What’s Google really up to in wireless? Might not be what you think.

Is Google really considering taking on the wireless industry? According to The Information, the company has talked to Verizon and Sprint about leasing wireless capacity to launch a mobile service.

Earlier this week, the blog cited unnamed sources who claimed Google has had talks with Verizon and Sprint over leasing access to their mobile networks in markets where Google has already deployed its Google Fiber 1Gbps broadband service. So far, Google has launched service in Kansas City. And it's building Google Fiber in Austin, Texas, and Provo, Utah. It also recently announced plans to expand the broadband service to several more cities.

Some other blogs have speculated, based on rumors and reports like this, that Google may be interested in becoming a mobile virtual network operator, or MVNO, meaning Google would buy access to these established wireless networks at wholesale rates and resell the service to customers, thus competing directly against wireless operators.

I know what you're thinking. How could Google be so stupid? Didn't the big cable companies try to do the same thing twice and didn't they fail both times?

Yes, you're correct. But before you start jumping to conclusions about the type of service Google may be offering, let me start by saying that the people in charge of Google's broadband strategy aren't dumb. The reports speculating that Google may try to enter the wireless market to compete directly with mobile operators, such as AT&T or Verizon, may be overstated. My theory is that Google is exploring its options and is much more likely to develop a strategy that looks more like what its cable rivals are doing today, rather than rehashing the cable industry's failed attempts to become traditional wireless resellers.

Not surprisingly, neither Google nor either of the wireless carriers would comment on the rumors.

Why a Google MVNO is a bad idea
Building a wireless business as an MVNO that uses another wireless operators' setup is a tough way to make money, even for a company like Google. For one, the very fact that you have to rely on another company's wireless network is risky. What's more, even though Google would likely get access to a far more extensive wireless network than it could build itself in a short period of time, the network assets that major wireless carriers offer MVNOs is typically limited.

For example, Verizon doesn't offer any of its wireless reseller partners access to its 4G LTE network. The only MVNOs that get access to Verizon's LTE network are the ones involved in the company's LTE Rural America project. This project lets rural operators use Verizon 700MHz spectrum to build out LTE infrastructure in parts of the country where Verizon doesn't find it cost effective to build a network. Then those operators become partners with Verizon, and their customers can roam onto Verizon and Verizon customers can roam onto their networks.

Sprint also limits the coverage its reseller partners can use to only the Sprint territory. It doesn't include access to any of Sprint's roaming partners, which reduces the overall footrpint the reseller has access to. Still, Sprint does offer access to its 4G LTE network.

And of course, there's the problem of competing with a company whose infrastructure you lease. If the bursting of the telecom bubble in the early 2000s taught the communications industry anything, it's that building a business around someone else's infrastructure just isn't a good idea.

What's more, Google has made it clear that it has no interest in selling traditional voice services. While the cable operators and telephone companies it competes against with Google Fiber all offer phone services as part of a "triple play" bundle, Google decided not to include telephone service as part of its offering because it didn't want to have to comply with all the regulatory requirements. My guess is that the company would be equally averse to complying with wireless voice service requirements.

The last thing to point out here is that selling wireless service doesn't even really make sense for Google, since it's unlikely to help the company's fiber broadband business. This was the harshest lesson the cable companies learned during their foray into the cellular market.

"When we started down this path eight years ago, we believed that if [we] didn't have a cellular service for the quad play, we'd be at a huge disadvantage," Comcast Senior Vice President of Strategic Initiatives Tom Nagel said in a recent interview discussing Comcast's wireless strategy. "But what we learned was that the 'quad' play wasn't really necessary."

Nagle said the quad play sounds good on paper but that the reality is people buy mobile phone and cable services in different ways.

"Wireless may be sold as a family plan," he said. "But each person signs up for the service at different times. Cable and broadband are household services."

What happened with cable anyway?
Nearly a decade ago, some of the biggest cable operators around, including Comcast, Time Warner Cable, and Cox Communications, formed a joint venture with Sprint called Pivot, which was supposed to deliver what everyone was calling the "quad play" of services: TV, home phone, broadband, and mobile phone service. The cable operators planned to use this joint venture to resell Sprint wireless service and bundle it with their other services. The idea was to create an even bigger package that would help cable operators retain customers.

But the joint venture was a miserable failure. Two and half years after it was formed, Pivot was dissolved when Comcast and Time Warner Cable pulled out. In 2008, the cable companies came together again to bankroll another joint wireless deal between Sprint and Clearwire with the expectation that they again would be able to resell wireless service. That was also a flop and the cable companies pulled out of the joint venture, leaving Sprint as the only stakeholder.

These same cable operators also paid $2.4 billion for licensed wireless spectrum in a 2006 FCC auction. They later sold this spectrum in 2011 to Verizon for $3.6 billion.

Even though the cable operators made a nice little profit from their spectrum sale to Verizon, overall they learned some difficult lessons during the past decade of wireless experimentation.

The biggest lesson was that reselling a wireless operator's mobile phone service was unnecessary and difficult. The second lesson was that having a mobile play is important, but not in the manner you'd expect.

"We realized that licensed spectrum was not critical," Nagel said. "Yet we still believe that making our products mobile is very important. Our mobile strategy today is taking this awesome high-speed data product and extending it. It's like taking your coax cable with you."

Indeed, Comcast, Time Warner, Cablevision, Cox, and Bright House Networks have pooled their assets together to build CableWiFi, the largest Wi-Fi network in the country. The network, which uses unlicensed wireless spectrum, consists of 200,000 indoor and outdoor hot spots. Today, the network is mostly concentrated on the east and west coasts as well as in big cities in the midwest. But the network is growing quickly. And with a recent FCC vote to free up additional unlicensed spectrum in the 5GHz band, these cable operators say they'll be able to do even more with their Wi-Fi network.

Lessons for Google
So what does all this mean for Google and its potential plans to launch a wireless service? My guess is that Google is likely looking at developing a mobile strategy more along the lines of what the cable operators are doing, rather than becoming a mobile carrier that directly competes with the wireless giants.

This makes sense for a few reasons. First, with Google Fiber, Google already has the necessary high-capacity backhaul in the ground that's necessary to support an extensive mobile network that acts as an extension of its Google Fiber service. And let's not forget that Google also has experience delivering community based Wi-Fi. The company launched a citywide Wi-Fi project in its hometown of Mountain View, Calif., and it has been operating a free community Wi-Fi network in New York City in the neighborhood surrounding its office in Manhattan.

If Google is indeed talking to Verizon and Sprint about leasing licensed wireless network capacity, my guess is that the company hopes to use these networks only as backup in areas Wi-Fi may not be able to reach. This type of service could end up looking a lot like the service that some interesting new MVNOs, such as Republic Wireless and Ting, are offering.

These companies are offering alternative and very cheap wireless phone and mobile data service. Republic has voice and data plans starting at as low as $10 a month. The company is able to keep prices so low by using Wi-Fi as the main form of network connection and then licensing Sprint's network capacity to handle voice calls and data sessions in areas where Wi-Fi isn't available. The idea is that customers will spend a majority of their time on the free Wi-Fi network and only a fraction of their time on Sprint's network, which Republic pays for by the megabyte.

While the cable industry hasn't yet partnered with a company like Republic or Ting, it's intrigued by their business model and the use of Wi-Fi to provide a low cost alternative wireless service.

"We know of Republic Wireless," Nagel said during our discussion about Comcast's wireless ambitions. "We've talked to them. There are a number of players in that space. And I think it's interesting. It's all about learning and exploring what's possible with Wi-Fi."

As for partnering with a mobile carrier --yet again-- to further its mobile broadband strategy, Nagel admitted Comcast has considered it.

"We look at 4G LTE all the time and evaluate how it might be used," he said. "It's not front and center for us right now. We definitely aren't sitting around saying, 'Boy, I wish we had an MVNO.' But you never know what will happen in the future. The one thing we do know is that Wi-Fi will be the foundation of whatever we do in mobile."

Maybe, just maybe, that's what Google is thinking too.