What changed
A blog post by 5Gstore, a US network-equipment retailer, dated May 5, 2026, summarised Ookla Speedtest data for US Starlink users: the median download speed rose from 100 Mbps in Q1 2025 to 132 Mbps in Q4 2025, then settled at 128 Mbps in Q1 2026. It reported that 22 states had median uploads of 20 Mbps or more, and that 44.7% of users met the FCC’s 100/20 Mbps benchmark in Q4 2025, up from 17.4% in Q1 2025.
Alaska was the one state below 100 Mbps, at a reported 87.6 Mbps median. Several Plains states, including North Dakota, South Dakota, Nebraska and Wyoming, plus New Jersey, saw download gains of 80–90% year on year.
When
Reported May 5, 2026, covering Q1 2026 data. We couldn’t open Ookla’s own article, so this is a press report from a single outlet; Ookla’s later half-year report (our September post) points the same way. This post is more than 90 days old and is kept as background.
Who is affected
Anyone weighing Starlink at home, and owners who get asked how fast it is. No plan, price or referral rule changes.
What it means for you
Median speeds are a fair way to describe a network, but they hide spread: half of tests are slower than the median. A referral works the same whatever speed the friend gets, but a friend who isn’t happy and cancels in the first two billing cycles ends the referral for both of you.
If you share your experience, share your own numbers and say where you live; that is more useful to a friend than a national median.
What to do now
Point friends to independent speed data for their state, and to the 30-day return window, rather than promising a speed.
What stays the same
The referral program, plans and prices are unchanged by a speed report.
What we don’t know yet
Ookla’s exact methodology for this release, and the figures for later quarters, which we’ll only cite when we can read a source.
How to check it yourself
Ookla publishes Starlink analyses on its own website; look for its Q1 2026 Starlink release and the H1 2026 US report. 5Gstore’s summary is the source we could read. For a single home, run Speedtest at several times of day over a week rather than relying on one result.