The Growing Splinternet

From FlaticonThe term splinternet refers to Internet service in a country that controls or censors content available to citizens. The best-known example of a splinternet is the Great Firewall of China. While there is a lot of different software and platforms available to Chinese citizens, many web platforms from outside the country are blocked, and citizens all understand that anything they do on the Internet can be monitored.  China is not the only splinternet. For example, the Russian government restricts Internet access to only  approved sites in a lot of the country.

Iran has always controlled the Internet to some extent, but in recent months has entered the realm of full splinternet. This started with public protests against the government. Citizens could communicate inside the country, but only through government-controlled apps. The government blocked citizens from viewing foreign websites and from sending pictures and videos outside the country.

Internet advocates are warning that the splinternet is spreading. Wired recently had an article that says that China is now exporting the technology that support their censorship techniques for the Great Firewall. The article claims the technology has been exported to multiple governments around the world. This is going to make it a lot easier for smaller countries to achieve the same control of the web as achieved by China.

AI is making it a lot easier for governments to track what people are doing on the web. AI can also be an effective tool for blocking websites and can help a government to identify people using any software that does an end run around web restrictions. In the past, people found ways around government restrictions. I recall that protesters in Hong Kong became adept at coordinating and communicating by setting up ad hoc networks that bypassed government monitoring.

While not exactly a splinternet movement, there is a significant effort in Europe to create telecom and cloud infrastructure that is purely European. There is a lot of demand from businesses for cloud solutions that are independent and fully within European control. As much as anything, this movement is an attempt to avoid the large U.S. software companies that largely control the web around the world.

An example of this new direction is the consortium recently announced by Orange, Deutsche Telekom, Telefónica, TIM, and Vodafone, They have launched the European Edge Continuum, which allows customers to deploy applications that are restricted to only use the networks of the five providers. This is nearly the opposite of the approach being taken in the U.S., where ISPs hand traffic to hyperscalers that route traffic in ways that are unknown to the ISPs and users.

It’s becoming obvious that there is a downside in this country to a web that relies on a handful of hyperscalers. Corporations are increasingly frustrated when they experience major outages due to software problems in distant data centers that are out of their reach and control.

The trends are not encouraging. It’s hard to think there won’t be an increase in splinternet-like activities from governments around the world. When that’s coupled with people and corporations that want to minimize the use of giant hyperscalers, it looks like a further segmentation of the concept of an open web.

Broadband Subscribers 4Q 2025

I recently looked at the reported broadband subscriber counts from the largest publicly traded ISPs for the end of 2025. Most of these statistics come from the quarterly reports of the ISPs. The table below looks at the change in subscribers for the fourth quarter and also for the whole year of 2025.

The story hasn’t changed a lot during the year. FWA cellular ISPs still dominate new net customer gains in the industry. Cable companies continue to lose customers. Telcos are growing by adding more fiber customers than they are losing copper customers.

There are a few interesting stories in these numbers. There were a lot of industry predictions that FWA cellular sales had peaked and would be slowing. However, the fourth quarter is the largest net customer gain ever for the market segment.

The additions for the telcos mask that growth in fiber customers is being diluted by continued losses of copper customers.

Breezeline says that it has turned around losses. The company had been losing 7,000 customers per quarter, and for the fourth quarter, it was down only 1,000. The company has hopes of getting back to net growth.

I don’t know how much longer making this table will make any sense. Just in the last quarter, Ziply Fiber and Consolidated Communications fell off the table when they went private. In the first quarter, Frontier will merge with Verizon and Lumen’s fiber customers will shift to AT&T. A little later this year, Cox, which is currently privately held, will be added to the Charter numbers. There are a number of aggressive fiber overbuilders that are privately held and not included in the table.

Closing the Senior Broadband Gap

An article in Telecompetitor cites statistics from AARP that show there is still a substantial broadband gap among seniors. According to data from the Older Adults Technology Services (OATS) organization that is part of AARP, 19 million seniors didn’t have a home broadband connection in 2023. That’s not a big improvement over the 22 million seniors who didn’t have home broadband in 2018. The article cites a statistic from AARP that only 61% of seniors over 75 have a home broadband connection.

AARP notes that the reasons seniors don’t have home broadband are similar to the reasons affecting other demographics. For example, only 53% of seniors with less than a high school education have broadband compared to 68% of seniors with a high school or higher education. Incomes also matter, and seniors with incomes under $25,000 struggle to afford broadband. AARP notes some improvement in senior use of broadband from 2018 to 2023. They noted that more seniors are using large-screen devices in 2023. The biggest improvement for seniors was an improvement in the use of smartphones and mobile services.

Interestingly, the AARP statistics didn’t consider 5G home internet or 5G mobile plans as broadband access. AARP believes that 5G has issues like unreliable service, data caps, and other problems, which means that a 5G connection isn’t reliable for some of the services that seniors need, like medical monitoring. 5G capabilities are improving.

The big cell carriers have been adding spectrum and making electronics upgrades, which are bringing faster and more reliable speeds. However, one of the issues that isn’t discussed enough is that fast cellular broadband speeds are only good within about two miles of a cell tower. Most urban households are within a mile of a cell tower and can receive download 5G speeds over 100 Mbps, and often much faster. But in rural areas, a lot of people live more than two miles from a cell tower. In rural areas, there are still large swaths of areas with little or no cellular broadband coverage. Cellular service is also engineered by the big three carriers to have relatively slow upload speeds, and it’s not unusual, even in urban areas, for upload speeds to be under the 20 Mbps definition of broadband set by the FCC. It’s possible that AARP will eventually bless 5G as acceptable broadband, but even then, only for customers within a few miles of a cell tower.

The senior broadband gap is significant since seniors are increasingly reliant on broadband. Seniors are like the rest of us and are seeing essential services moving online. Money and banking have largely moved online. Local, state, and federal governments are moving forms and portals online, and often as the only way to communicate. Homebound seniors need access to online shopping. Businesses everywhere are converting to online portals. For example, portals are being used to communicate with mortgage companies, veterinarians, dentists, plumbers, etc.

The biggest need for home broadband for seniors is increasingly becoming health care. Healthcare providers are almost universally shifting to the use of online portals to communicate with patients. The portals are used to schedule visits, communicate results of medical tests, and answer patient questions. Anybody without the skill or ability to access medical portals is at a major disadvantage.

Seniors are also increasingly reliant on telemedicine, where they can have virtual visits with physicians without having to make an office visit. Broadband is also needed for monitoring medical devices that are increasingly used to track at-risk patients who suffer from diabetes or heart issues, and for general monitoring after patients are released from surgery.

I’ve talked to a lot of digital equity folks in the last year, and a lot of them are focused on seniors. These groups help seniors find affordable broadband connections and take advantage of any discounts or subsidies that can help pay for a connection. Digital equity folks are helping seniors obtain computers, tablets, and smartphones and teaching them how to use them. I find myself repeatedly ruing that the federal government decided to nix a lot of the grant funding to help with this effort. Luckily, a lot of the folks engaged in this work are finding ways to continue the effort without the federal funding assistance.

States Addressing Affordability

I recently wrote a blog about the new legislation approved in New Mexico that will fund a low-income program similar to the now-defunct federal Affordable Connectivity Program (ACP). The New Mexico program provides a subsidy of up to $30 per month to help qualified households afford a home broadband connection. The new LITAP program is being funded by a monthly surcharge of around $1.50 added to telephone, VoIP, and cellular bills.

State legislatures across the country are recognizing that good broadband adoption is vital for a robust economy and have tackled or are considering plans to make broadband more affordable.

In June of last year, the Oregon legislature passed HB 3148, which strengthens support for low-income broadband customers. The legislation increased the monthly subsidy for low-income households for broadband from $10 to $15 per month. On Tribal lands, the subsidy is increasing to $40. When added to the federal Lifeline subsidy, residents get a total reduction of broadband bills of $24.95. The program also allows a one-time subsidy of $100 to help low-income households buy a computer or tablet.

Like in New Mexico, the Oregon program is funded through the Oregon Telephone Assistance Program (OTAP), which is funded from the existing Residential Services Protection Fund (RSPF) that is supported by surcharges on retail telecommunications bills. All ISPs in the state are required to participate in the plan.

California also has a new low-income broadband plan. The California LifeLine Home Broadband Pilot Program is a three-year program designed to make home broadband more affordable. The program officially launched on January 26 of this year.

The California program offers a $20 discount for qualified homes to help pay for standalone broadband. The subsidy is $30 if used for a bundle of broadband and telephone service. The program also covers up to $39 to cover the installation cost of enrolling with an ISP. The California program is also funded by surcharges on customer bills through the California Lifeline program.

The Connecticut legislature also passed legislation that implements a plan in reaction to the end of the federal ACP plan and that creates affordable broadband for qualified low-income households. Connecticut took a different approach from the other states and is mandating that ISPs offer at least one affordable plan for low-income households.

The Net Equality Program requires ISPs to offer a broadband plan to eligible households for a rate of $40 that must deliver speeds of at least 100/5 Mbps for the first year of the plan and 100/20 Mbps after that. The plan goes into effect this month on a voluntary basis and becomes mandatory for ISPs by October 1 of this year. The $40 rate can be adjusted annually by the rate of inflation. The program is administered by the Office of Consumer Counsel.

If Congress is unwilling to tackle broadband affordability, then states will have to step up. I have to think other states are discussing the topic, and I’d love to hear from folks about other efforts.

Disaster Regulatory Policy

There has been a lot of talk over the last year at the state and national level about lowering the regulatory and paperwork barriers for ISPs building rural networks funded by BEAD grants. I don’t know any rural ISP that doesn’t think this is a good idea. Any efforts that can simplify processes like permitting, pole attachments, and environmental studies would be a big benefit for a rural construction effort.

There is another situation that probably warrants easier and simpler paperwork. Many communities have disaster and emergency plans that describe ways to ensure public safety and that government functions after a natural disaster. I think these plans should include ways to make it easier to restore electric utilities and communications networks after a disaster.

This topic came to mind as I am watching the stakeholders in Western North Carolina work to identify issues that hindered recovery and repair efforts after the region was devastated by Hurricane Helene. The damage from Helene was the worst imaginable for the communications networks. Power went down across the region. A lot of cell towers were knocked out of service. Over 1,700 miles of fiber were destroyed, including the backbone fiber paths that brought broadband to and from the area. Some roads were washed away by flooding, while many other roads were blocked by downed trees and landslides. Huge numbers of utility poles were knocked down or badly damaged.

As you might imagine, the local government’s ability to conduct anything resembling business as usual disappeared with the storm. Government communications were out like everybody else, and the local government properly turned its attention to public safety and getting people out of harm’s way.

Normal processes that would work after a less extreme disaster were not sufficient after this major disaster. For example, nobody at the local level was prepared to deal with the huge numbers of repair crews that poured into the area a few days after the storm. One issue that slowed down the out-of-town repair crews was the paperwork and contracts that had to be executed before a crew could begin to work. Local governments needed to identify incoming crews so they could eventually be compensated. Somebody had to verify that incoming linemen were certified to work in the power space. Incoming crews needed to provide proof that they carried the needed level of insurance. Even after an emergency, the local government wanted to make sure that crews understood any local ordinances and rules related to rights-of-way, safety, and other related issues.

Another issue that quickly arose was the paperwork and processes normally required for permitting and for rights-of-way. Permits are normally required for construction activities like digging up any portion of a roadway or erecting new utility poles. As every ISP knows, permitting means dealing with multiple government entities. The federal government controls Interstate highways, and in this region, the roads through national parks, forests, and the Blue Ridge Parkway. The State controls access to state highways and roads through State parks. The county controls access to county roads, and cities control access to city streets that are not under one of the other jurisdictions. This region is also full of privately owned roads in rural areas. There were also the usual complications associated with working to cross bridges and railroad tracks.

There were also a lot of other questions that are being raised. For example, does it make sense to replace a washed-away cabinet or hut in a flood plain without elevating it or locating it outside of the flood plain? What’s the right way to rebuild infrastructure where the old infrastructure was completely washed away by a flood?

Local governments, utilities, ISPs, first responders, and many other stakeholders in the region are working this year to understand what could have been done better. For example, a lot of the required paperwork for incoming crews could have been done online in the cloud. But that would have required some agency outside of the region to process the paperwork. It would mean incoming crews from around the country would have needed to know where to find such paperwork. Some of the normal requirements for paperwork could have been relaxed, and that probably means creating an ordinance that gives city and county officials the ability to declare a communications emergency that would ease rules for permitting and construction. Perhaps the biggest challenge of all is making any needed changes across multiple layers of government.

I’m impressed with the work being done in this region to look at improving our response to the next disaster. The discussions are going to continue throughout this year. The issues being discussed are things that local governments across the country should be thinking about – before they are struck by disaster.

Broadband Shorts March 2026

The following are a few topics I found interesting but which are two short to need a full blog.

Acquisitions Changing the Broadband Landscape. We’ve recently seen the closing of a number of major mergers and sales that are changing the broadband landscape.

  • On January 20, the sale of Frontier to Verizon closed. This $20 billion blockbuster sale brought 2.2 million fiber subscribers and eight million passings. Long-time followers of the industry are somewhat amused to see Verizon buy back millions of passings it sold to Frontier in the past.
  • On February 2, AT&T closed the sale of over 1 million fiber customers from Lumen, which brought four million fiber passings. This included customers in major markets like Denver, Seattle, Salt Lake City, Las Vegas, Minneapolis-St. Paul, Orlando, and Phoenix.
  • On March 10, the sale of Starry to Verizon closed. While bringing only 100,000 customers, the acquisition also brings Starry’s proprietary technology that uses 28/39 GHz millimeter wave spectrum to deliver wireless broadband, mostly to MDUs. The speculation is that Verizon will use the technology to expand to MDUs outside of its fiber footprint.
  • The huge merger between Charter and Cox Communications is still pending. The merger recently got approved by the FCC and still needs approval from several states. Cox would bring around 6 million broadband customers and 12 million passing to Charter, making the combined company the largest ISP in the country.
  • GFiber just announced a merger with Astound Broadband that would spin GFiber from Google Alphabet.

Action in the NDIA Suit. The U.S. Department of Justice sought to dismiss the lawsuit filed by the National Digital Inclusion Alliance (NDIA) that challenged the administration’s refusal to disperse the grant funding approved by Congress from the Digital Equity Act. These grants were aimed at tackling digital inclusion efforts that included bringing broadband devices to those that need them, training people how to use computers and broadband, and training for broadband-related jobs. The NDIA suit was first filed in early October 2025. I note the DOJ motion since the agency has had a low success rate in defending executive actions that killed various other federal grants. I think there is still a chance that this funding will eventually be awarded as intended.

AI Fueling Surge in Deepfake Spam. Hiya, a service that provides apps to block spam calls, released its State of the Call 2026 report, which says that AI is fueling an increase in spam calls. A survey of over 12,000 consumers across the  U.S., UK, Canada, France, Germany, and Spain showed a rise of deepfake calls, which use AI to mimic voices that are familiar to those being called. One in four Americans said they received a deepfake voice call in the last year. Americans said by nearly 2-to-1 that spammers are winning the battle over the FCC, which is trying to squelch spam calls.

AT&T Partnership with Amazon. AT&T, Amazon Web Services (AWS), and Amazon Leo announced a broadband collaboration this week that integrates AT&T into the AI and cloud capabilities of AWS. AT&T will become the preferred vendor to provide connectivity to AWS data centers. Amazon LEO has an existing arrangement with Verizon to bring fiber to ground stations, and it will be interesting over time to see if that business shifts to AT&T.

AT&T will partner with Amazon Leo to provide satellite broadband connectivity to some AT&T broadband customers. This is an interesting solution that could help AT&T more easily walk away from rural copper networks. AT&T also wants to bring satellite backup broadband to AT&T business customers.

Light Spectrum Licensing

There is an interesting spectrum battle going on between cell carriers and satellite companies. The heart of the contention is spectrum in the Upper Microwave Flexible Use Service (UMFUS) bands, specifically the 24 GHz (GigaHertz), 28 GHz, upper 37 GHz, 39 GHz, 47 GHz, and 50 GHz bands. These frequency bands are generally referred to as millimeter wave spectrum.

Satellite companies use some of this spectrum today via a shared arrangement with cellular companies that have purchased some of this spectrum in FCC auctions. Satellite companies are seeking greater use of this spectrum to communicate between satellites and ground stations, which is a growing concern as the number of different satellite providers and the overall number of satellites in the sky increases. Satellite companies want access to more shared spectrum using a process the FCC calls ‘light licensing’, which is a process to register new ground stations with relatively little paperwork.

The big cellular carriers don’t want any expanded use of light licensing. Like many spectrum disputes, at the heart of the issue is money. AT&T, T-Mobile, and Verizon each spent roughly $2 billion in millimeter wave spectrum auctions, and these companies would like to require satellite companies to come to them to negotiate individual site licenses for the spectrum, presumably as a way to recover some of their investment in the spectrum.

To some degree, at least so far, the millimeter wave spectrum has been a bust for cell companies. Millimeter wave spectrum comprises very short waves that can deliver immense amounts of data. You might remember the ads on TV in 2019 when Verizon touted gigabit speeds on the new Samsung Galaxy S10 smartphone. It was impossible that year to watch a sporting event without seeing these ads multiple times.

At the time, Verizon had deployed millimeter wave spectrum in a few downtown areas of a handful of cities in the hope that faster speeds would lure customers to buy a more expensive subscription to get access to faster cellphone speeds. The trial was largely a bust since it turned out that millimeter wave spectrum is blocked by everything in the outdoor cellular environment, even the body of a cellular customer. But the real issue was that the public had little appetite to pay for faster cellular speeds and still doesn’t today. Cellphones don’t run apps that need superfast downloads, and existing 4G networks at the time easily delivered video streams.

Cellular carriers argue that letting satellite companies use the spectrum for more ground stations diminishes their licenses. Satellite companies counter by pointing out that there are almost no outdoor uses of the spectrum today, and most applications are being used to wirelessly move a lot of data indoors.

This will change to some extent now that Verizon has purchased Starry, which is the only carrier to make extensive use of millimeter wave spectrum outdoors. It seems likely that Verizon plans to expand the Starry model and technology, which is only used today in a few markets.

This fight is a good example of the spectrum battles we are going to see over the next few years that tend to pit large, important constituencies against each other. I expect a lot of loud spectrum battles as the FCC tries to find 800 MHz of mid-range cellular spectrum for auction – almost all spectrum in this range is already used by somebody other than the cell carriers.

The FCC is forced to choose, in this case, between cell carriers and satellite companies. My bet is that the satellite companies will be allowed to use more spectrum under the light licensing rules. This is partly because they currently seem to have a favored status at this FCC, but also because the cell carriers have never utilized the millimeter spectrum in any meaningful way. This also might be a precursor to a future where the FCC decides that multiple parties can more easily share spectrum.

Fixing Federal Permitting

The House of Representatives recently approved H.R. 5419, a bill introduced by Representative Tom Kean of New Jersey. Titled the Enhancing Administrative Reviews of Broadband Deployment Act, the bill would attempt to improve the process of getting permits for broadband on federal land. The bill has been referred to the Senate Committee on Energy and Natural Resources.

Specifically, the bill would direct the Department of the Interior and the Department of Agriculture to conduct a comprehensive review of administrative barriers involved in reviewing applications to use federal rights-of-way. The two agencies would have a year to report their findings of the review to Congress, along with a plan on how the agencies can provide the needed staffing to review rights-of-way applications in a timely manner. The bill correctly recognizes that applications to build broadband on federal lands can be incredibly slow, to the point of creating hurdles that can stop broadband projects.

Looking back through old blogs, I see at least two other attempts by past administrations to improve access to federal lands for broadband.

In 2012, President Obama signed Executive Order 13616 that tried to make it easier to build fiber on federal lands. In a process that sounds similar to the one contemplated by the new legislation, the EO established a Broadband Deployment on Federal Property Working Group that was given a year to create a report to describe how to streamline applications for broadband on federal lands. The GAO was tasked with creating a standard set of forms and fees to apply for building broadband on federal lands. Finally, the EO ordered that all federal agencies implement a Dig Once policy that required placing conduit as part of any road projects on federal land.

In 2018, President Trump signed Executive Order 13821 that directed all federal agencies to seek to reduce barriers to capital investment, remove obstacles to broadband services, and more efficiently employ government resources related to broadband. The EO directed the General Services Administration to evaluate the effectiveness of the GSA Common Form Application used to get permission to locate broadband facilities on Federal land. Finally, the EO specifically directed the General Services Administration to complete the creation of a common form and master contract for placing wireless facilities on federal lands.

This repeating cycle of relooking at the same issues makes me wonder if this will ever be solved. And even if this legislation finally solves the issue, it will be far too late for BEAD implementation. BEAD grants are finally starting to be awarded this year, and those grants will have a four-year shot clock. For now, NTIA says that it does not expect to extend any grants beyond the four-year window.

This is not to belittle the legislation. Perhaps Congress can make progress on the issue when executive orders have failed. It just seems like the timeline contemplated by the legislation won’t provide solutions in time to make meaningful changes for broadband projects that will likely be underway this year. I hate to be pessimistic, but I have to wonder if the legislation or any executive orders will filter down to the dozens of agencies in charge of rights-of-way in various parts of the federal government.

Where’s the Growth?

I work for a few ISPs that are lucky enough to be working in counties that are seeing explosive population and new housing growth. Growth has always been uneven across the country, and I decided to take a look at the parts of the country with the fastest growth.

One of the best ways to understand housing growth is to look at permits to build new homes. Historically, 99% of permits to build new single-family homes results in a new home. About 80% of permits for multifamily homes turn into a condominium or apartment.

According to the Census Bureau’s Building Permit Survey, the top five states for permits for single-family homes in 2025, as measured in permits per 100,000 residents, were South Carolina (56), North Carolina (48), Florida (45), Idaho (45), and Delaware (43). At the bottom of the list were New York (4), Connecticut (5), Illinois (6), Rhode Island (6), and Massachusetts (6).

The states with the most permits for multi-family homes, as measured by permits per 100,000 residents, were Washington DC (60), South Dakota (33), Colorado (28), Florida (27), Arizona (27), and Idaho (26). The states at the bottom of this list were Mississippi (2), Rhode Island (3), Michigan (4), Alaska (4), and New Mexico (5).

Since my state of North Carolina had the second-fastest growth in single family-home permits per 100,000 residents, I thought I’d look deeper at the statistics in the state by county. As you might imagine, in most states, there is a wide variance between the counties with the fastest housing growth and the slowest. The following statistics for North Carolina are for total housing permits per 100,000 residents for 2023. The counties with the fastest new housing permits per 100,000 were Brunswick (254), Iredell (115), Warren (110), Pender (100), and Franklin (97). There were 8 counties with fewer than 10 new permits per 100,000 population, all in the eastern part of the state.

North Carolina has less variance in counties from top to bottom than most states, and there is healthy growth happening across most of the state. Other states that have growth in most parts of the state include Washington, Arizona, Florida, Utah, Virginia, and New Hampshire.

None of these statistics says anything about 2026, because the housing industry is infamous for seeing cycles of big booms and busts. The biggest booms since 1970 have been in 1972, 1977, 1984, 2005, and 2021. Some of the worst years for new housing were 1975, 1981, 1990, and 2009.

The housing industry had high hopes that 2026 could be a high-growth year since mortgage interest rates at the end of 2025. Realtors report pent-up demand for people who would like to move or buy a new home. But other factors like the tariffs that affect lumber prices, stagnant new job growth, labor shortages in parts of the country, and overall inflation could also result in a bad year for new housing.

One of the biggest mysteries in the housing industry is understanding the relationship between new housing growth and population growth. The two issues clearly are related, but which comes first? Does building new homes attract new residents, or do developers react to people who were coming anyway?

ISPs that find themselves in the middle of housing booms should count themselves as lucky. It’s something that can disappear almost as quickly as it appeared, but taking advantage of growth is one of the easiest business plans.

The Human Touch

Fierce Network recently had an interesting article about Consumer Cellular. This is a cellular MVNO that you might not have heard of. For those not familiar with the term, an MVNO is a cellular company that buys wholesale minutes and data from one of the large cellular carriers and markets the cellular product under its own brand name. Consumer Cellular operates on the AT&T network.

The company was founded over thirty years ago, aligns with AARP, and has always marketed to older cellular customers. Consumer Cellular stresses affordable cell plans, and its average revenue per customer is around $30 per month. The most interesting thing about the company is that it is growing while many other MNVOs are shrinking. The company has grown to over 4.4 million customers and is still seeing continued customer growth.

Consumer Cellular credits its success to the fact that it doesn’t use AI chatbots to answer online customer inquiries, as is being done with most other cellular companies. Consumer Cellular answers all customer calls with a live person.

The company is also rapidly expanding retail stores at a time when some of the bigger companies have been closing retail stores. Consumer Cellular opened its first retail store in Florida in 2022 and now has over 70 locations. The company has plans to expand to between 200 and 300 stores. The company sees big benefits from the retail stores. Customers who enroll at the stores have lower churn. The company says that people who visit a store also tend to buy higher-end phones and subscribe to larger monthly plans after they can see the options in person compared to online.

Most ISPs can’t duplicate Consumer Cellular’s success. The company is unique because of its older demographic. Consumer Cellular has embraced something that many smaller ISPs have known for years. Small ISPs universally tell me that one of the key secrets to success is that they know their customers. Smaller ISPs have thrived on having customer service calls answered live and having knowledgeable technicians. I know ISPs competing against big cable companies who say their live and responsive customer service is one of their primary advantages when competing against a big cable company.

However, this is changing, even for many smaller ISPs. Not all customers want to talk to a person. Consumer Cellular is unique because of its older demographic, but ISPs serve customers of all ages. Many customers would rather communicate using chat, text, social media, or email. I see many of my clients expanding the ways that customers can communicate with them. The ultimate customer service platform would give each customer the choice of how to communicate. But that can be expensive for ISPs that don’t have a lot of customers.

At least one big ISP went to the opposite extreme. For a long time, Starlink had no live customer service reps, and customers were expected to interface through an AI chatbot. The web is full of reviews from Starlink customers who got frustrated because they couldn’t talk to a person and walked away from the company. However, the company now has some live customer service reps based in Latin America that can be reached after customers start with a chatbot. Starlink has an unusual customer service challenge since it operates worldwide across a huge range of languages.

Some of the biggest ISPs are already experimenting with AI bots on customer service websites, meaning customers are chatting with a computer instead of a person. If they are not already being approached, I expect smaller ISPs will be seeing salespeople offering to integrate AI into their processes. I would caution smaller companies about taking this plunge. There is a lot of rumbling from large companies that AI customer service isn’t acting the way they were hoping for. The AI industry is still young, and there is likely going to be a shakeout at some point where a lot of the companies offering services using somebody else’s AI platform could disappear overnight. Small ISPs should think hard before giving up their local advantage and letting computers talk to their customers.