SpaceX Wants to End Rural Subsidies

SpaceX has now told the FCC multiple times that the agency should begin the process of phasing down and eliminating all rural high-cost subsidies from the Universal Service Fund. The company filed comments asking for the end of subsidies in FCC Docket 26-96, which is looking at reforming the High-cost fund for an all IP Future, and in Docket 10-90 that was looking at Universal Service Fund reform. In both filings, SpaceX says that its proliferation of low-orbit satellites means that rural subsidies are obsolete and are no longer needed. The company characterizes anybody who disagrees with it as a “subsidy-dependent incumbent”. SpaceX goes on to argue that the money currently spent on rural subsidies should be redirected to programs that lower the cost of broadband for low-income families.

It’s an interesting policy question that raises more questions about satellite broadband than it does about the rural companies that are receiving broadband subsidies. For example, the CAF II subsidy is being provided by a number of rural ISPs who use the subsidy to justify building and maintaining rural fiber, similar to what was done with the FCC RDOF.

I think the SpaceX request raises multiple policy questions. First, is satellite really providing broadband? Earlier this decade the FCC declared that Starlink was not providing broadband when the agency canceled the $885 million award Starlink had won in the RDOF reverse auction. At the beginning of this year, Ookla said that only 45% of Starlink customers were meeting the FCC’s definition of broadband at 100/20 Mbps. But Starlink’s speeds have been slowly improving, and will be boosted over the next five years as the company replaces the current V2 satellites with the new V3 satellites that provide 10 times more download capacity and 22 times more upload capacity.

As I have argued many times, the FCC’s 100/20 Mbps standard is no longer a reasonable definition of broadband. Recent data from OpenVault showed that 43% of U.S. households now subscribe to speeds faster than 500 Mbps download and 70% of homes subscribe to speeds of 200 Mbps or faster. It’s hard to look at these statistics and think that 100 Mbps is still an adequate definition of a good download speed. However, the real issue for satellite broadband is the upload speed. A large percentage of the homes cited by Ookla as having fast download also get far faster upload speeds than 20 Mbps. Fiber overbuilders are mostly providing symmetrical data speeds, meaning gigabit upload speeds. Cable companies across the country are upgrading the upload path and are delivering speeds of 50-100 Mbps at a minimum, and in some networks, gigabit upload speeds. The whole country learned during the pandemic that 20 Mbps upload is not adequate to support more than one user at a time – something the FCC has steadily ignored.

Another issue raised by the SpaceX request is whether Starlink provides a good substitute for voice service. Much of rural America lives in areas with little or no cellular service. The companies that receive high-cost subsidies all must pledge to offer voice service so that rural residents have access to 911. SpaceX argues that anybody with a satellite broadband connection can buy VoIP. That’s a true statement, but it raises two more policy questions of availability and affordability.

The companies receiving high-cost funding agree to service every home and business in their footprint. It’s pretty clear that Starlink can’t make that claim. While the company won’t disclose the number of homes it can serve in any given geography, there are examples, like last year’s Burning Man, where Starlink got overwhelmed by too many customers in a rural area.

I think the most important factor is affordability. Starlink’s price of $130 per month is a lot higher than other rural broadband options. Companies that receive high-cost support sell broadband for a lot less than $130, but more importantly, will sell voice-only to customers who don’t want broadband. It’s not reasonable that a rural household that can only afford voice and not broadband should have to spend $130 per month for satellite service.

In my mind, Starlink has to offer a different set of products if they want to declare that they can match the benefits that companies that receive high-cost funding bring to customers.

  • Starlink needs faster upload speeds, which it probably will eventually get as it phases in the V3 satellites.
  • Starlink would have to be able to serve everybody in a geographic area if it wants to see the subsidies end in a given area. I can envision a regulatory test similar to the one that cable companies have used to get regulators to declare a market to be competitive. High-cost subsidies should only stop in an area where every customer in the area could be served by a satellite instead.
  • Finally, Starlink would have to be a lot cheaper. If SpaceX wants to see the rural high-cost subsidies ended, it should offer lower broadband prices in those markets, including a really low-cost connection for voice only.

Growth of Mobile-Dependent Households

Roberto Gallardo, an Associate professor at Purdue University, has released an interesting policy paper, An Evolving Digital Divide: Measuring and Understanding Mobile-Dependent Households (MDH). The paper summarizes Gallardo’s research, using U.S. Census data, to compare households that rely on cellular for broadband connectivity in homes before and after the pandemic.

The paper cites several interesting statistics. Data from the U.S. Census American Community Survey (ACS) shows that the percentage of homes with no Internet access decreased from 17.6% in 2017 to 6.5% in 2025, meaning that 12.5 million homes gained broadband over that period. However, over the same period, the percentage of homes that only access the Internet with a smartphone increased from 7.5% in 2017 to 11.5% in 2024, an increase of six million households. The study also looked at device ownership. The percentage of homes that only have mobile devices (no laptops or desktops) in 2017 was 9% of 2017 and increased to 15.2% by 2024, an increase of 9 million homes. These statistics show that, while home broadband usage has increased dramatically by 12.5 million homes from 2017 to 2024, a lot of that gain was from the growth of mobile-dependent homes.

Over the years, I recall seeing articles in the popular press that speculated about these trends and concluded that people are increasingly happy with mobile broadband and don’t need a big-screen broadband option. Gallardo cites other studies that come to a different conclusion. He cites a study by Napoli and Obar in 2014 that coined the term “mobile Internet underclass” to describe the inferior nature of broadband access through a cellphone using the small screen and data plans with small data caps. That study concluded that treating mobile access as equivalent to fixed broadband risks rendering the growing mobile-only population invisible. He also cites a study by Quaglione, Matteucci, Furia, Marra, and Pozzi in 2020 that says that mobile broadband is a good complement to having home broadband, but is not a substitute for home broadband.

Gallardo goes on to undertake a deep data dive into the Census data for 2019 and 2024. He looked to see if he could discern any differences in mobile-dependent households that are related to metrics like metropolitan versus rural, race, those between 18 and 34, those over 65, households employed in blue-collar occupations, household incomes, home ownership, and homes with no Internet access available.

Some of the most interesting findings from the data related to mobile-only households include:

  • Metropolitan areas saw a significant increase in the percentage of MDH households.
  • There was no statistical change for those between 18 and 34.
  • There was a decrease in the percentage of MDH households over 65
  • There was an increase in the percentage of MDH for homes with no broadband availability.
  • There was a drop in the percentage of MDH in blue-collar households.
  • There was a higher percentage of MDH in low-income households.

Gallardo’s research didn’t look at the causes of the shifts, but it’s not hard to speculate about the reason behind some of the statistics. The increase in mobile-dependent homes in urban areas and in low-income homes is likely mostly linked to affordability. The decrease in mobile-dependent homes in those over 65 and blue-collar workers might be due to increased home broadband availability,

Gallardo reaches some interesting conclusions based upon the statistics. First, infrastructure alone does not solve the digital divide issues that push homes to use mobile broadband instead of home broadband. He concludes that if the goal is to improve access to broadband, then infrastructure construction should be paired with demand-side intervention programs like providing devices, digital literacy training, and affordability programs. At least for now, much of the digital equity funding is in limbo, and we saw Congress allow the ACP program to die on the vine. Gallardo says, “Without affordability support, the path of least resistance for many households is a mobile plan.”

Fending Off the End of the World

By now, everybody’s seen the recent announcement that there is a possibility that AI could lead to the end of mankind. While it’s likely that this announcement is mostly hype, there are some worrisome characteristics of AI that have to be taken seriously. For example, the hack of Hugging Face showed us that AI agents can be used to break through firewalls and hack into sophisticated networks. Many security experts have been bracing themselves for the pending introduction of hacking using quantum computers, but it looks like an equally terrible threat might already be here.

It’s worth pointing out that AI agents are not anywhere close to being self-aware and are not going to run around on their own hacking networks. Somewhere behind every hacking effort is a person who set the AI agents loose. Unfortunately, whether the origin of the hacking is a malicious foreign government or an overambitious teenager, the results can be the same. While the headlines are about the possible end of the world, the real immediate threat is the ability of AI agents to disable the many trusted computer networks that are in the background of our lives.

If we have problems with AI hacking, it’s not going to look like Skynet from The Terminator. Instead, it’s going to look more like Live Free or Die Hard, where commercial networks, banking systems, and traffic lights stop working reliably. That’s going to feel a bit like the end of the world, because the things we rely on every day are at risk of being broken. Anybody who has been through a major natural disaster knows what that feels like. I’ve had first-hand experience with two catastrophic hurricanes, and it’s daunting when power, water, cellular, and broadband all die, because losing those things kills other important functions like public safety and the food supply chain. Let’s face it – we are totally dependent on cloud systems. Look at the panic that ensues when cloud programs go out of service for a few hours.

I want to examine the possibility of AI hacking by looking at the example of hacking the least-connected network we all rely on – the water system. Water systems use SCADA (Supervisory Control and Data Acquisition) software that lets water system operators collect real-time data from field sensors and lets operators adjust flow rates, tank levels, and water pressure. Wells and pump stations use PLCs and RTUs , which are computers and software that automatically react to local sensor data. Water processing and sewerage plants have become increasingly automated.

If you go back just thirty years, practically nothing in a water system was computer-assisted outside of some customer billing. But we’ve been busy over the last several decades introducing software that increased efficiency and reduced the need for people. It’s ironic that, faced with the possibility of AI hacking, the improvements we’ve implemented in recent decades might now become the biggest vulnerability. We’re all vulnerable if our water network is disabled.

The obvious fix for AI hacking is to isolate a network from outside interference. We’ve been trying to do that for the last decade by placing networks behind firewalls and using a range of cybersecurity measures to fend off hackers. The scary scenario we’re suddenly facing is that those cybersecurity measures might be worthless if the hacker is a series of AI agents. I think the announcement by AI executives that AI could indeed go amok is really an admission that AI can likely overcome existing cybersecurity measures and break through firewalls.

If we can’t rely on existing firewalls and other cybersecurity measures, there are two alternatives. The old school way would be to go retrograde and return to the practices of thirty years ago when there were no computers and software anywhere in a water network. That’s pretty extreme, and I suspect that most water operators wouldn’t know how to go about disassembling computer controls without crashing the water system.

The other fix is to completely air-gap a water network, which means completely eliminating all outside connections to the system. That’s also a drastic step. It would mean cutting broadband lines and disabling the ability of smartphones or tablets to interface with the network. It means eliminating all cloud-based software and loading all software needed to operate the network locally. It probably means building a small data center to control the network locally. It means cutting off electronic links to neighboring water networks. It means more manual effort to accomplish tasks that have been automated.

If I operated a water or electric utility, I’d be having this conversation soon. The warning that AI could bring my current network down might means that all my current cybersecurity measures might be inadequate. The recent announcements of rogue AI agents, even if mostly hype, change everything related to network security.

A New Type of Churn

I think every ISP I’ve talked with this year is seeing higher churn. For those not familiar with the term, churn is the measure of how many customers an ISP loses. Churn is inevitable. It’s churn when a customer moves and drops service. It’s churn when a customer changes to another ISP. It’s churn when a customer doesn’t pay their bill, and the ISP cuts off service.

The industry churn rate varies widely by the size of the ISP and by markets. The biggest national ISPs have a churn rate in recent years between 1% and 1.5% monthly. The three large cell carriers have churn just under 1% monthly. These large businesses know that a lot of their churn comes from competition. When customers have multiple choices for satisfactory broadband or cell service, they are more open to changing providers to react to special pricing.

Smaller ISPs typically have far smaller churn rates. I know a number of small fiber ISPs with historic churn rates between 4% and 6% annually. Many of these fiber ISPs don’t face competition across their entire footprint, although almost everybody has some competition these days. Every ISP has customers that move away or die. There are always exceptions to the rule, and I know ISPs operating in college towns and near military bases that have always experienced much higher churn than average.

Many ISPs have programs to try to reduce churn. They might have a win-back program to try to talk customers out of changing to a competitor.

The headline for the blog is a new definition of churn. Some ISPs who have been trying to understand higher churn have dug deeper into the data. What they are telling me is that most of the increased churn is due to household economics and customers who can’t afford to pay the monthly broadband bill. Nonpayment has always been a big component of churn, but it is now moving to the top of the list.

Most households want to keep the broadband connection, but in tough economic times they are going to value keeping food on the table more than broadband. We can’t forget that dropping broadband can be a significant savings for a household, particular when it also means dropping streaming services, which might cost as much or more than the broadband bill. Most people who drop broadband have some partial alternatives like using cellphones or computers at the office.

When ISPs dig deeper into economic churn, they are finding something new. They are finding that people who drop broadband because they can’t afford it want to come back. They might pay for broadband for six months and drop it for three or four months. When they are flush again, they come back. That’s a new category of churn that we haven’t seen much of in the past. These are households that value the broadband connection, but who just can’t afford it. If I had to coin a name for this new class of churn, it might be something like occasional customers.

If you look around the world, there are models for selling to occasional customers. ISPs in places like Nigeria sell broadband by the day, week, or month because they understand that many households cannot commit to a steady subscription. ISPs there understand that getting some revenue out of a connection over the course of a year is far better than getting nothing. If an ISP drops a customer permanently, they strand the cost of the drop.

There are a number of ISPs in the U.S. who recognize this phenomenon to some small degree. For example, some ISPs offer seasonal rates for snowbirds or college students who are only in the market for part of each year. But I haven’t heard of any ISP that makes it easy for customers to come and go when they can afford to connect. In fact, many ISPs make it hard for customers to come back.

One of the good things about the way that most ISPs sell broadband is that they make customers prepay for the coming month. One way to deal with occasional customers would be to sell clearly-labeled pre-paid broadband that automatically expires at the end of the billing period if the next monthly bill hasn’t been pre-paid. Customers with this plan could come and go as they can afford to pay.

While there have always been occasional customers, what’s new today is the large number of customers who are struggling to pay their broadband bill. There are a lot of smart marketing people in the industry, and I’m sure some of them will find clever solutions to the problem. Rather than shun customers who have trouble paying, ISPs should provide a way for them to pre-pay when they can afford it. That’s way smarter than just losing the customers to permanent churn and stranding network costs.

Satellite Shorts September 2026

AST SpaceMobile is making progress with offering satellite cellular service. The  company launched three massive satellites in August, bringing the fleet to 13 satellites. These are much larger than ASTs original satellites or ones being used by other carriers. They are essentially floating central offices. The FCC also gave the company the go-ahead in August to begin testing its satellite-to-smartphone service. AT&T and Verizon have both inked deals with the company to be their satellite partner for cellular.

More Countries Entering the Satellite Race. STC Group has partnered with Astranis to deploy a dedicated satellite to serve Saudi Arabia. The country wants to strengthen its communications capabilities independently of the other big satellite fleets. This has become a priority for Saudi Arabia, as a complement to its major investment in terrestrial fiber networks.

The Vietnamese aerospace company VinSpace has signed a contract with SpaceX to launch its first satellite into orbit. The country hopes to develop a space industry and deploy multiple satellites in the future.

Deorbit-as-a-Service. The U.S. Space Development Agency recently chose designs to consider from three companies that have business plans to launch satellites that will capture and deorbit dead satellites left in orbit. This is a nascent industry with challenges, because nobody has ever been able to snag a dead satellite. Some people in the industry describe most existing satellites as “ungrabbable”.

More Spectrum for Satellites. The FCC is scheduled to vote at the end of September on providing more than 1,000 MHz of spectrum in the 12 GHz and 42 GHz bands for high-speed satellite broadband. The spectrum could be used for home broadband, airplane and ship connectivity, and traffic-routing between satellites and earthstations. The FCC seems to be approving new spectrum for satellite uses at a much faster pace than it ever considered for other uses of spectrum.

Next-gen Starlink Satellites. In July, SpaceX launched twenty of the next-generation V3 satellites as a test of the satellites’ solar arrays and antennae. The satellites were allowed to deorbit soon after the test. The company currently has scheduled a launch on September 28 of the first operational V3 satellites. The launch has been delayed from the original goal of the first quarter of 2026. The satellites are much larger than the V2 satellites and require the Starship rocket to launch. Starlink says the new V3 provides 1 terabit of download and 160 gigabits of upload speeds, a huge improvement over the V2 satellites. The use of the Starship rocket will also increase launch capacity. In the past, the maximum number of V2 satellites per launch was 29, but with Starship, up to 60 V3 satellites can be launched at a time.

Europe Moves Forward with IRIS2. Europe is collectively launching hundreds of independent LEO and MEO satellites to create a European sovereign satellite network. All vendors for the project are European, with  Airbus, Aerospacelab, Thales Alenia Space, and OHB building the satellite constellation, and Eutelsat, SES, and Hispasat handling ground functions. IRIS2 is not intended as a competitor to Starlink for home broadband, but will be used to provide secure satellite connections for government, defense, and emergency purposes. This seems to be part of the larger movement for Europe to wean itself from American tech companies.

Connected Vehicles. Toyota and Iridium have done a demo, in conjunction with Deutsch Telekom IoT, to send voice messages to and from cars using the Iridium satellite network. This was the first test of the technology the company wants to use to provide connectivity for automotive, logistics, remote utilities, smart agriculture, emergency response and other IoT applications in areas that are out of the reach of cellular networks.

Danger of Space Collisions. Darren McKnight, a space debris expert at LeoLabs in California, said recently that we are already ripe for a major collision event in space. He says the biggest danger comes from the orbiting derelict rocket bodies that are still in orbit. These mostly sit at altitudes a little higher than LEO satellites, and a collision of rocket remains would rain debris down on all satellites below.

Fixing the Broadband Maps

There is a lot of talk about improving the FCC broadband maps. I published a blog a few weeks ago that described a bill passed by the Senate that would require a new review of the FCC broadband maps. The bill is labeled the Modernization, Accountability, and Planning (MAP) for Broadband Funding Act.

I’m all for changes that would create an accurate broadband map. But I’m fairly certain that the FCC is not going to consider making the changes that would actually improve the maps. The biggest problem with the current maps is that speeds claimed by ISPs are sometimes far different than what is being delivered. Fixing that would require the FCC to actually look at what’s in the maps and challenge ISPs that exaggerate speeds. I can’t imagine the agency adopting new rules that would saddle itself with that responsibility.

There is an existing tool that could be used  to compare claimed speeds to actual speeds. Ookla records millions of speed tests across the country for all ISPs. The FCC could buy the Ookla speed test records to compare against ISP claims. With AI and other new analytic tools, it wouldn’t be hard to compare Ookla speeds to FCC claimed speeds.

If speed tests were considered, ISPs would instantly yell that speed tests aren’t accurate enough to use to judge them. In many ways, they would be right. People often take speed tests when their broadband isn’t performing right. There are lots of homes where speeds are slower due to WiFi issues. A computer sitting close to the WiFi router will show faster speeds than a computer at the other end of a home. Many customers subscribe to lower speed tiers, which, by definition, don’t show the capability of the ISP.

But when large numbers of speed tests are used correctly, they are a valuable tool for understanding an ISP’s real performance. I’ve been lucky enough to work with counties that have purchased the full Ookla speed data set. Those tests provided them with enough data to do meaningful comparisons.

One of the things I learned in working with speed test datasets is that you must concentrate on the fastest speeds rather than average speeds. Doing so eliminates most of the concerns I discussed above. If an ISP claims data speeds of 100/20 Mbps to the FCC, but there are no speed tests over the course of a year that hit that speed, the ISP is exaggerating speeds. This kind of analysis can be used to ferret out exaggerated speed claims. This also worked in both directions – I’ve often seen faster speeds than the speeds claimed to the FCC, which means the ISP is underreporting speeds.

Speed tests can also be used to test broadband coverage claims. I was working with one county where a WISP was claiming coverage over more than half of the geography of the county. The Ookla speed test data for that ISP showed only a single speed test over the course of a year. It was clear that the WISP’s coverage claims were aspirational, not actual, and the County was able to use the Ookla results to challenge the broadband maps in the first BEAD map challenge. Without the Ookla tests, the County could not have made a map challenge, because the NTIA challenge rules wanted the County to find customers using the ISP in question and convince them to take speed tests. It’s obviously impossible to test an ISP that is not actually present.

Speed tests show all sorts of other interesting facts about ISP coverage. I’ve written several times in the past about how the speed for FWA cellular wireless quickly diminishes with distance from the tower. An FWA carrier might be delivering 300 Mbps to a customer within a short distance of a tower, but after a mile, speed drops to 100 Mbps. By two miles speed drops to 50-75 Mbps, and at three miles drop to the range of 25 Mbps. Most FWA speed claims in the FCC map will show 100 Mbps across the entire area around a tower, and much of that speed claim is overstated.

The FCC already spends a lot of money annually trying to get the mapping fabric right. But it’s pretty clear that the FCC doesn’t look at the speed data claimed by ISPs. The only time the FCC uses this data is when it published its annual report to Congress about the state of broadband. The FCC benefits by overstated maps since they tell a better story than what’s really available in rural America. If the FCC really wants to get serious about broadband mapping, why not buy the Ookla data twice per year to compare to claimed speeds?

Love Those Volunteers!

Today’s blog talks about volunteerism in the broadband/telecom industry. I was prompted to write the blog when a friend sent me the picture to the right of cookbooks published as a fundraiser by the Louisiana Chapter of the Telephone Pioneers of America. The Telephone Volunteers was a non-profit created by AT&T in 1911, first led by Theodore Vail, the famous AT&T president who oversaw the big growth of the company. The organization claims 800,000 members today. The Pioneers are organized into state and local chapters that tackle a wide range of local charitable causes. When I worked at Southwestern Bell, they gave employees one day off each year to participate in a non-profit activity organized by the Telephone Pioneers.

I know a lot of readers of this blog work for non-profits, and I know a number of people who donate time to broadband-related charitable causes. I’m involved in two non-profits in North Carolina, but this blog is not about me, but about the idea of using our accumulated industry knowledge to give back to the community.

I happen to live in an area in Western North Carolina which has very active broadband-related activities that are always looking for volunteers. There are groups here who refurbish smartphones, computers, and tablets to provide to those in the community who most need them. There are groups here who do hands-on digital training on how to use devices and apps, and how to navigate the Internet safely. There is a group here who provides smartphones and support for those getting out of long-term incarceration. There is a non-profit that advocates for policies that support broadband access. There is a lot of volunteer work in this region dedicated to resiliency and figuring out how to be better prepared for the next big storm.

I’m aware of many other volunteer efforts around the country. I know several people who help their local school systems navigate the E-Rate program to get funding for connectivity. There were a lot of people who volunteered to get households subscribed to the ACP program before Congress let the program lapse. I know several people who have been working for years to get affordable telephone rates in jails and prisons. There are several major cities with ad hoc wireless networks cobbled together by volunteers that provide broadband to households who can’t afford it.

I saw mountains of volunteers step forward in a lot of counties when they saw the opportunity to get faster broadband. There were people distributing and processing broadband surveys to understand community demand. There were a lot of people who worked on getting people to take speed tests for the map challenge processes. I witnessed a new cooperative being formed that was largely accomplished by dedicated volunteers. I remember working for cities where a lot of volunteers stepped up to pass ballot measures that authorized municipal broadband in their city.

Volunteerism is becoming increasingly important. Federal and state monies that were used to help people afford broadband have died or been cut back. I know many readers of this blog have a vast amount of telecom experience. If you have talents and haven’t thought about volunteering, you should consider it. There is a lot of fun to be had sitting in a room with a bunch of other tech geeks rebuilding computers. There is a lot of satisfaction from helping somebody learn how to safely use the Internet. You’ll probably never have more fun than an evening teaching computer safety to a group of older adults. No matter where you live, there is somebody nearby tackling broadband issues. If not, then start the effort!

Using Unlicensed Spectrum for D2D

The FCC recently adopted a Notice of Proposed Rulemaking that asks if it makes sense to give satellite providers access to unlicensed WiFi spectrum to support Direct-to-Device cellular service. The NPRM specifically looks at the use of 2.4 GHz and 5.8 GHz. The NPRM specifically rejected the idea of using 900 MHz for satellite service.

The FCC seems all-in on the idea of satellite cellular service and has taken a number of steps recently that favor this budding industry. However, Commissioner Anna Gomez, the only Democratic commissioner, voiced concern about how using this spectrum might affect the many devices that rely on WiFi spectrum today.

It seems like a valid concern. WiFi has turned out to be the most important and valuable spectrum we have. The whole broadband world seems to run off WiFi modems in homes and businesses. We have outdoor WiFi hotspots to provide outdoor broadband access. WISPs largely rely on WiFi spectrum to deliver fixed wireless broadband to rural homes.

It’s easy to understand why the FCC would consider this. The agency is facing huge demand for spectrum from multiple industries. Cellular carriers want more spectrum to get ready for the upcoming introduction of 6G. Cellular carriers also need to feed the rapidly growing FWA home broadband industry. Satellite companies have an insatiable demand for spectrum since they envision eventually launching hundreds of thousands of broadband satellites. And now, the satellite cellular companies want to blanket the Earth with competing networks to deliver outdoor cellular service, and the Direct-to-Device (D2D) carriers have already requested approval to launch tens of thousands of satellites to provide outdoor cellular service. There are more requests for spectrum than what is available, so the FCC seems willing to raid the public WiFi bands to support a new industry.

This is not the first time that the FCC has allowed others to share WiFi spectrum. A number of years ago, the FCC allowed cellular carriers to use WiFi spectrum in busy markets. I recall a report from university students in Chicago who had documented that cellular usage of WiFi had killed outdoor WiFi on their campus. I don’t know if cellular companies are still dipping into WiFi now that they have been able to get more spectrum, but this was a failed effort that favored the carriers over the normal users of WiFi.

What seems to be different today is that the FCC is unabashedly pro-satellite. They seem to be giving satellite companies everything they ask for. It’s not unusual for regulators to make it easy for new technologies – the FCC took a hands-off approach to the creation of the Internet. But there seems to be a lot of risk in allowing satellite companies to broadcast WiFi from space. This idea is clearly favoring satellite carriers over WISPs.

I talked to several wireless engineers about the proposed use of WiFi spectrum, and they all voiced concerns. The satellites would blanket the countryside with new WiFi spectrum. The engineers told me that it is impossible for this to be completely neutral for a WISP. Every outdoor receiver will perceive the new use of the spectrum as interference. They said there was no way to tell how bad this might be without trying it, and that sounds like a scary experiment if it turns out that the impact is severe. It’s easy for the FCC to allow a new use of spectrum, but it’s unlikely that it would be undone quickly if there is a problem.

This is out of character for the historically slow and cautious approach that the FCC has always taken with spectrum ideas. And that’s even more the case when we’re talking about WiFi spectrum, which has proven its value many times over.

Remote Work Is Here to Stay

Working from home is still one of the most important uses of broadband, particularly in rural communities. Roberto Gallardo, who writes in Substack as the “Data Whisperer”, recently wrote an article discussing the change over time of those who work from home. For those who don’t know Roberto, he’s an Associate Professor and the Vice President for Engagement in the Agricultural Economics Department at Purdue University.

Everybody remembers the explosion of working from home during the COVID-19 pandemic. A lot of people decided that they liked working from home, and the percentage of those doing so has not dropped to anywhere near pre-pandemic levels.

The article starts out by comparing the percentage of those who worked from home in 2014 and 2024. Overall, the percentage of those who are working from home more than tripled over the decade. In 2014, 4.4% of the U.S. labor force worked from home, and in 2024 that had grown to 15.1%.

There is still a big disparity in the percentage of home workers in urban versus rural areas. In 2024, 16.1% of urban residents work from home while around 8.4% of those living in rural areas work from home. Gallardo credits the difference to two reasons. First, there is a higher percentage of jobs in urban areas that can be done from home. He says that over 36% of jobs in urban areas could be done from home, while only around 28% of jobs in rural areas can be done remotely.

But there is also still a difference due to the lack of good broadband in rural areas. While there has been a lot of grant funding aimed at rural America, a lot of it hasn’t been built yet. Even where there is broadband that supposedly meets the FCC definition of broadband, there are many rural ISPs that are not delivering a good enough upstream path to reliably work from home. For example, a recent Ookla report said that only about 45% of Starlink customers have speeds that meet that FCC speed definition.

One of the most interesting parts of his article is an interactive map that shows the percentage of people who work at home in each county in 2014 and 2024 compared to the number of jobs that could be done remotely. Gallardo acknowledges that a lot of remote workers work for somebody in another county or state, but this is still an intriguing map that shows the level at which remote work has become an important part of local economies. The map shows that remote work has been more accepted in New England, the Pacific Northwest, Colorado, and the Mid-Atlantic. Interestingly, the states with the highest level of remote work in 2014 had bigger gains by 2024 than other states.

As most people will do, I looked at my own county in North Carolina first. The county had 15.4% of eligible workers working remotely in 2014, which grew to 34.8% in 2024. I can attest that remote work is common here, and there are numerous people on my own block who work from home. But examining the map shows a lot of counties where the percentage of those working from home is below 10%, some at only half that level.

I’ve heard stories about how getting good broadband can transform communities. I talked to a cooperative in Alabama who told me that there were practically no rural home workers before fiber was built, but that many homes in the coop territory now have people working remotely. They said that higher incomes from remote work are transforming the economics for families and the local community, and that the fiber is also attracting some people who already work from home to move to the county.

Starlink’s Other Network

The Benton Institute recently wrote an article that reminded me that Starlink is operating a second satellite network, separate from the one used to serve residential and business broadband customers. The company announced this network, which it dubs Starshield, at the end of 2022. Today, this network is now highlighted on the SpaceX website.

Starshield is touted as a military-grade network, which in layman’s terms means a separate network that doesn’t share any traffic with commercial customers. In the U.S., the National Reconnaissance Office (NRO) has launched several hundred Starshield satellites. This is the federal agency that has always operated military and spy satellites. The British Ministry of Defense is also said to be migrating to the Starshield platform. Other militaries around the world that use SpaceX, like the well-known use of Starlink in Ukraine to direct drone traffic, use the normal commercial Starlink satellite platform. Benton reports that Starshield and Starlink are not yet fully separate, and that a Starlink outage affects Starshield performance, but it’s likely that the company is working towards total independence.

SpaceX touts several benefits of Starshield. They can use whatever level of encryption of traffic required by a government customer. The satellites can also carry custom payloads, which likely means spy cameras or other devices. SpaceX also says it can integrate Starshield into other satellite networks.

Since nobody I know will even be using a government spy satellite, you might wonder why I am writing about Starshield. I’ve been wondering lately about Starlink’s claim that it is interested in providing backhaul to cell sites and bigger broadband connections to rural businesses. This has always puzzled me because, at the local network level, a satellite network is no different than any other network. If Starlink were to connect to a handful of cell sites in a given rural area, the bandwidth used for the cell sites would reduce the capacity available to residential customers. There is only so much bandwidth available in a localized satellite footprint.

We know this is true, because even today, there are stories of localized Starlink areas getting overwhelmed. There were reports during the Burning Man festival last year that nearby Starlink customers complained about loss of bandwidth when festival attendees brought along their Starlink Mini satellite dishes.

This makes me wonder if Starlink plans to have separate satellites just for backhaul. We know what that might look like. OneWeb, owned by Eutelsat, provides big bandwidth to European government locations and large businesses. The company currently has around 650 satellites in orbit to support bigger bandwidth users.

It wouldn’t be hard for Starlink to carve out satellites that specialize in large customers. The company currently has 11,100 satellites in orbit and should be close to its original approved constellation of 12,000 by the end of the year. The FCC has already given the company permission to grow the constellation to 15,000. And since SpaceX is never satisfied, it’s already asked the FCC for permission to double the constellation to almost 30,000 satellites. It doesn’t seem like it would be much of a challenge for Starlink to carve out 500 satellites from the large constellation to specialize in backhaul. This would not only open up new revenue opportunities but would also mean not diluting bandwidth delivered to its millions of residential customers. This is all pure speculation on my part, but its an idea that makes sense.