Google Fiber is bailing of Louisville, Kentucky because it screwed up its fiber build there. In an attempt to move quickly and save money, Google forgot the iron law of engineering:
Good, fast, cheap. Pick any two.
Google went with fast and cheap, and it turned out not so good. The problem was microtrenching, and its little brother, nanotrenching. Which particular techniques were the problem isn’t clear, but the result is. According to Google’s blog post yesterday…
We’re not living up to the high standards we set for ourselves, or the standards we’ve demonstrated in other Fiber cities. We would need to essentially rebuild our entire network in Louisville to provide the great service that Google Fiber is known for, and that’s just not the right business decision for us.
The lessons we’ve learned in Louisville have already made us better in our other Google Fiber cities. We’ve refined our micro trenching methods and are seeing good outcomes elsewhere.
But it’ll cost too much to rebuild its plant in Louisville, so it’s adios. According to a story on WDRB.com (via a link on Google’s blog post), the epoxy compound that Google was using to fill up the shallow slits it dug in streets for its fiber – that’s how microtrenching is done – failed. The fix they planned to use was to go back, scrape the epoxy out of the slits and refill them with asphalt. Reading between the lines of Google’s blog post, that technique didn’t work any better – the implication is that Google would have to rip everything out and start over again if it wanted to keep doing business in Louisville.
Google says it going to move ahead with fiber projects in other markets. The list includes Kansas City, Austin, Provo, Charlotte, Atlanta, Orange County, Salt Lake City, Raleigh-Durham, Nashville, Huntsville and San Antonio.
The list doesn’t include San Francisco, or any of the other Bay Area cities where it’s relying on Webpass, an Internet service provider – primarily wireless – it acquired in 2017.