There is no question that this has turned into one of the oddest years for broadband during my career. We’ve seen Digital Equity grants killed. We’re seeing the spending for BEAD being cut in half. And maybe oddest of all, we’re seeing States make sizeable BEAD grant awards to Kuiper, although the company isn’t close to having its first broadband customer.
You might think we should have learned a lesson from when Starlink was a big winner in the RDOF reverse auction. The FCC eventually killed those awards after it determined that Starlink was not ready to fulfill a major commitment to serve large numbers of locations in specific geographies.
As of the date of this blog, Kuiper has 153 working satellites in orbit. It has scheduled launches of an additional 72 satellites before the end of the year. It’s worth noting that previous planned launches have all been seriously delayed.
Kuiper was granted permission in July 2020 to deploy a constellation of 3,236 satellites. The satellites will be deployed at three altitudes of 370 miles, 380 miles, and 390 miles. The company says it will begin beta testing when it reaches 578 satellites deployed at 390 miles. To put this into perspective, Starlink launched commercial service when it reached 1,260 satellites. Even with that number, early Starlink customers complained about short service lapses between satellites. Kuiper is under pressure from the FCC to have 1,618 satellites in orbit by mid-2026 to maintain its spectrum licenses.
For the last month, I’ve been perplexed by the magnitude of the BEAD awards being made to Kuiper. As of the date of this blog, Kuiper has tentative BEAD awards for 324,000 locations, second only to Starlink at 427,000. The next biggest award winner is Comcast with 233,000 locations. These are tentative awards, and NTIA is still reviewing and may reject some of the tentative awards to fiber, which would likely increase the awards to satellite.
Kuiper is not nearly as ready as Starlink was with RDOF. The RDOF reverse auction closed at the end of 2020, and Starlink invited selected customers from its waiting list to try the service in January 2021. Starlink started taking pre-orders nationwide in April 2021. Starlink could finally reach every part of the lower 48 states in March 2022.
One of the oddest things about Kuiper is that nobody knows how fast the service will be until it is deployed. Early Starlink customers received speeds that were faster than advertised, but speeds went downhill quickly as customer additions outpaced satellite deployments.
Starlink is only now on the cusp of delivering consistent 100/20 Mbps broadband. According to a report from Ookla, Starlink speed tests in the second quarter of 2022 showed a median download speed of 53.95 Mbps, meaning half of customers had speeds faster than that speed, and half were slower. Median upload speeds in that quarter were 7.5 Mbps. In the first quarter of 2025, Ookla reported that Starlink had climbed to a median broadband speed of 104.71 Mbps download and 14.84 Mbps upload, nearly double the speeds in 2022. The Ookla report said that only 17.4 % of Starlink customers fully met the FCC definition of broadband of 100/20 Mbps per second, with the limiting factor for many customers being slow upload speeds.
I saw a recent quote from a State broadband manager, when asked why he made an award to Kuiper, said it was because they bid the lowest cost per passing. That seems like a cynical response, and it makes me wonder if State Broadband Managers have thrown up their hands and are just following NTIA’s rules without questions.
The chances are good that Kuiper will complete the constellation and will eventually deliver satellite broadband. But history has also shown that new technology companies are often late in meeting commitments and sometimes fail altogether. The BEAD grant process is taking a big chance that Kuiper will meet its obligations and that speeds will be reasonably fast – something that nobody can know until it happens.
The only difference to residents between a location that receives a BEAD subsidy for LEO service and one that does not is the possible “free” satellite terminal, i.e., roughly $300. For low-usage areas, Starlink seems to have periodic free terminal offers (including right now), so it’s even lower for those areas. Thus, the impact of BEAD grants on these locations is near-zero, regardless of the LEO provider chosen by the state broadband office. (Conceivably, BEAD users will get priority in areas with limited capacity, but there are currently only a few of those.)
look at how BEAD is set to be distributed. Any unserved BSL is getting funded, and starlink and kuiper have 0 claimed BSLs covered. That means they are going to get funding for BSLs they bid to cover that they *already cover* and *already serve* ie existing customers. So it doesn’t matter to them in the short term, they’re getting that BEAD money if a potential customer signs up or not.
Getting that paying customer is now an entirely different thing, the ‘build’ funding and the recurring revenues essentially unrelated.
And they still need to market to get those customers against the $40-50 cellular fixed because after they spend some billions of dollars on more satellites and capacity, they need income for the 1-2 per day failure rate once the BEAD money runs out. They really have to onboard huge numbers of customers for network maintenance, even at a ‘loss’ in a sense, because all those birds and all that capacity is sunk.
I suspect there’s a bubble forming that is at risk of popping in 5-7 years, ie EOL on the birds launched during bead. When either a bailover of some sort is required or a price hike or reduced service levels etc.
“Past Performance”, i.e. an evaluation of a contractor’s history on previous contracts to predict future success, used to be a standard prerequisite for most, if not all, federal contracts. I don’t know why a company with zero past performance, or proof that their product could meet the minimum requirements, would even be allowed to participate in BEAD. The whole way this program has been rolled out, across both administrations, has been disgraceful.
We went this route, not asking for money just to prove we had services. I submitted scheduled speed tests for customers as evidence and it was rejected.
The government just isn’t very good at these things, and as a result you should expect poor outcomes.