Does Facebook’s Terragraph Compete with 5G?

In May 2018, Facebook and Qualcomm announced a partnership where the two companies will collaborate in proceeding with Terragraph trials using Qualcomm’s 802.11 ay prestandard technology. Terragraph was presented during Facebook’s F8 Developer Conference in 2016 and proposes to use 60 GHz to provide gigabit connectivity to underserved urban areas. Facebook has published impressive antenna and infrastructure concepts built on the Wireless Gigabit (WiGig) standard. They aim to couple off-the-shelf hardware with Facebook’s formidable software skills to create a flexible and cost-effective urban connectivity network. The deployment model depends on the availability of metro-area fiber; Terragraph nodes are deployed in urban street furniture or buildings to provide seamless connectivity across an urban area. Facebook is running a trial in the San Jose area (which started late in 2016) but has not yet announced any further information about the success or challenges faced in this trial.

Qualcomm’s involvement in the Terragraph project is certainly important news, but it does not mean Facebook will succeed in the ultracompetitive urban connectivity domain. In fact, past initiatives by Facebook and Google, including in rural areas that are largely uncontested, have not been successful since neither hardware commoditization or software expertise can solve physical world problems. At the same time, Nokia offers an 802.11 ad product called Wireless PON; Lattice Semiconductor offers an 802.11 ad product for 60 GHz; and Qualcomm has in the past announced collaboration with enterprise Original Equipment Manufacturers (OEMs) (including Microtik) to improve enterprise Wi-Fi connectivity. Starry and Marvell also offer a 37–40 GHz product for Fixed Wireless Access (FWA). Although the WiGig market has not reached economies of scale, it does illustrate healthy activity.

Terragraph’s deployment model at 60 GHz is not new; in fact, there are Radio Frequency (RF) vendors that have been selling similar products (albeit built on proprietary technology rather than on WiGig) aiming at connecting urban and rural areas. For example, Siklu’s Etherhaul product is positioned toward street-level connectivity, using a mesh network at 60 GHz. On the other hand, 802.11 ay (to be standardized during 2019) can create economies of scale much faster than proprietary technologies and will naturally be more cost-efficient. Qualcomm is also an advocate of 802.11 ay (and 802.11 ad) and continues to drive support for the WiGig standards in its high-end System on Chips (SoCs). The Facebook initiative is an ideal partnership for Qualcomm and can raise awareness on a global scale, which may raise interest from other regions.

Similar to previous initiatives (e.g., Facebook Aquila, Google’s Project Loon, Google Fiber), Terragraph is expected to deploy in a large-scale trial in order to provide expertise for urban connectivity rather than compete head on with fixed or wireless FWA services by telecom service providers. It may create competitive pressure for telcos in the regions in which it is commercially deployed, but it is not expected to disrupt the mass market. It also comes to the market at a time when Verizon is pushing aggressively to deploy 5G FWA across the United States.

In the United States, Terragraph is designed to work in the unlicensed 60 GHz band, whereas 5G New Radio (NR) will be deployed at 28 GHz. The 5G specification has a considerable advantage in terms of radio propagation, although both systems will face deployment challenges since mmWave systems require much denser deployments compared to sub-6 GHz networks. At the moment, it is reasonable to assume Terragraph will remain limited to the U.S. market and in trial mode in the San Jose area for the foreseeable future and will not gain significant market traction before 802.11 ay is standardized. Facebook also aims to use commoditized hardware, which will require a flourishing chipset/device ecosystem and economies of scale and is not expected to be out until 2020.

However, Terragraph’s biggest problem—assuming it will reach commercial status—will not be hardware or software but the physical challenges of deploying infrastructure in the field. It is also the same reason urban small cells have not taken off as expected, since placement, power, and backhaul were significant barriers to reach deployments of thousands of small cells. Exactly the same challenge will face Facebook’s deployment; this is something Verizon is currently assessing with Samsung for 5G at 28 GHz for FWA.

In summary, ABI Research does not expect Terragraph to compete head on with Verizon’s and AT&T’s 5G FWA use cases. It may create some competitive pressure through media attention, but Facebook does not have the necessary logistical and operational skills to deploy Terragraph en masse, and advanced software cannot reduce the cost below what operators can achieve.