Starlink pricing, Telesat delay and Amazon Leo slip favour SpaceX
- Good news: Starlink's dynamic pricing is showing on bills, with rises capped at about 20%, a new way to lift ARPU; the evidence is user-reported.
- Good news: Telesat Lightspeed slipped by three months and its demonstration satellites will ride a Falcon 9, which is launch revenue for SpaceX.
- Mixed: Amazon Leo dropped its mid-2026 target and has only 3.6% of Starlink's in-orbit fleet, but has the capital for a long war of attrition.
What we watchSegment data in the Q4 10-K, to show dynamic pricing's real effect on ARPU.
Read the full notes
Three independent signals strengthen SpaceX's position at once. (1) Starlink's dynamic pricing by regional supply and demand is starting to appear on bills. Rises reach a ceiling of about 20%, with prices up where capacity is tight and down where it is idle, a new way to lift ARPU.
Read more
(2) Its main rival, Telesat Lightspeed, has slipped by three months, and its demonstration satellites will now ride a Falcon 9 rideshare. A rival's constellation build-out is itself launch revenue for SpaceX. (3) Amazon Leo has quietly dropped its mid-2026 target and has only 3.6% of Starlink's in-orbit fleet. But with its AWS channel and Globalstar's direct-to-phone spectrum, it is the only rival with the capital to fight a war of attrition over the medium to long term.
Key evidence
- In the US Midwest, Starlink's 100Mbps residential plan rose from $50 to $60.50 a month (+23%) and Residential Max from $120 to $143 a month (+19%). Some users in other areas received price-cut notices. Source: How-To Geek, 30 Oct, compiling self-reported Reddit and Facebook posts
- The rises mostly hit the ceiling of about 20% that PCMag mentioned. The sample is user-reported rather than company data, and the real ARPU effect will have to wait for segment data in the Q4 10-K
- Telesat Q3 results, 29 Oct: global Lightspeed service is delayed by three months to the end of Q2 2028 (the Canadian Arctic service target is unchanged). Two demonstration satellites missed their launch window and will move to a Falcon 9 rideshare, launching in about June 2027
- Telesat has signed for 14 dedicated Falcon 9 launches and has not yet booked a 15th. Its Q3 loss was C$157 million, and about $2.4 billion of debt matures in December. Source: Via Satellite, SpaceNews and SpaceQ
- Falcon 9 production is being cut to make way for Starship. Ars Technica (30 Oct) reported that industry analysts think the launch capacity squeeze is worse than expected
- Amazon Q3 call, 29 Oct: Jassy said Leo is on track to begin initial service this year (The Stack noted that the earlier mid-2026 target had been quietly dropped)
Show 4 more
- As of 28 Oct, Amazon Leo had about 400 satellites and 14 launches. Leo's production and launch costs will only start to be capitalised from Q1 2027, and most are still expensed in Q4
- Business service in the Baltic states will open in 2027
- KeepTrack, 30 Oct: Starlink has launched 13,069 satellites in total, with 11,190 in orbit, of which 11,176 are working (11,144 on 25 Oct)
- Leo has only about 3.6% of Starlink's in-orbit fleet
Catalysts to watch
- Segment data in the Q4 10-K will show the real effect of dynamic pricing on ARPU
- How Telesat handles its $2.4 billion of debt due in December
- Launch of Telesat Lightspeed's demonstration satellites on a Falcon 9 rideshare in June 2027
- Amazon Leo's deployment pace in 2027
- The point in Q1 2027 when Amazon Leo starts capitalising costs
- How long the Falcon 9 capacity squeeze lasts, and how fast Starship takes over
Risks
- Dynamic pricing evidence is user-reported only, with no company data to back it
- Rises have hit the ceiling of about 20%, so there is limited room for further increases
- Amazon Leo has an AWS channel and Globalstar's direct-to-phone spectrum, and over the medium to long term has the capital to fight a war of attrition
- If the Falcon 9 capacity squeeze persists, SpaceX's own Starlink launch cadence will also suffer
Sources 5
- How-To Geek, 30 Oct
- Via Satellite, SpaceNews and SpaceQ: Telesat Q3, 29 Oct
- Ars Technica, 30 Oct: Falcon 9 production cut
- The Stack: Amazon Q3 call, 29 Oct
- KeepTrack, 30 Oct
A brief is a lead for our research, not a verified conclusion or advice.