Quick Thoughts: Yes, Comcast, There is a Threat From Wireless Broadband
In my recent post about the “over the top” threat to cable, I wrote briefly about the potential for 4G wireless networks to compete directly with cable modem service for fixed residential broadband customers. For cable industry executives, wireless broadband is something akin to the bogeyman – they’ve been hearing about it since they were new hires twenty years ago and they don’t believe it. In the cable worldview, no existing network is fast enough to compete for video-capable broadband and it is too expensive to build one that can. Since Verizon decided to stanch its FiOS investment, this is the status quo for more than 70% of Americans (Exhibit 1).
Without competition, Cable MSOs can manage the “over the top” threat. Since all of the customers on each 500 household cable node share the same 152Mbps of Internet bandwidth, and since a single stream of HDTV quality video can require 7Mbps, only so many users can watch Netflix at the same time before the system capacity is tapped out and everyone sees performance degradation. Rather than invest to divert system capacity from Cable TV to expanding Internet service, Cable MSOs discourage Internet video by metering usage and hiking the price for high usage households. If “over the top” video gains enough traction to stimulate “cord cutting”, that is discontinuing channelized TV service in favor of Internet video only, MSOs will just hike the fees for high speed internet to make up the difference – given the rising cost of content, TV is a much lower margin business for cable operators than Internet anyway. Either way, the Cable industry will win, right?
Uh, no. The biggest flaw in the Cable wins argument lies in the assumption that no credible competition for high performance residential broadband will emerge. Yes, today’s 4G wireless networks do not have sufficient capacity to handle the more persistent demands of residential users and the immense cost of building out blanket national coverage is driving high pricing and usage caps that would seem to preclude their use for residential applications. However, the status does not usually remain quo for very long in wireless (Exhibit 2).
The LTE-Advanced specification was ratified as meeting the ITU standards for 4G in November 2010. Essentially, this verifies that the proposed technology is able to deliver speeds of at least 100Mbps for mobile users and 1Gbps for limited mobility users. The differences between the current LTE deployments and LTE-Advanced are substantial, although the specification allows for LTE devices to operate over LTE-Advanced networks and vice versa. Specifically, LTE-advanced supports the use of as many as 8×8 antennae for MIMO (Multiple Input Multiple Output), while LTE tops out at 4×4 and is currently deployed at 2×2 (Exhibit 3).
This technology offers a roughly linear improvement, meaning a near four-fold improvement in system capacity and device through put vs. current LTE networks. LTE Advanced is also expected to make use of more sophisticated error correction and other techniques to eke out modest (20-30%) additional throughput. Furthermore, LTE Advanced allows for the aggregation of up to 100 MHz of spectrum within a single cell, five times the maximum for LTE. This relationship is also fairly linear.
Today’s LTE systems, such as Verizon’s widely available service, offer roughly 120 Mbps in aggregate capacity per cell, ballpark comparable to the 152Mbps offered over each cable node. A full 8×8 MIMO, 100MHz LTE advanced system could increase total capacity by more than 15x per cell, to nearly 2Gbps. While this would obviously dramatically increase the per cell backhaul requirements, these technologies are also evolving quickly enough to accommodate the progress. Given the 7Mbps per stream requirement for HDTV, each cell site could handle more than 250 simultaneous streams. This is definitely fast enough and capacious enough to be a viable alternative to cable broadband. Of course, there are objections from the cable crowd.
“Yes, but LTE Advanced is years away and will likely be delayed!” LTE Advanced was ratified in November 2010 as meeting the ITU standards for true 4G – 100Mbps mobile download speeds and 1Gbps limited mobility downloads, amongst other requirements – and commercial equipment is expected by year end 2012, with network deployment in early 2013. While the full capability of LTE Advanced, may not be included in the first deployments due to current semiconductor processes, the speeds and capacities should progress steadily akin to Moore’s Law. Note that 3G networks were launched worldwide in 2001-2002, and have improved user speeds by more than a full magnitude in the decade since. Current industry roadmaps suggest hitting the full 4G capabilities by 2015 in commercially available equipment is a reasonable expectation.
“Yes, but there isn’t enough radio spectrum available to implement LTE Advanced!” At this moment, this is partly true, but change is afoot. The FCC has identified 415MHz of spectrum that it believes can be made available for wireless broadband over the next 5 years, which if the FCC is successful, would more than double the spectrum in use for commercial wireless services. 120 MHz of this spectrum is being wastefully used by television broadcasters, and is awaiting Congressional approval for incentive auctions that would motivate these station owners to vacate expeditiously with a portion of the proceeds (Exhibit 4). We believe that this will be approved. Another sizeable piece of spectrum is in use for commercial satellites and could be effectively shared with terrestrial applications. Both LightSquared and Dish are pursuing LTE projects using these bands. While there are currently objections that both networks would create interference for adjacent spectrum users, including GPS, these problems are technical in nature and can be resolved through refinement of the technology. The FCCs proposals do not include Clearwire’s 150MHz of spectrum in the relatively high band of 2.5GHz. Clearwire’s holdings are greatly disadvantaged for mobile service, as the maximum area covered by a cell at its frequencies is less 20% of the size that Verizon or AT&T LTE sites can cover at 700MHz. For fixed residential broadband, this distance limitation would be far less of a liability.
“Yes, but wireless networks are so expensive, operators cannot be cost competitive!” The big investment for wireless network operators is coverage. AT&T and Verizon have both built out more than 50,000 cell locations across the U.S. for their 3G networks. Upgrading a base station to support LTE costs $40-70K and adding a new LTE base station runs just over $100K (Exhibit 5). If a new cell tower is to be built, that can add $250-500K to the cost, assuming an appropriate location can be secured. Other one-time costs, such as upgrading backhaul and backbone networks can add further capital items. So if one is to build a nationwide LTE Advanced network, it would cost well into the billions of dollars. BUT WAIT! Who said every network has to provide blanket coast-to-coast coverage? For fixed broadband purposes, I can roll out LTE Advanced on a patchwork basis, cherry picking attractive communities. My first cell site will cost me $100-200K to build and will give me up to 100 square miles of coverage assuming that I have a chunk of the 700MHz spectrum block and that I am serving a flat topology. Once I put up that cell site, I can begin selling to every household in my coverage radius – potentially tens of thousands of households, depending on population density. I don’t really have to put up another cell site until the first one starts to run thin on capacity, but with up to 2Gbps to divvy up, my economics can look pretty good. A carrier already providing LTE on a national coverage basis may choose to upgrade to LTE advanced only in attractive residential markets – I suspect most users requiring multiple streams of HD video are not traveling at vehicular speeds – and achieve even better economics than a new build carrier. Once built, we note that all broadband carriers have the same marginal costs for carrying data – zero.
Adding it all up, we are confident that wireless will challenge wireline broadband. A decade ago, telephone executives pooh-poohed the idea that cable could ever take 10% of their residential market share, much less the nearly 25% that they have today. The idea that anyone would drop their fixed line telephone in favor of their cell phone would have been considered ridiculous. Now, nearly 30% of American adults have done so. Ask Verizon CEO Ivan Seidenberg who said “I’ve seen the movie” referring to his own company’s struggles to combat its own “cord cutting” against new technology enabled rivals. The lesson: ignore the relentless march of technology at your own peril.