IBM, and others. Given its recent contracts, the company is well ahead of competing equipment makers.

Yet Tropos faces some difficult tests before it can realize its vision. The new, large-scale projects in San Francisco and Philadelphia will get the technology out of dress rehearsal and in front of a major audience. These launches will be key to the company's fate. As hundreds of other cities look on, contemplating whether to install their own cheap broadband, and as a phalanx of massive data carriers like Verizon and Comcast glower over what may be a new threat, Tropos will march out onstage. Says CEO Ron Sege: "The best thing we can do is make sure the big cities do well, for everyone to say, 'Oh, my God, it works."

"What Stops the Internet From Being Everywhere?"

In San Francisco, there is a new café every year that has "the best coffee in town." At the moment, it's Ritual, a chic place in the Mission District with leather couches, wireless Internet, and PowerBooks on every table. The two founding engineers of Tropos--Narasimha Chari, who goes by "Chari," and Devabhaktuni "Sri" Srikrishna--are sitting at a small table, drinking lattes and reflecting on recent news. About a year ago, the mayor of San Francisco put out a request for proposals, looking for the optimum plan for "unwiring" the city--that is, for creating a citywide Wi-Fi network. Just the day before, out of a half-dozen contenders, the selection had been announced--and Sri and Chari's list of big wins had gotten one municipal contract longer.

But the two men, both 32, scarcely stopped to rest. That's because each successive contract brings them closer to answering a question that's intrigued them since they met as undergraduates at Caltech about 15 years ago: "What stops the Internet from being everywhere?"

The magic of the box occurs when you mount it on a lamppost. Install 30 of them per square mile, and you've created a wireless network that can transmit data all over a city.

The inquiry arose out of mutual concerns about India and other developing countries. As a brainy boy growing up in Calcutta, Chari would take long excursions through the city searching for textbooks containing just the kind of math and science materials you can download in seconds today from the Internet; he knew that connecting people in poor and remote regions could be a profound form of change. Sri, for his part, had a deep desire to be useful and an appetite for solving engineering problems. So while attending graduate school in the late 1990s (Sri at MIT, Chari at Harvard), the two men would hang out in the bars around Cambridge and talk about how to get the Internet everywhere on the planet.

The intellectual challenge soon became as enticing as the moral one. It was a problem of cost efficiency: How could you bring the power of computer networks to villages hundreds of miles from the nearest cable TV, places where people can't even afford phones? It was a technical problem, of bouncing signals around in the air over large areas and then back to the nearest data wires. And finally it was a problem of overcoming natural physical limitations: the distance transmitted signals could travel, for one, and the amount of stuff that can be sent simultaneously. "It's just a very fascinating subject," says Sri. "We never really set out to start a company."

Any solution had to be dirt cheap. Even in the United States, broadband is so expensive, both to provide and to purchase, that its growth has not kept up with consumer appetites. Today many rural areas around the country have no high-speed data services, simply because it costs so much to dig up the streets and lay wire. Jupiter Research, a market research firm, estimates that 35 percent of Internet users in exurban or rural areas can get only dial-up connections. In some cases, the necessary conduits reach town, but jackhammering the last bit of pavement to serve a smattering of houses is more of a burden than it's worth. "There are some places where the economics are prohibitively expensive," says Brian Blevins, a Verizon spokesperson.

For Chari and Sri, the alternative to digging would have to be radio, and while drinking beer and poring over dense technical books, they came across a radio technology developed in the 1970s for military uses. The technology worked on battlefields, but its inventors and the engineers who came after assumed that it wouldn't scale. Sri and Chari thought otherwise. They suspected that if you could program the nodes of these radio networks cleverly enough, teaching them to move information around quickly, you could make the network as big as you wanted.

Their idea was a variation on the principle of the bucket brigade or steppingstones. If you can't get the signal to reach all the way to the wired Internet, make it hop from one transmitter to another until it does. And give it some basic rules for finding the most efficient pathway there.

Here at Ritual, for instance, e-mail data comes in over wires to a base station or router somewhere in the room and then heads through the air to the nearby laptop. Everyone in the café is just one hop from the wired Net. This configuration requires every user to be within about 100 feet of the device that's plugged in, and it's why wireless broadband is generally limited to offices and cafés. But what if you told that router to select another router for passing along its message, and told that router to select yet another after that? If you taught those routers to make efficient choices that wouldn't require arduous processing, eventually the Internet would spill out into the streets.

Sri and Chari got hold of some Wi-Fi gear--a cheap type of radio technology recently introduced to the enterprise market for office environments--and started playing with their routing ideas. They mounted antennas on cars and tooled around Cambridge, testing the performance of nodes programmed to obey their new steppingstone rules. "When we started doing this," Chari says, "people laughed at us, saying Wi-Fi is an indoor technology. But our approach has always been, don't take anyone's word for it."

The two men soon realized that they were no longer solving a math problem: They were developing a product. So they picked up and left Boston for northern California. They hooked up with two friends of friends who understood finance and formed a company. It was not a particularly opportune time. "In 2001, we were out there looking for funding. It was awful," says Chari. But Bill Gurley, whose firm, Benchmark Capital, invested early in companies such as eBay and Red Hat, liked their ideas. "I don't think anyone at that time was thinking about municipal wireless," Gurley recalls. "But what was keeping Wi-Fi from going outside?"

Even in the united states, More than a third of Internet users in exurban or rural areas can get only dial-up connections.

Well, nothing. In the United States, most towns already own the infrastructure for suspending 14-pound boxes in the sky: lampposts, traffic lights, telephone poles, city buildings. The Tropos routers themselves cost only about $3,500 each. So with 30 per square mile installed in a city like San Francisco, you'd spend about $5 million on boxes to serve more than 700,000 citizens. According to a report by PricewaterhouseCoopers, building a fiber network costs $2,000 "per home passed," in the industry's argot; providing DSL costs a few hundred dollars. Compare both with Philadelphia's estimate that the cost per home passed of its Wi-Fi network will be $30. On the user end of the equation, the hardware economics look even better. The Wi-Fi cards that early adopters were sliding into their laptops in 1999 went for about $2,000 apiece. Today the devices are preloaded into nearly all new computers and cost less than $10 each. Right now, as Chari and Sri drain their lattes at Ritual, there are an estimated 50 million Wi-Fi-ready computers out there.

So Bill Gurley got onboard. He liked the open standards of Wi-Fi technology and how quickly the price on the user's side was dropping. He loved Chari and Sri's vision of teaching routers with limited range and capacity how to build bucket brigades and choose the most promising pathways, based on the condition of the network. "It's very elegant," Gurley says. He also liked the growth potential of the market and the focus on software. "As a venture capitalist, I love everything about the Tropos model," he says.

In January 2002, Benchmark Capital ponied up $2.2 million for the young company to work with. Other VC firms followed, including the Intel Communications Fund and Siemens Venture Capital. And so did Ron Sege.

Good Enough Beats Best

Ron Sege (pronounced seh-gee) is a tall stick of a guy with blue eyes and blond eyelashes, whose elaborately stitched jeans were meant for a younger man. At 49, he is on his second wife, his second batch of kids, and the fourth small company he intends to make large. In a sense, Sege is a Web 2.0 guy all around, bringing hard-earned experience to a young company with a still-unproven business model. As he puts it, "I've seen this movie before."

Sege began working in technology in the 1980s, but really hit his stride in the '90s, as a manager at 3Com, the company that spawned Ethernet technology. 3Com had a few hundred employees when he arrived; by the time he left in 1998, he had 4,000 employees under his leadership alone. He learned a lot about growth in his 10 years at 3Com, but more interesting was what he learned about the power of lowering your standards. "From a tech perspective, good enough beats best," he says. Ethernet, the protocol that allows office PCs to share databases and printers and storage in a small local network, was far from perfect. "But it was inexpensive, easy to use, and anybody could design to it." Sege learned the beauty of this approach to business--float a quick and dirty product, let users and other product developers improve on it, and push it as a dominant shared platform. "Wi-Fi has many of the same attributes," he says.

After 3Com, Sege took a job as executive vice president of Lycos, one of the first Internet portals, where he helped engineer an Internet-bubble buying spree that included acquisitions of,, and Wired Digital. "That was my media mogul period," Sege says with a laugh. He left Lycos in 2001 and joined Ellacoya Networks, a company based in Merrimack, New Hampshire, that creates software to help broadband providers ease congestion in their networks.

Bill Gurley, tipped off by a Benchmark partner who'd worked with Sege in the past, saw in the Ellacoya CEO someone who'd ridden small companies through significant growth and who understood a good deal about data networks. He contacted Sege and told him about Tropos. The company made sense to Sege. Taking off-the-shelf indoor base stations and sticking them up on power poles--that was a formula he understood. Sri and Chari had already come up with the tricks, the proprietary algorithms for handling data traffic and monitoring the system from one main PC, which would set Tropos apart from its direct competitors. (The company has 30 software patents and patents pending.) In 2004, Sege came onboard--"to do all the stuff not involved with writing software."

At first, that meant selling Tropos boxes and software to a small but eager market the start-up had identified: police and fire departments. After September 11, the consequences of poor emergency communications became painfully clear to city leaders nationwide, and many municipalities were attempting to do something about it. What few civilians realize is that their heroes with hoses and their men and women in blue have always relied on only one of their senses for passing information: their ears. They use the same two-way radio technology today that police departments adopted in the 1930s. Some forces have introduced computers into their cruisers for searching DMV or criminal databases, but these hookups are as slow as your first dial-up modem. Forget about downloading a mug shot. Maps, surveillance videos, traffic updates, real-time messaging? Impossible. What emergency responders need is broadband. And it has to be broadband that's everywhere, broadband that moves.

Tropos could deliver that. Sege traveled the country, giving presentations to police and fire departments, steadily signing up customers. Oklahoma City bought Tropos technology to build a network for its police department covering 620 square miles. In Milpitas, California, about 10 miles from the Tropos headquarters, a 40-node Tropos mesh allows police to look up DMV photos and monitor video surveillance of high-crime areas.

So Sege and his team were surprised in the spring of 2004 when they got an order from Chaska, Minnesota, a Twin Cities suburb that wasn't looking to serve its police force. The town's city council wanted cheaper connectivity--for all of its residents, who were stuck paying $45 per month for high-speed access from Sprint and Time-Warner Cable. The goal was to provide broadband access for all of its citizens for no more than $20 a month. "Tropos was selling a system for public safety departments. Our IT guys thought, 'Why couldn't you do 3,000 connections instead of 300?" says Chaska's city administrator, Dave Pokorney.

For Tropos, this was exhilarating. Chaska had come up with this plan on its own, with no help from Tropos, which was focusing its efforts on public safety. The company had helped create networks designed to serve the general public, but only in parks or other circumscribed areas. Chaska was out ahead of them--and within three months, the city had a real-life metroscale network available to anyone in town.

Sleeping Giants

Everyone at Tropos agrees on what made the company take off. It happened in August of 2004, when Philadelphia, the largest municipality to date to do so, announced plans to blanket the city with Wi-Fi. The idea was to deliver cheap, and possibly free, broadband Internet access to the 1.5 million souls--digital haves and have-nots alike--who lived within the city's 135 square miles.

This was a bold, pioneering step, lauded by civic groups and techies around the country. But the news hit one party particularly hard: Verizon. At the time, the vast majority of Philadelphians who wanted fast connections to the Web had been coming to Verizon for DSL. Now the company would have a new competitor. The proverbial sleeping giant was caught off guard.

It's one thing to build a wireless network for 8,000 households in the suburbs of Minnesota. But it's something else entirely to do so in one of the nation's biggest metros.

Verizon's lobbyists marched straight to state lawmakers in Harrisburg and demanded action. And they got it. A telecommunications bill that had been lingering around the capital for more than a year suddenly came up for a vote, and it had a brand-new provision attached to it. The measure said that Pennsylvania cities intending to create high-speed data networks must give the dominant local phone company the right to build first. If the incumbent proceeded within 14 months, the city would be required to drop its plans. For the leaders of Philadelphia, that meant doing nothing for more than a year before getting their project under way. It also meant that cheaper service--some subsidized for the poor--would happen only at the whim of Verizon.

But the prospect of an Internet cloud floating through every park and into the city's overlooked neighborhoods had already intrigued many Philadelphians, and the state legislature's intervention galvanized people to protect the idea. "The school district, the nonprofits that wanted to serve poor neighborhoods, even our tourism organizations saw the potential," says Dianah Neff, Philadelphia's chief information officer and a 14-year veteran of Silicon Valley businesses. "When the legislation came up, we put the pressure on. We had 3,000 people call, write, and e-mail the governor."

Tropos, which already had been tapped to install two pilot projects in public parks, watched the events unfold. Sege hired a Washington lobbying firm, which showed up in Harrisburg, attempting to sway leaders to spare local governments from restrictions. In late November 2004, just as the bill was approved, Philly's Wi-Fi enthusiasts got a break. "It was almost like diving to get the catch in the end zone," says Sege. The state agreed to exempt Philadelphia from the requirements. (All other Pennsylvania municipalities remain bound by it.)

The way Sege sees it, Verizon's in-your-face tactics were the best thing that had ever happened to the start-up. The giant telecom's reaction made dozens of other cities take notice. If Verizon was so ruffled, people seemed to think, then Philadelphia must have been on to something interesting; the technology's potential must be real. "The phone was ringing off the hook," says Sege. Cities around the country, from Minneapolis to Tempe, Arizona, began announcing plans for wireless networks. Several months later, the technology was validated by another waking giant when Cisco announced it would begin building routers for muni Wi-Fi. Tropos sales went from 90 municipal clients in all of 2004 to 75 in just the first half of 2005.

The next step in the Philadelphia project was to respond to the city's RFP, and Tropos now had to get down to details. The company had the gear and the software for monitoring and troubleshooting the network, but there was a lot the small company was lacking. Customer service for one thing. And billing. And consumer sales. Rather than build those capabilities in-house, Sege began searching for an established Internet service provider with which to partner. EarthLink fit the bill. The ISP, based in Atlanta, had thrived as a middleman, buying wholesale dial tone, wrapping it up in an attractive brand, and selling it to Internet surfers. But as the world shifted to faster wires and fiber optics, EarthLink had little to offer. Unlike the phone companies, it owned no connections into the home.

In January 2005, Bill Gurley paid a visit to EarthLink's board of directors. He presented his case for a partnership, in which Tropos would provide infrastructure--the actual broadband network--and EarthLink would handle customer support and sales. In response to Gurley's presentation, EarthLink sent a team to visit Chaska to see for themselves if the new technology worked. The group toured the town and climbed under tables testing the network's reliability. They interviewed folks in bars. And they were sold on it. "Municipal Wi-Fi is really important for us," says Donald Berryman, EarthLink's president of municipal networks. "It's one of the top three investments we're making in future products. It can help us control our destiny because we'll own the network." Tropos and EarthLink have since landed deals with five cities and have proposals out to five more.

But Will It Really Work?

Not surprisingly, the Bells and other data-access providers haven't backed down. Since the maneuver in Pennsylvania, giants like BellSouth and Comcast have fueled a fight against muni Wi-Fi across the country. Lawmakers in Ohio, Virginia, Kansas, and Oregon, among others, have proposed legislation to keep local governments from building their own networks or at least make it more difficult for them to do so. Fourteen states, including Florida and Colorado, have already passed restrictions. "We have not supported a ban on municipal networks," says Verizon's Brian Blevins. "But we've felt where there's vibrant competition, the networks can undercut and disrupt a market that's working very well."

Critics of muni Wi-Fi argue that if local governments participate in building broadband networks, they'll exploit unfair tax and regulatory advantages, irresponsibly drain public coffers, and mismanage the services. To counter the legislative gambit, Sege and others have taken to evangelizing in Washington, D.C., and state capitals. They've made some progress. In June 2005, Republican Senator John McCain of Arizona and Democratic Senator Frank Lautenberg of New Jersey introduced a federal bill in answer to the activity in the states. The Community Broadband Act of 2005, still in committee, would "preserve and protect the ability of local governments to provide broadband capability and services." Says one Lautenberg staffer: "The senator doesn't think there should be obstacles--we're 16th in the world in terms of broadband penetration." A bill awaiting a vote by the House, on the other hand, would create barriers--for instance, requiring cities to partner with a private company. A restriction like that, though seemingly innocuous, would have prevented Chaska from building its network.

These policy struggles are not the only hurdles Tropos is facing as it lunges for profitability in 2007. There are big technical questions. It's one thing to build a wireless network for 8,000 households in the suburbs of Minnesota. But it's something else entirely to do so in one of the nation's biggest metros. "Nobody's demonstrated that you can have 135 miles of Wi-Fi," says Julie Ask, a research director at Jupiter Research. Radio signal is notoriously unpredictable. When your cell phone drops out every time you round the corner of Elm Street, that's because the mobile provider didn't predict a problem there. Home devices from cordless phones to baby monitors might cause interference. Tempe, Arizona, where Tropos competitor Strix Systems provided 500 wireless routers, discovered that signal wasn't getting through house walls beyond 150 yards from the routers. Many Tempe users found they needed an additional $100 device to receive and send data from indoors. Tropos could face similar problems.

Dozens of municipalities have joined in, but there is not much of a record. "As a mayor, why wouldn't you say, 'I want to bridge the digital divide'?" says Ask. "EarthLink wants to point to Philadelphia and say, 'Hey, it works,' but until there's proof'�" After a city government invests $20 million, no users will be happy if their connections go down or their webpages load slowly. The last thing Tropos needs is for annoyed customers to head back to Verizon.

Another looming question is what business models will work. Will consortia like the EarthLink-Tropos team for San Francisco prove easy for cities and profitable for the participating companies? Will the Bells hedge their bets and start offering their own systems? Will cities build their own public Internet utilities, just as many today deliver power without the help of private entities? In any of these scenarios, Tropos' business doesn't change. The Bells, the city governments, the ISPs--they'll all need to buy boxes from someone. As experiments are made and the best models emerge, Sege insists that Tropos will stay relevant.

First, of course, he has to deal with Philadelphia, which is building its 15-square-mile test area this summer and plans to roll out the full network in 2007. "I honestly believe that a lot of people are waiting to say, 'We told you it wouldn't work," Sege says. Philadelphia CIO Dianah Neff doesn't seem to mind that tension. "There's a lot of pressure on Tropos and EarthLink. But that's to our benefit because they're trying really hard," she says. "It's like you live in a fishbowl. It's not just other cities, but the world that's watching."

Martha Baer is co-author of Safe: The Race to Protect Ourselves in a Newly Dangerous World. This is her first story for Inc.