Oct 28, 2021: The Information’s Creator Economy Summit


Inside the Battle Over Google Fiber

When Google was planning to launch its Fiber broadband and TV service, Fiber executives had ambitious hopes of signing up around 5 million subscribers in five years, said a person close to Google's parent, Alphabet. But by the end of 2014, more than two years after service began, Google had only signed up around 200,000 broadband subscribers, said a former employee. The current number isn’t known, but it’s still well short of initial expectations, said another person close to Alphabet.

Now the company is rethinking its approach by shifting service to wireless, as has been widely reported. Wireless is a much cheaper way to offer broadband service than digging up streets to lay fiber cables in cities across America.

But that’s only part of the story. Last month, Alphabet CEO Larry Page ordered Google Fiber’s chief, Craig Barratt, to halve the size of the Google Fiber team to 500 people, said the second person close to Alphabet. (The Google Fiber unit is now known as Access.)

The Takeaway
Alphabet’s decision to shift its Google Fiber rollout to one based on wireless reflects a compromise among senior executives at the company. Alphabet’s founders had grown unhappy about the cost of the Fiber rollout. Meanwhile Fiber chief Craig Barratt had contemplated leaving, unhappy over how Fiber would be affected by the Alphabet restructuring.

Mr. Page has also told Mr. Barratt to reduce the current cost of bringing Google Fiber to customers’ homes to one-tenth the current level.

Keeping Google Fiber going using cheaper technology and with lower overhead reflects a tenuous compromise among a group of senior executives at Alphabet with mixed feelings about the project. On one side, Alphabet co-founders Mr. Page and Sergey Brin aren’t satisfied with the pace of Google Fiber’s rollout or the costs.

Mediator Role

CFO Ruth Porat, though known as a cost-cutter, has played a mediator role. She has told Mr. Page that Google Fiber has a solid business model that can succeed, and needs to be given time to work, said the person close to Alphabet. “She is in the middle saying, ‘Hey, relax, this is a complicated business, let’s see what they can do with the budget they have,” the person said. Still, “the Fiber group is on a pretty tight leash and getting a lot of feedback that they should solve the problems with technology,” using wireless.

As for Mr. Barratt, he was seriously contemplating leaving Google a year ago, according to people familiar with the situation. He was unhappy about the restructuring of the company into Alphabet, which carved off startups like Google Fiber into separate units with their own profit and loss. Mr. Barratt worried that would make it harder for Fiber to get resources and his dissatisfaction was “an open secret” at the time, one of the people said. He ended up staying.

The episode is the latest example of how the restructuring of Google into Alphabet hasn’t gone smoothly. It also suggests that Google Fiber remains on shaky ground.

Mr. Brin’s discontent with Google Fiber became apparent at an offsite meeting in Napa, California, late last year, where the Google X unit and Mr. Barratt’s Fiber unit were in attendance. Mr. Brin said that Fiber was a low-margin business and said Alphabet would have to find less expensive methods than digging up streets to deploy fiber optic cables in order to continue expanding coverage, said the person, who was at the event. His comments stunned the Google Fiber team, who hadn’t previously heard one of the founders express their attitudes about the project.

The main problem is that costs were much higher than Google expected. The original 2010 business model for Fiber put together by then-CFO Patrick Pichette was too rosy, said people who saw it.

Building a new fiber service is enormously costly. Verizon, for instance, spent $23 billion to lay fiber cables for its FiOS project over the past decade, looking to reach 18 million homes. Google had hoped its project would pass 40 million homes, said a former employee.

“The Fiber group is on a pretty tight leash and getting a lot of feedback that they should solve the problems with technology.”

“It’s extremely expensive, even by Google standards, to do what they’re doing,” said the person. “When you have millions of paying subscribers, then you start making money, and in the long term it makes the investment worthwhile. But it’s a very different style for Google to spend money up front on infrastructure and recoup it over time.”

Viral Explosion

Mr. Barratt’s idea had been to expand Fiber to as many cities as possible to get a viral explosion of subscribers and revenue, which Google could then use to expand to other cities, said the person.

One former Google Fiber employee said any hopes that the service could have been profitable were dashed by an overly aggressive rollout plan. While Google started in Kansas City in 2012, it quickly expanded its ambitions and now is in seven cities, with plans for 16 more. The speedy expansion meant the Google Fiber team was unable to develop a systematic approach for bringing the technology to new cities, which could have lowered the costs, said the person. “Every city was a fire drill, and we weren’t able to learn from our mistakes,” said the person.

Google saw using installation contractors as key to its plan to ease future Fiber rollouts to multiple cities because the company wouldn't be doing the work itself, said the former employee. But Google delayed Fiber rollouts to some cities for several months because it had issues with getting the contractors to complete the work on time, said the person.

Google Fiber is now focused on using wireless technologies developed in-house to bridge the last 100 yards between the fiber lines and utility poles to customers’ homes. In June, Google acquired Webpass, a company that develops high-speed wireless internet access technology.

—Jessica E. Lessin contributed to this article.

Kevin McLaughlin has been a reporter at The Information since 2016, covering cloud computing, enterprise software and artificial intelligence. He is based in San Francisco and you can find him on Twitter @ KevKubernetes.