Why is Google Making Phones, Anyway?

on October 6, 2017

Google-branded phones own about 1% share of the smartphone market, have limited carrier distribution, and won’t make real money for several years. Amazon and Microsoft both tried, and failed, on phones. Even Google’s initial phone foray, with Motorola, was a bomb. As my Techpinions colleague Jan Dawson pointed out in his excellent piece yesterday, Google doesn’t even seem all that serious about selling phones.

So, why is Google making smartphones? In fact, they just doubled down, acquiring HTC for $1.1 billion (a bargain) and introducing two new Pixel models.

The historic argument was that for Google, the more screens on which to view Google ads and do Google searches, the better. But Google is getting plenty of traffic from iPhone as the default search engine on iOS, as well as from Android devices. So selling a few million Pixels won’t really make a material difference to Google’s mobile search business.

My view is that Google’s commitment to being in the phone business is part of a broader strategy, with three central elements. First, there are those who believe that Pixel is an important component of Google’s commitment to hardware, along with Home speakers, Clips camera, Daydream VR headset, Pixelbook laptop, and even the AI-powered Pixel earbuds. But I think Pixel phones are necessary to some things that Google wants to accomplish in the wireless and connectivity arena. No, I don’t think they’re going to buy Sprint or T-Mobile. But I think they do want to play a broader role in connectivity, which includes multiple forms of wireless. Google continues to invest in Project Fi, their hybrid Wi-Fi mobile offering that works best on Google-centric phones such as Nexus and Pixel.

Google is also playing an increasingly pivotal role in the evolution to 5G. The company has a good chance of being selected by the FCC as one of the Spectrum Access System administrators for the shared 3.5 GHz spectrum (CBRS). Google has also been part of an industry push to make spectrum in the 3.7-4.2 GHz band available for 5G. Alongside this effort on the ‘mid-band front”, Google is part of a group of 30 key industry players pushing the FCC to set aside more spectrum for unlicensed (AKA Wi-Fi) users in the 6 GHz band, which sits in proximity to the 5 GHz band used for Wi-Fi. Also on the “5G” front, the Pixel 2 incorporates the latest LTE advances—3 Way Carrier Aggregation, 4×4 MIMO, and 256 QAM—which will allow Google to learn more about what some advances on the road to 5G might mean.

Google is investing in other forms of connectivity as well, from Google Fiber in the U.S. to Project Loon, an effort to bring the Internet to unserved areas. A test this summer kept one of its Loon balloons over Peruvian airspace for fourteen weeks. Google also needs to keep its eye on Facebook, which itself is involved in some of these groups but is also leading the Telecom Infrastructure Project and doing some other super-secret stuff.

Second, I think that being more directly involved in phones is important to Google’s AI push. Mobile devices are going to be a big part of how users experience AI, in ways different than they might from PCs or other types of hardware. And even though Google search is the default on iOS and Android, Google Assistant faces much greater competition from Apple (Siri) and Samsung (Bixby), plus of course Microsoft and Amazon. So Google needs its own hardware, from PCs to home speakers to phones, to work on some of its important AI efforts, such as Assistant. One of the new/unique aspects of the Pixel 2 is the Active Edge, which allows the user to squeeze the side of the phone to launch Assistant.

Third, Pixel represents a relatively cheap way to do some public R&D around several initiatives. If you think about it, Pixel is really a big public beta. They get HTC for $1.1 billion – a relative pittance in Google/Silicon Valley terms, plus they can now use Google powered phones to test things from Project Fi to AR (Google’s ARCore framework is fully active on the Pixel 2), to some nifty tricks on Assistant, in a way that they can’t on other Android-powered devices. And the risk is relatively low, by comparison. Something messes up, and they upset a few million Pixel/Nexus customers, not a few billion iPhone and Galaxy customers.

With a relatively modest investment, at least in Big Internet Player terms, Pixel phones allow Google to be masters of their own Android domain, providing a walled-off mobile testbed for new capabilities related to connectivity, AI, AR, and other concepts.