Direct-to-Cell Innovation Will Expand Broadband Access to All Americans

On November 26, the FCC issued an order that granted low-earth orbit (LEO) satellite broadband provider Starlink authorization to provide Supplemental Coverage from Space (SCS) and operate on certain spectrum bands for direct-to-cellular (direct-to-cell) operations, under certain conditions. Direct-to-cell, sometimes called direct-to-device (D2D), is the technological capability of connecting satellite broadband networks to standard terrestrial mobile cellular wireless smartphones. Starlink reportedly has an agreement with nationwide mobile wireless network provider T-Mobile, under which it will provide mobile Internet connectivity in the US exclusively to T-Mobile for one year. 

Additionally, AT&T and Verizon reportedly have entered into commercial agreements with LEO satellite network provider AST SpaceMobile. AST SpaceMobile will be using spectrum in the 850 MHz band licensed by AT&T and Verizon, whereby AST SpaceMobile will provide direct-to-cell capability and thus enable mobile wireless broadband coverage to 100% of the geography of North America. 

As I wrote in a December 2023 blog post, smartphone access to satellite broadband networks is a stellar example of the broadband market's dynamism. Near-future commercial availability of direct-to-cell capability by competing mobile wireless broadband providers in partnership with LEO satellite network operators is innovative, enhances competition, and doubtless will improve access to broadband for Americans. 

 

Indeed, the important potential improvement in broadband access enabled by direct-to-cell innovation should factor into the FCC's forthcoming Section 706 Report as well as its forthcoming Communications Marketplace Competition Report. In assessing progress in deploying advanced capabilities in a reasonable and timely fashion to all Americans and in analyzing market competition for broadband services, the Commission should take a forward-looking analysis rather than rely on static snapshots in time from the past. 

 

For the Commission, direct-to-cell capability ought to serve as a reminder that private market investment and innovation drive the improvement and expansion of broadband networks far more than slow-moving subsidy programs such as the Broadband Equity, Access, and Deployment (BEAD) program that draw from the public treasury – and ultimately from US taxpayers. During the next Trump Administration, the FCC should return its focus to promoting private network investment and innovation and to eliminating rate regulation and other burdensome, costly restrictions that harm market competitiveness and fail to meaningfully benefit consumers. 

 

This year, the Free State Foundation filed public comments and reply comments with the FCC in its current Section 706 report proceeding. FSF also filed public comments and reply comments in the Commission's current Marketplace Competition Report proceeding. 

Maine Satellite Plan Casts Doubt on BEAD Program Approach

As reported by Broadband Breakfast (subscription required), the Maine Connectivity Authority (MCA) has announced that it will allocate upwards of $5 million toward the purchase of Starlink terminals for every remaining "unserved" location in the state. What's more, the contract with Starlink "is expected to include capacity guarantees to ensure that the state-purchased terminals can connect to internet service at the newly established speed benchmark of 100/20 Mbps."

Universal access? Check. Speeds that satisfy the FCC's recently updated "broadband" benchmark? Check. A price tag roughly one-twentieth that of fiber? Check. So why, then, will the MCA spend an additional $278 million in federal subsidies from NTIA's Broadband Equity, Access, and Deployment (BEAD) Program to connect "underserved" locations to fiber?

Despite their marketplace-proven ability, technologically speaking, to deliver the speeds that consumers demand – and that the Infrastructure Investment and Jobs Act, the statute that created the BEAD Program, specifies – from day one NTIA has discouraged the use of BEAD Program subsidies to deploy non-fiber broadband distribution platforms.

As I pointed out in "BEAD Program Technological Neutrality 'Fix' Falls Short," an August 2024 Perspectives from FSF Scholars, even recent changes to NTIA's BEAD Program rules approving the use of satellites and unlicensed spectrum do so only under very limited circumstances – to be specific, when the price tag for fiber exceeds a state-specified price ceiling aptly labeled the Extremely High Cost Per Location Threshold (EHCPLT).

Other distribution technologies – cable broadband, fixed wireless access using licensed spectrum, and so on – likewise are eligible for BEAD Program subsidies only if the fiber cost exceeds the EHCPLT.

The MCA's announcement that it will make Starlink terminals available to all 9,000 unserved locations in the state at a cost of just $599 per location, plus free shipping and professional installation, highlights the degree to which NTIA's approach leads to inefficiencies and waste.

Evidence that proves this point can be found in the very same press release announcing the purchase of Starlink terminals: "[i]n 2025, MCA will facilitate the investment of an additional $350 million in broadband infrastructure through the [BEAD] Program to serve the remaining 5% of locations in Maine that currently have slow and unreliable internet service."

According to my back-of-the-envelope math, $350 million in total subsidies works out to almost $12,000 per "underserved" location – that is, locations with Internet access at speeds equal to or greater than 25/3 Mbps but less than 100/20 Mbps. That amounts to a nearly 20X premium for a fiber-based solution as compared to the cost of satellite-based service.

Incidentally, Volume 2 of Maine's Initial Proposal, which was approved by NTIA in June, does not identify a specific EHCPLT. Instead, it indicates that the MCA intends at some point in the future to set the EHCPLT so high that, in virtually all cases, fiber will win the day:

If it is determined that a small number of locations in a given PSA should be served with alternative technologies allowed through the EHCPLT process to ensure maximum impact of BEAD funding, MCA will consider allowing non-fiber service to a minimal number of locations. All other locations in the PSA will otherwise be served by FTTH.

Accordingly, the extent to which satellites and other non-fiber distribution platforms will be eligible for BEAD Program funding likely will be extremely limited.

The BEAD Program's underlying congressional goal is to connect locations still without "broadband" – Internet access at speeds of 100/20 Mbps – in a cost-effective manner. The fact that, in Maine, far more federal taxpayer dollars will be spent on "gold-plated" fiber infrastructure to upgrade "underserved" locations than what is being spent to connect "unserved" locations strongly suggests that NTIA's approach is fundamentally flawed.

Satellite Broadband Competition and Access is Improving, and FCC Policy Should Promote That

A December 14 article in PCMag reports that satellite broadband provider HughesNet is now offering residential subscribers services with advertised download speeds of up to 100 Mbps. This is up from prior offerings of up to 25 Mbps and 50 Mbps downloads. The improved capabilities are the result of the high geostationary orbit Jupiter 3 satellite this summer, which apparently has now been tested and is ready for service. 

According to the FCC's 2022 Communications Marketplace Report, "[a]s of year-end 2021, satellite operators served a combined 1.7 million subscribers in the United States." And subscriber numbers for GEO satellite broadband services offered by HugheNet and ViaSat do not appear to have grown but have perhaps declined slightly in recent years. (GEO providers as well as LEO entrants were subjects of my March 2018 blog, "Satellite Broadband Services Will Enhance Competition and Reach New Consumers.") But HughesNet's satellite service upgrade is a shot in the arm to broadband competition, and the Jupiter 3 will better enable HughesNet's geostationary orbit (GEO) broadband service to compete with Starlink's low earth orbit (LEO) broadband service. 

 

Speaking of LEO broadband, PCMag also reports that a recent filing with the FCC shows that Starlink now serves approximately 1.3 million subscribers, or about 59% of the total satellite broadband subscriber base. And it is reported that Starlink recently received approvals to conduct testing of the cellular Starlink system that will transmit data to unmodified smartphones using T-Mobile’s licensed spectrum in the 1910-195 and 1990-1995 MHz bands. 

 

(Much, much more could be said about the FCC's treatment of Starlink, and expect Free State Foundation scholars to weigh in on that in early 2024. For now, one ought to consider reading the order released by the Commission on December 12 and the separate statements by its members, including Commissioner’s Brendan Carr and Nathan Simington)

 

Future commercial availability of smartphone access to satellite broadband networks is another example of the broadband market's dynamism. For further background, see my March 2023 blog post, "Big Announcements on Deployments to Direct-to-Device (D2D) Satellite Services." Importantly, these broadband innovations that enhance competitiveness and expand access to unserved and underserved Americans ought to be promoted with a light touch regulatory policy and not by turning those services into heavily regulated public utilities, which the Commission has proposed to do. FSF's comments filed with the Commission on December 14 of this year make the case against imposing public utility regulation on broadband services – including satellite broadband. In order to be able to ensure that all Americans have access to broadband, it is essential that the Commission promote competition and innovation by satellite providers, not suppress it. 

Starlink Hits 400k Subscribers, Introduces Portable Dishes

Starlink recently stated to the FCC that it has 400,000 global customers, nearly tripling from the 145,000 customers it had when I last blogged about Starlink in January. This figure includes subscribers across 36 countries, and while Starlink does not specify its US-based subscribership, its broadband service is currently available in 48 states.

Starlink also recently introduced a "portability" feature. For an extra $25 per month, Starlink subscribers can bring and use their satellite dish wherever they want. Portability should be useful for activities like camping, road trips, and generally improving Internet access in rural areas and other hard-to-serve terrains.

Image Copyright: Jud McCraine. Reshared without alterations.

Portability could increase broadband competition on the basis of product quality because it makes Starlink a hybrid between fixed and mobile offerings. The feature could differentiate Starlink from fixed broadband offerings by adding dual fixed-or-mobile capability to a single subscription. And the feature could also differentiate Starlink from existing mobile offerings by enabling mobility without the need to stay in range of wireless infrastructure.

Starlink's Performance Shows Prudence of Swift FCC Approval

Recent Ookla data measuring the performance of Starlink's low-Earth orbit (LEO) satellite broadband network shows the prudence of the FCC's swift approval of LEO constellations. SpaceX's Starlink, the only operational satellite broadband network in LEO providing home broadband, posted industry-leading speeds and latency for Q3 2021. Most notably, Starlink's speeds and especially its latency are generally comparable to that of many fixed broadband networks, showing that, once sufficient LEO satellites are deployed, LEO constellations, realistically, can provide high quality connectivity to rural and hard-to-serve areas where building physical networks is cost-prohibitive.

Starlink posted 87.25 Mbps / 13.54 Mbps median broadband speeds with a median latency of 44 ms, far outpacing competitors HughesNet and Viasat, which operate broadband networks at higher orbital altitudes than LEO. Starlink's speeds are over 4.5 times faster than HughesNet (19.30 Mbps/ 2.54 Mbps) and Viasat (18.75 Mbps/ 2.96 Mbps) for both download and upload. And while fixed networks have superior speeds, Starlink comes far closer to the median fixed network speeds of 119.84 Mbps / 13.54 Mbps than its higher orbit competitors, which offer speeds that are roughly a fifth or sixth of median fixed network speeds.

But latency is where Starlink especially stands out. Starlink's median latency, 44 ms, is about 15 times smaller than the median latencies of its satellite competitors, and only 2.9 times larger than the median latency for fixed networks. High latency – meaning signal delay – has long characterized satellite networks, and this makes them less attractive options than terrestrial and wireless broadband. Broadband with high latency makes certain applications that depend on split-second connections like video games, voice and video calling, and livestreaming less usable and reliable. As Commissioner Brendan Carr noted in a 2018 statement approving part of Starlink's constellation, "[LEO satellites] promise lower latency connections because they typically orbit only a few hundred miles above Earth, as opposed to many thousands." Starlink's network is delivering on that promise at an early stage.

Starlink's high speeds and low latency are important because satellites generally provide global coverage, which makes Starlink a serious contender for providing affordable, high-quality service in rural and other hard-to-serve areas. Users simply need to install a dish on their home to receive satellite broadband, avoiding the expensive and often cost-prohibitive infrastructure buildouts needed to reach rural and other hard-to-serve areas. This is precisely the consumer demographic Starlink targets with its marketing. Its website states: "Starlink is ideally suited for areas where connectivity has been unreliable or completely unavailable." So far, Starlink has 145,000 customers, a figure that's quickly grown from the 90,000 it reported in July 2021.

Ookla does note that Starlink's service performance varies by region in the United States, ranging from about 45 Mbps download in some areas to 145 Mbps download in others. But as Starlink continues to launch satellites, its coverage and network capacity will improve. Starlink has FCC approval to deploy 12,000 satellites, but so far it only has roughly 2,000 in orbit. SpaceX has already launched more Starlink satellites this month.

Even though Starlink hasn't deployed its full constellation, its broadband performance is already stronger than median fixed speeds in multiple developed countries. Starlink's median speeds for Q3 2021 in Australia, Belgium, Germany, and the United Kingdom outperformed median speeds for fixed networks in those countries. This is largely because fixed broadband is much slower in those places than it is in United States, according to Ookla's measurements.

Starlink, the first operational LEO constellation, first applied for FCC approval in 2016. Nothing guaranteed a swift approval, and the Commission could have chosen to wait longer before approving the application. Such delay often denies lifechanging technologies to Americans. Thankfully, the Commission approved LEO constellations for Starlink and potential competitors Amazon, OneWeb, Boeing, and others, including some that already have exited the market.

Free State Foundation Director of Policy Studies Seth Cooper supported Starlink's approval back in 2018. I previously highlighted a planned backhaul deal between Amazon and Verizon that will start once Amazon's LEO constellation reaches orbit in the next few years.

Starlink's early performance metrics suggest that the FCC was smart to quickly approve LEO satellites. The Commission's swift approval and support for market competition has allowed Starlink and potentially others to serve the unconnected Americans our federal government has spent billions trying to reach over the last decade.