There was an article published last week by Dr. Christopher Ali in Tech Policy.Press that asks if we should be making widespread broadband grants to Starlink and other low-orbit satellite technologies. Dr. Ali is Professor of Telecommunications in the Bellisario College of Communications at Penn State.
I highly recommend reading his paper. I was particularly taken by his conclusion. He says, “There is an important difference between technological neutrality and technological equality. LEO and fiber are not equal, and any policy that treats them as such will widen the very divide we have spent decades trying to bridge.”
I have been making this same observation about almost every grant program in the last decade, but just not as eloquently as Ali. As Ali points out, his comments are not a criticism of Starlink. Like him, I’ve talked to dozens of rural folks who absolutely rave about Starlink. For rural households, finally getting access to working broadband twenty years after the rest of the country has been a revelation. They can finally work at home, join Teams calls, and take online classes – things that the vast majority of Americans take for granted. Starlink should absolutely be a part of BEAD to reach remote locations, but should it be deployed to other locations?
Ali’s real issue is with the folks who set grant rules. Consider BEAD. The rules were established in 2020, and grants were not expected then to be finally completed until the end of 2028. Consider how much broadband has changed in the country just between 2020 and 2025.
- In 2020, Ookla said the median broadband speed in the country was 86 Mbps download and 12 Mbps upload. In March of this year, Ookla says that median broadband speeds in the country has increased to 287 Mbps download and 53 Mbps upload.
- In the second quarter of 2020, OpenVault said the average U.S. household used 359 gigabytes per month of download data and 25 gigabytes of upload data. OpenVault says at the end of 2024 that consumption had grown to 652 gigabytes of download and 46 gigabytes of upload.
The policy folks who set the BEAD rules set the broadband target performance for BEAD just a hair above the national average broadband performance in 2020. We’re only half way to the completion of BEAD grant construction and the country has already more than doubled the 2020 national broadband speeds and consumption. It’s not a stretch to predict that by 2028 the average U.S. home will be consuming more than a terabyte of data each month.
If the authors of the BEAD grant rules had looked just a decade forward, they would have set the BEAD performance standard to something like 400/100 Mbps. It doesn’t seem like a big policy stretch to think that valuable grant money ought to build networks that match the average market performance when they are completed. As Ali mentions, the biggest issue with LEO satellites isn’t even speeds, but capacity. Will the LEO companies be able to provide broadband to the many millions of households who will have no other broadband options?
It’s obvious why the folks in Congress picked wimpy BEAD standards. They are politicians and were under tremendous pressure from ISPs to not be excluded from BEAD dollars – and even under more pressure to not declare cable company networks as underserved. I remember the furor from cable companies in 2020 that lobbied hard against the BEAD upload speed requirement of 20 Mbps. That was because, at that time, most of them had upload speeds closer to 10 Mbps. It’s amazing what only five years of market pressure has done, and cable companies are upgrading urban upload speeds to 100+ Mbps with quick mid-split upgrades and have plans to get to gigabit upload speeds with DOCSIS 4.0.
WISPs didn’t have the same market power as cable companies in 2020, but they fought hard to make sure that the requirement for BEAD didn’t climb above 100 Mbps. But after only five years, they got access to a lot of new spectrum and can buy gear that will deliver 500 Mbps or faster broadband.
As ALI points out, LEO technology barely meets the 2020 definition of broadband that was codified in BEAD, and it is not a forward-looking technology – it is not equal to fiber or even to fixed wireless. And yet, the NTIA is doing mental gymnastics using an argument about technology neutrality to give more money to satellite technology. Perhaps the critics of satellite technology will be proven wrong, and satellite providers will improve technology so that by 2028, they will be delivering forward-looking speeds and coverage. But if not, we’ll be making grant awards in 2025 to implement 2020 broadband.
As always, Doug Dawson, you are correct. We agree that Starlink is doing a great job. I’m happy to see long-time underserved, finally served. Capacity is the issue. Serving a few houses in the neighborhood is one thing. Serving the neighborhood is quite another. All slows to a crawl.
Deploying obsolescence is absurd. The current standard will guarantee the US stays way behind many other countries.
https://eldotelecom.blogspot.com/2025/03/origin-of-tech-neutral-shift-for-bead.html
None of this should be surprising. Dates back to the 1990s when federal policy was to promote always on broadband speed over agonizingly slow dialup.
That remains the objective today notwithsanding the Biden administration’s view of fiber as critical infrastructure and allocating subsidies in the 2021 infrastructure legislation. More “broadband” connections at faster speeds than in the previous few years the KPI. LEOs meet that objective.
I agree Starlink is a fantastic product, and I also know it has and will have capacity and sustainability issues. Recently a reddit post highlighted a $250 congestion fee in Sacramento CA, not far from us. And there is the whole subject of the 5 year life cycle of the birds in space.
Putting that all aside, the clamor for faster and faster speeds is misleading. The article mentions monthly GB used and median speedtest numbers. Neither represent the clients day to day experience with an ISP. Let me show an example
#1 We had a client on a legacy 10 Mbps service. (side note, they only had a 2.4 WiFi router and every client device was WiFi in the house) They consumed around 1TB a month in data
#2 We upgraded them from 10 Mbps to 100Mbps. While contributing greatly to the “median speedtest” numbers, their actual day to day usage didn’t change.
The reason for the upgrade was not due to customer complain. Their old service was delivered on 5 Ghz and was overgrown with trees, while still able to satisfy 10 Mbps. The new service was CBRS Tarana and we push a minimum plan of 100 Mbps on that tech. They didn’t really care, happy before, happy after. Now 10 Mbps is too low for most families, I do get that. But 100 Mbps is going to satisfy for a long time, many years. I do not believe that the people in charge were wrong for picking that number rather than something in the 250-500 range. Where they were completely wrong is spending even one dollar on last mile, instead of putting around 30-50% of that money all into better regional fiber connectivity so the many companies excelling at last mile could have access to cheaper fiber connectivity at more locations. We have wireless technology that easily, reliably delivers 1Gbps point to multi-point, all it needs is a solid backhual and fiber to the tower makes that really easy.
I agree with trendaltoews
I do think there’s a big change in how much average data consumed by residential services when they hit certain high water marks for sustained speeds. For instance, if you can regularly get 35Mbps then you may start getting Netflix 4K video, under 30Mbps and you’ll be at 1/4 the bitrate on 1080P. That jump in average use will not be repeated by Netflix or Amazon for a very long time even with 10Gbps service, as most consumers aren’t looking for 8-16K TVs or getting access to that content.
There was also a sea change on in video game distribution. Games went from being delivered via disk to being downloaded. Even if a disk is purchased, the games need 100GB+ downloads to play, and likely weekly updates that are 10-100GB. That doesn’t contribute to average use, but could easily add 1TB to monthly use.
Further, the low hum of the work-from-home folks is small but aggregates up. 1-5Mbps is all they typically use but if 20% of a customer base is doing that it really brings the numbers up.
I’m just saying this to balance the ‘Technology Equality’. That term only applies in context to whatever conversation you’re having. Mobile data on your phone is dramatically superior to 1Gbps home service… when you are in your car. And 150Mbps Starlink is dramatically superior to the 1G circuit that is 15 miles from your home when you’re at home.
The 100×25 service from a provider that you can get, that puts effort into actually delivering that to you reliably and at a fair price, that service checks all the boxes.
And just a comment about the ookla tests. That’s just what you can get, that’s the speedometer on your car. It’s not the amount of data you need to ‘feel’ like your service is fast. 100×25 done well feels fast. 500×35 done poorly feels slow. “Technology Equality” is such a narrow scope, ignoring that a service is the aggregate of it’s parts. local wifi, the last leg, the switch/router network, the head end, QoE/QoS management. Knowlagable, fairly paid, good moral staff being a non-technical but pivotal part of the system.