Google Fiber Was Doomed From the Start
The internet access answer won’t come from private markets, but rather from policies that make for competitive networks.
By Susan Crawford
Mar 14 2017
Just a handful of newsflashes have come home to me in such a way that I never forgot where I was when I heard them. Most were disasters, like the Challenger explosion or the attacks of September 11. In February 2010, I was sitting in my office in Ann Arbor when another event made the list — but this one surprised and delighted me. I cheered. Google had announced its fiber experiment, a plan to wire at least 50,000 homes with fast, bountiful connections. Finally, someone was going to try to unstick the monopolistic, stagnant, second-rate market for high-capacity internet access in the US.
Last month, there was a shakeup at Alphabet’s Access division (the new name for what was originally called Google Fiber). It named a new CEO, Greg McCray, and news outlets reported that hundreds of Access employees were being shifted to other parts of the Google empire. The former CEO, Craig Barratt, had already announcedlast year that Access would “pause” plans to launch fiber networks in several cities. Taken together, these news reports all seem to signal that Google is dumping the idea of fiber and moving decisively into wireless access solutions.
The bumpersticker from defenders of the status quo is that this means the Google Fiber experiment was a disaster. That’s simply not the case.
What this set of events does usefully and colorfully signal is that we need an entirely different approach to the country’s desperate need for world-class data transmission.
We do need fiber, everywhere. But we’re talking about basic infrastructure when we talk about fiber. And it is not in any private company’s short-term interest to make that basic fiber infrastructure — which amounts to a substantial upgrade to the last-century copper and cable lines with which Americans are now stuck — available to everyone at a reasonable price.
Google’s retreat is all about the bottom line. It wanted an unrealistic rate of return on basic infrastructure. It wanted to see rapid cost declines per subscriber, like the Moore’s Law changes in productivity that have taken place when digital technologists squeeze costs from other legacy businesses.
But although the cost of fiber — the glass itself — has fallen through the floor, and the gear needed to deliver signals over fiber has gotten cheaper over time, 80 percent or more of the cost of installing fiber is labor. The high upfront cost of getting all that labor to rip up the streets and hang wires on poles can’t be paid back in just a few years. The cost of that labor isn’t going down right now.
Basic physical infrastructure is like that. It requires long-term vision and patient capital — think horizons of 10 years or more, rather than two or three. In return, investments in basic infrastructure will pay steady, reliable returns until the sun explodes. And the spillovers from those investments in terms of economic growth and social justice for everyone in a community are routinely extraordinary. Think subway systems, railways, and telephone networks: Fiber optic internet access for everyone, and for every other form of infrastructure, is the most important recent addition to that list of physical networks.
As Google’s comfort with longterm returns waxed and waned, it labeled its fiber project an “experiment” (2010), then a “business” (2012), and finally a “bet” or “moonshot” (2015). Now it’s hoping to avoid many of those labor costs by experimenting with what can be done using its Webpass wireless access unit, rather than solely installing cables itself.