SpaceX is urging federal regulators to refine how they treat its satellite internet constellation in the Federal Communications Commission’s annual broadband performance assessments given the overall quality of service possible, especially when looking at latency and availability.
7379
In a Sep. 1 ex parte filing with the commission (PDF), SpaceX argues that removing “actual … LEO service” skews this nation’s progress report and sends investment elsewhere.
In a lengthy filing, the company tells the commission that its two‑year-old Starlink service has more than 2 million active U.S. subscribers and offers median peak‑hour download speeds of nearly 200 megabits per second — proof, according to SpaceX, that modern satellite broadband qualifies as meeting the FCC’s “advanced telecommunications capability” standard.
The push comes as the FCC again ponders how it counts satellite systems in its Section 706 reports, which historically targetted fixed terrestrial technologies like fiber and cable, based on satellite’s limited capacity and lower take rate. But SpaceX argues that such assumptions are outdated, and it alludes to an upgrade planned for V3 satellites that will expand capacity and provide gigabit‑class service.
Why the Section 706 Report Is Significant
Section 706 directs the FCC to assess whether broadband is being deployed to all Americans in a reasonable and timely fashion. The findings of the report affect benchmarks, rules to determine who is eligible and how policy makers direct their scarce dollars. If satellites counts aren’t accounted for, SpaceX contends, many entire areas could be erroneously labeled “unserved” when a viable satellite option exists — influencing the roster of grant-eligible neighborhoods in everything from state build plans to federal competitive-scoring.
The proceeding and the FCC’s moving speed target baselines also intersect. The agency’s current minimum for “advanced” service is generally considered to be 100/20 Mbps. Pulling satellite all the way into that framework would require more of a shift toward nuanced accounting of peak‑hour performance, capacity by service cell, latency and reliability: metrics where historically fiber has shone due to its relative consistency but for which the LEO constellations are getting much stronger.
SpaceX’s Argument: Scale Availability
SpaceX is framing the argument in terms of scalability. There is no matching fiber in terms of performance, but deployment to sparsely populated last‑mile areas is slow and costly. Industry estimates typically peg rural fiber builds at $5,000 to $10,000 per passing, with higher outliers in difficult terrain. A satellite constellation works the other way round – capacity is added on orbit, and through ground gateways, we have a country coverage nationwide in an instant; that’s connectivity for households in rural areas everywhere.
The firm highlights a steady increase in median Starlink speeds as the subscriber tally has climbed, crediting gains to changing satellite generations, laser inter‑satellite links and additional spectrum utilisation. It argues that counting satellite in federal reporting will be more realistic in terms of true availability for telemedicine, education and small business in areas where fiber builds are years away.
Critics Challenge “Fiber-Equivalent” Language
Rural constituencies and advocates for fiber remain unconvinced. The National Rural Electric Cooperative Association has asked the FCC to take into account LEO capacity limitations, cautioning that service in a given area may not consistently meet the 100/20 Mbps standard at all times if many of them signed up simultaneously. The WTA – Advocates for Rural Broadband have also pointed to times when Starlink paused adding new sign‑ups in certain geographies as evidence of capping cell‑level congestion.
NTCA – the Rural Broadband Association makes the case that “substitutability” should rest on a greater basis than theoretical availability and should focus on consumer‑perceived parity of speed, latency, reliability, and worthiness. Unaffiliated network testers in the United States have largely recorded mid- to high‑megabit-per-second U.S. median downloads for Starlink at 20th to 40th percentiles, with regular latency in the range of between 40 and 60 ms — much better than geostationary satellite, but below most single‑digit-millisecond experience on fiber. Peak‑hour performance can also be influenced by weather sensitivity and traffic management strategies.
Here’s What’s at Stake in the Battle Over Technology Neutrality
How the FCC categorizes satellite might have ripple effects across funding programs. The agency’s Broadband Equity, Access, and Deployment program for $42.45 billion in the Commerce Department prioritizes future‑proof buildout but gives states flexibility for hard‑to‑reach areas. If LEO service were to be accepted as being the same as fiber for availability measurement purposes, more places may potentially be labeled “served,” resulting in less of a need for costly fiber construction and altering strategies from state to state with respect to where capitals are deployed. Some public‑interest advocates also fear that redefining availability could water down the argument for retaining Universal Service Fund high‑cost support in rural markets.
For its part, SpaceX argues that a technology‑neutral approach should involve judging outcomes for consumers — speed, latency, reliability — not pre‑choosing winners.
Fiber advocates argue that multi‑gig service – now and well into the future -, symmetrical capacity, long service life and low ongoing costs make fiber the most cost‑effective long-term use of public funds, even where time-to-service is not as fast.
What to read and watch
The commission might be headed in a number of directions. It could opt to treat LEO service fully in its availability counts; go for a middle ground where satellite is given credit with weightings applied for capacity and latency; or establish an upper tier of service that doesn’t get to count places as “served” unless the performance is demonstrably still hitting predicted peak capacity levels during peak hours. Look for a renewed emphasis on empirical evidence — state mapping data, third‑party speed tests, subscription caps and real-world congestion patterns — rather than theoretical maximums.
(Wherever the FCC lands, its central question — should the nation’s broadband scorecard reward immediate, good‑enough connectivity in hard places, or be focusing first and foremost on infrastructure that scales to multi‑gigabit, low‑latency service over decades? SpaceX hopes that the answer to this question is “both”— and for Starlink to count accordingly.