It wasn't so long ago — three years, to be exact — that communities all over the country were slavering over the prospect of getting fiber-based, blazing-fast broadband Internet service from Google.
More than 1,000 communities had fallen all over themselves to meet the company's terms, which included exemptions from some city regulations, expedited permit approvals and — the key — Google's right to choose which neighborhoods to endow with its fiber technology and when. This was a "demand-driven" model designed to ensure that the most densely populated and affluent areas got served first.
Hopes that Google would establish a nationwide model for fiber Internet service were dashed last month, when the company suddenly declared a "pause" in its plans to lay fiber in as many as 18 municipalities, beyond the eight metro areas where it already is building or has completed its system. The disappointed suitors will have to wait for this strategy to play out or move ahead on their own. They would be well advised to keep an eye on San Francisco.
That city is contemplating what would be the largest and most ambitious public broadband system in the country, with the ultimate goal of running fiber into every home and public building.
"Plain and simply, the Internet has to be viewed as a utility," San Francisco Supervisor Mark Farrell, who is spearheading the program, told me. "Just like water and power, people cannot function in today's world without access to the Internet. Over 100,000 San Francisco residents don't have access. Bridging that gap needs to be a priority, and it's up to individual cities to move forward."
San Francisco has some advantages other communities lack. It's geographically compact and densely populated, with a technophile citizenry.
But the city of 865,000 residents is facing a daunting challenge. No city of comparable size has deployed a system of fiber to every building, the board of supervisors' legislative analyst's office warned in March. The technical and physical obstacles to laying fiber citywide could sink the project at birth. The cost, which could come close to $1 billion, could raise public and political opposition. Existing Internet providers such as Comcast and AT&T can be expected to fight the project through lobbying and lawsuits.
Still, this is a city sensitive to its digital inadequacies. As of 2014, the legislative analyst reported, only 2.6% of San Francisco residents had gigabit-per-second access, the gold standard for broadband connectivity. To this day, service in a city closely identified with high technology is spotty and often atrocious.
"I feel like I'm always paying more for less," says Joaquin Siopack, 43, who works as a data analyst out of his home in the hilly Miraloma Park neighborhood. He recently abandoned AT&T's slow DSL Internet service in favor of a line from independent provider Sonic Internet delivering a pathetic 10 megabit-per-second connection — more than twice the speed he got from AT&T for $60 a month, about 40% more than he paid AT&T.
Neither Sonic nor any other service has plans to run fiber into his neighborhood any time soon. Cable service from Comcast would be an option — but he'd have to wire his home for the connection himself.
Some time in the next few weeks, San Francisco will hire a consultant to lay out the options for citywide fiber service. The choices include building its own network and acting as its own Internet service provider to provide gigabit speed access to every building in the city; offering incentives to coax private firms into building the physical network and providing service; or sharing construction and operation of the system with commercial partners.
Also uncertain is whether the city would build the system as a utility, serving all buildings, or replicate Google's demand-driven model by laying fiber only to where it knows customers exist.
Then there's the question of how the system would be paid for — by taxpayers, by all residents and businesses charged an average utility fee estimated at about $43 a month, or only by users paying a monthly fee. The range of potential construction costs, the legislative analyst reported, runs from more than $867 million for a city-built utility-style system, to zero, if the city chooses to turn the chore entirely over to private enterprise. But in that case, there would be no guarantee that all residents would be served.
"Every option is on the table," Farrell says, with one caveat: "Regardless of the method, I believe the public should own this infrastructure."
That's an important point, because America's reliance on private enterprise to build its broadband infrastructure has left the country mired in the digital equivalent of the Stone Age. The San Francisco legislative analyst reported that the average Internet speed nationally was 21.2 megabits per second in 2014, which ranked 19th among the top 20 countries in the world.
Competition among the cable and telecommunications firms is so meager—the nation's cable industry is a collection of regional monopolies — that they have felt almost no pressure to offer customers state-of-the-art connectivity or to reach out to underserved communities.
Broadband service is almost the very model of a public good that should be provided by government, because the profit incentives of private industry work in favor of limited service delivered to customers only selectively. Running fiber is a capital-intensive task: concentrating on high-rise apartment buildings occupied by tenants who will pay premium prices for lightning-fast service spells profits; reaching out to a sprawling community of low-income users, many of whom don't own computers, doesn't. The harvest is the "digital divide," in which modern technology bypasses already disadvantaged citizens.
Only a handful of public entities have taken matters in hand. Interest is beginning to stir—on election day, 26 cities and counties in Colorado opted out of a 2005 law — sponsored by the cable industry — prohibiting them from providing telecommunications services to the public. That brought to 95 the number of Colorado municipalities considering going it alone.
Yet of the more than 1,000 fiber networks in the U.S., only 143 are municipally owned, including six in California. The pioneers typically are small, affluent or tech-savvy communities, such as Santa Monica and Santa Clara, that tend to provide service chiefly to downtown business districts and public buildings, and see their programs as economic development tools.
That explains why Google Fiber initially looked like a godsend. When the company announced last December that it was exploring placing Los Angeles on its roster of "potential" fiber cities, Councilman Bob Blumenfield, who has been pushing to ease the way for would-be fiber providers in the city and had met with Google officials, declared that L.A. might be on its way to become "the world's premier gigabit city." (AT&T had announced the same week that it would start rolling out its GigaPower fiber service in Los Angeles County.)
Then came Oct. 25, a sort of Black Tuesday for believers in Google as a broadband savior. That day, the company announced the departure of Fiber's CEO, Craig Barrett, and the "pause" in its operations. "I'm hoping it's a pause and not a stop," Blumenfield says. "But all we know at the moment is that it's not happening right now." Los Angeles last year asked private companies to submit proposals for a citywide broadband system, but hasn't settled on a model.
Google will continue to build out its system in the eight communities where it's already established. But for other cities it's shifting its strategy toward wireless Internet service and to a target market of apartment buildings rather than single-family houses. This reflects its acquisition in June of WebPass, which runs fiber lines into neighborhoods, provides wireless connections from there to apartment buildings, and wired service from the buildings' rooftops to the tenants inside.
Google's strategic shift shouldn't have come as a surprise, says Christopher Mitchell, a broadband expert at the Institute for Local Self-Reliance. Signs abounded that the company, now called Alphabet, didn't see broadband service as a core business and was becoming more interested in other ventures, including self-driving cars. Its insistence on building on its own terms should have been a warning for companies that thought it would be the answer to their digital dreams.
"When you're begging for something from a big company like Google," Mitchell says, "you take what you can get."