Google has consistently had a goal of increasing the speed of broadband and lowering the costs, presumably so its cloud and over-the-top products are more usable. They’ve famously been readying a FTTH access network in Kansas City. This development, especially the pricing they’ve announced, has caused trepidation and fear in the hearts of service providers. At $70.00 per month for Gigabit access the offering is substantially less than any offering from competitors in Kansas City or anywhere else in the US. However, a recent Forbes article cites the actual cost of installs, as well as time it takes to build, as reasons to question exactly why Google is taking the step of becoming a broadband provider. Clearly it takes a substantial amount of time to deploy any kind of wired broadband, and FTTH is often more difficult than other alternatives because fiber splicing is a less common skill than copper termination–whether on coax or twisted pair wiring.
Google could certainly have its sights set on becoming a large or even dominant service provider. If its pricing is sustainable then incumbent providers will be in for a very difficult time, since Google’s approach attacks the most profitable part of their offering. Its an open secret in the industry that cable operators and telcos offering video don’t make much (if any) profits on their video offerings. This is because modern headend infrastructure is prohibitively expensive, and the content providers (ESPN, TBS, FX, etc.) charge per video subscriber without regard to who watches their channel(s). Google’s video offering is largely in line with other service providers in terms of pricing and line up, demonstrating that even Google can’t dictate terms to the content providers. The difference between Google’s offering is the large jump in synchronous speed and the low cost of the broadband access. If Google expands their Gigabit service, the pricing will force competitors to drastically lower their per-sub revenue to compete, not to mention the increase in costs if they want to come close to matching the speed. So, from that standpoint it looks like Google has a winning product strategy, but taking that product into other markets will take a very long time. The Kansas City network has been in planning and deployment since the announcement last year, and getting a substantial number of homes and businesses connected will take much longer.
Given that Google’s overarching goal is increasing broadband speed while decreasing costs, we have to ask if the best way to accomplish that is for Google to become a service provider in multiple markets. Certainly from a rate of change perspective it may not be the “right” approach. Service providers in Kansas City haven’t yet made pricing/speed changes there, and we wouldn’t expect providers in other cities to do anything now. This means for Google to substantially change the broadband market by becoming a service provider they would need to both build out in many more cities, plus get a significant percentage of the broadband subscribers in those cities before other operators really adjust their pricing. While Google is now set to wire most of Kansas City, they may not see a significant base of subscribers until well into 2013. I suspect they may be looking at other methods to effect widespread change.
So, what else could Google do? We probably need to understand how Google reduced their cost to provide broadband service. They did it by attacking the only portion that’s really affected by Moore’s Law. Patrick Pichette invoked Moore’s Law during a presentation introducing Google’s Fiber product, but few people seem to have grasped why this is really important. It tells how Google is thinking about broadband. They understand that they aren’t going to revolutionize how home installs are done or significantly change the economics of pulling fiber. What they can attack with Moore’s Law are access equipment costs. Google took the approach of designing their own CPE (in home) and aggregation gear (what’s in the Fiber Huts). This choice allows Google not only to dramatically lower the cost of their gear, it also allows them to choose what to include and exclude.
Today the vast majority of broadband providers purchase their aggregation from companies like Cisco, Motorola, Alcatel, Arris, and Calix, and all of them are looking to make a profit. These same companies also often make the gear (whether cable, DSL, or FTTH) that goes in the customers’ home. It’s even more interesting that Google, who purchased Motorola for the mobile division, shows no desire to keep the division that makes access and CPE gear. If rumors are to be believed, Google is in the process of shopping the CMTS, set-top-box, and other cable gear unit. If Google were interested in being an equipment manufacturer, I would expect them to keep that division and expand it into FTTH gear.
Google has been a master at disrupting markets, and I think they’re poised to do so again. Since Google doesn’t seem to want to make equipment for broadband providers, but they do want to dramatically lower the cost, they can take a similar path as they did with Android in the smartphone space. In fact, for set top boxes and some other CPE gear they can probably use Android itself, but it’s not really suitable for a high speed network OS (operating system). Google clearly has an OS on their aggregation gear. Whether it’s built from Android or a more stripped down Linux kernel doesn’t really matter, because they own it and can open-source it easily. By open-sourcing the networking OS and releasing reference designs, Google can provide the same sort of supply chain that Nvidia and AMD (ATI) do with video cards–with Google providing the design and software and other companies doing the manufacturing and distribution.
This would open up competition and dramatically lower costs in an area of networking that is dominated by a small number of players. What’s more, Google could do this in less than a year and help unleash a plethora of alternative, fiber-based broadband providers. This seems to be the only fast method to accomplish their goal of reducing the cost of truly high speed broadband, and if I am correct, then it’s not service providers that need to be concerned, but rather the equipment manufacturers that sell to them. Google had to become a service provider so they could understand the problems and find solutions.