Rumors say Google wants to build hardware from scratch, but what will that accomplish?
Over the weekend, The Telegraph reported that Google has plans to release a Google-branded phone that will "see Google take more control over design, manufacturing, and software." Google is apparently sick and tired of the iPhone "dominating" the high end of the smartphone market, and the company appears to believe a Google-built smartphone can solve this problem in a way that a Nexus device cannot.

The report further says that the Google Phone will appear "by the end of the year" and that it will exist in addition to the Nexus program, which the report says is "expected to continue this year with handsets made by Taiwanese company HTC."

I'm skeptical.

Three phones in six months?

We already know a good amount about the upcoming Nexus devices, if the current rumors are true. There will be two HTC-built devices, codenamed "Marlin" and "Sailfish." We have solid rumors from the usual "reliable sources," but we also know the phones exist because we can see Google make references to them in AOSP. Those two phones are also expected to be released this year (the Telegraph report agrees with this), so if the Telegraph is right, that's three smartphones from Google in the next six months.

This isn't the first time we've heard of a Google plan to take more control over its smartphones. Earlier this year, an Information report said Google wanted to take a more "Apple-like" approach to hardware development. That report claimed Google would design a phone itself and contract manufacturing out to Foxconn or another ODM. Google does exactly this with the "Pixel" line of devices, which so far has resulted in two Chromebooks and one kind-of-accidental Android tablet.
The approach doesn't seem efficient, but it also wouldn't be out of character for Google. This is a company that happily floods the market (and confuses consumers) with tons of similar products that compete with each other. Consider Google's communication "strategy," which has given us Google Wave, Google Buzz, Google Talk, Google Voice, Google+ Messenger, two Android SMS apps, Google Hangouts, YouTube Messages, Google Spaces, and soon Google Allo and Google Duo. It would definitely be in character for Google to release a hardware project that overlaps and competes with Nexus.

Building from scratch

What would Google gain by building its own smartphone? In the wider Android ecosystem, update speeds are an embarrassment, phones are full of crapware, and OEM skins ruin the consistent software design, but all of these problems are already fixed if you just buy a Nexus phone. Google totally controls the software side of a Nexus phone. It has launched Nexus-exclusive features in the past. It can do whatever it wants and make whatever changes it wants. A "Google Phone" would offer no benefits in the software department.

The only reason Google might want to build its own phone would be to fully take over the one thing it doesn't have total control over today: hardware design. Google has some input on the hardware of a Nexus device today, but it seems to come late in the design process. Every Nexus device is built in cooperation with an Android OEM, but the Nexus handsets don't stray far from existing models. For 2015, the Huawei-built Nexus 6P was clearly related to Huawei's flagship, the P8. The 2014 Motorola-built Nexus 6 was rumored to be a last-minute project between Google and Motorola, and apparently it was completely a Motorola design (the same hardware was rebranded as the "Moto X Pro" in China). For 2013, the LG-built Nexus 5 and the LG G2 were so closely related that with a little modification it was possible to swap parts between the two.
Does Google truly need to design hardware from the ground up, though? Today's flagship smartphones aren't all that different from one another. They all have the latest Qualcomm SoC, a big and dense display, a great camera, and a battery that never seems to last long enough. The biggest difference is the exterior design and materials, where you get to pick from plastic, glass, or aluminum. I've never heard anyone complain about the Nexus 6P hardware (other than the usual Snapdragon 810 complaints), and I doubt many people would have a problem with a Nexus phone based on the HTC 10. In short—today's hardware is fine.

So building your own from the ground up only makes sense if it's going to be radically different. But what radical changes can you make to a smartphone at this point? Google has a few projects in the works. There's Project Tango, a smartphone with extra sensors on the back that allow it to map the world in 3D, but a few additional sensors wouldn't require a ground-up redesign. Then there's Project Ara, a perpetually delayed modular smartphone, which truly would require a unique design. But Ara's clunky (and now watered-down) modular system, along with serious questions about battery life, don't really match the "high-end" iPhone-beating product The Telegraph describes.

Don't rock the Android boat

Google has been in the "build it yourself" position before; it used to be a phone manufacturer when it owned Motorola. When Google sold Motorola to Lenovo, it was reportedly part of a peace treaty between Google and Samsung. The same week that the Motorola sale was announced, Google and Samsung signed a 10-year cross-licensing deal. Even when Motorola was part of Google, Google tried to placate OEMs by placing a "firewall" between Motorola and the Android team, but apparently OEMs still worried about Moto getting preferential treatment.

The "Nexus" line was also part of this balancing act with OEMs. By co-branding a product with an OEM, Google could claim that it wasn't "directly competing" with its partners, just teaming up with one. Rotating build partners was part of this strategy, too. Google wasn't playing favorites; everyone gets a turn!

Google has always tried not to "rock the boat" when it comes to Android OEMs. And why should it? Android ships in four out of five smartphones, and we'd be willing to bet that the vast majority of its users rely on a ton of Google services.
A Google-made phone would fly in the face of all this. It would bring back all the bad feelings that appeared when Google bought Motorola. HTC in particular would be in an awful position, building a pair of Nexus devices that few people would want, since its devices would play second fiddle to the Google Phone. The same could be said of Lenovo if Google built a Project Tango phone. The Phab 2 Pro is supposed to be the Tango launch phone, and a Google Phone with Tango would steal all of its thunder.

The reality is that Android has defeated just about every challenger to the licensable mobile OS throne. Blackberry still won't admit it, but Blackberry OS is dead. None of the upstarts—Firefox OS, Tizen, or Ubuntu—got any traction. Amazon has found out that forking Android was harder than it looked with the Fire Phone. The closest thing to an Android challenger was Windows Phone, but even that is experiencing a Blackberry-like descent into irrelevance.

Maybe Google feels that it has a strong enough market position that it can now put the screws to OEMs and start competing directly with them. It's a future that OEMs keep planning for, and many of the big players have explored Android contingency plans at some point. It's hard to imagine Google feeling a need to do this, though. With the current Nexus setup, Google—primarily a software company—has all the control over software it wants. Building the hardware from scratch doesn't seem likely to make a huge practical difference, but it does cross a red line that OEMs have had a problem with in the past.