As some of you know, we are working on the City’s most powerful proprietary AI system to extend our work and make our own content more readily accessible. Importantly, “Rafael" learns and helps us integrate data, colour and news into our house view, allowing us to interpret events consistently and maintain a coherent investment approach. Rafael runs entirely inside our own environment on Sarria’s own research, and nothing leaves the house. Let us know your thoughts on the unedited example below:

Wolfgang’s Question: Good Morning. I am reading that the first rendezvous has taken place and Eutelsat have put the below press release out. I am now struggling a little with the stacking of satellite programs. How many LEOs are there now, how many more have we been anticipating, and how many more are they planning on shooting up there? What will that cost and when?

Answer: · pulling our colour Eutelsat
· reading the write-up Eutelsat
· reading the write-up Eutelsat
· reading the write-up Eutelsat
· table Eutelsat

Rafael: Good morning. Let me pull the current colour and model before reading this through our lens.Good morning. Let me also pull the OneWeb constellation write-up and the IRIS2 table before answering.Good. I have what I need. Here is the constellation stack, and then the read on the press release.

The LEO stack, where we stand, and what this announcement changes

Current constellation — what is in the air now

OneWeb Gen-1 comprises 648 satellites in low Earth orbit at approximately 1,200km altitude. That is the operational fleet today, giving roughly 90% global coverage. These are the satellites generating the LEO revenues we track in the model.

Already committed before today

Two tranches were already contracted and in our model before this press release:

Gen-1 replenishment (440 satellites, "new GEN1"): already procured, being delivered through approximately 2026-27, to refresh and extend the existing constellation. Technology-compatible with IRIS2. We had modelled this as the bulk of the OneWeb capex through 2029 — approximately EUR 2bn in total between 2024 and 2029, cross-subsidised from ESA.
IRIS2 shared infrastructure (EUR 2bn commitment): agreed in December 2024 when SpaceRISE signed with the EU. 264 LEO satellites plus 18 MEO, with Eutelsat leading the LEO segment. Completion originally guided 2030; capacity revenues not before then. Our model treated this as an equity option — not in the numbers, flagged as very long-dated.

What this announcement adds

Three things change versus what we had:

1. Scope increase on IRIS2: The constellation expands. The original IRIS2 had 264 LEO plus 18 MEO. This announcement confirms 264 dual Mil-Ka/Ku band LEO satellites plus an additional 66 Mil-Ka band satellites — so 330 LEO in total versus the 264 we had modelled. Our IRIS2 table carried 246 LEOs; the press release says 264 dual-band plus 66 Mil-Ka = 330. Throughput rises to over twice the current OneWeb capacity, which the announcement characterises as a commercial Ku LEO fully operational by mid-2032 (two years later than the 2030 figure we were working from).

2. Revised investment quantum: Eutelsat's total IRIS2 outlay is now EUR 2.23bn in shared infrastructure plus EUR 1.16bn in commercial infrastructure (Ku payload and commercial ground) = EUR 3.39bn total, phased 2027-2034. Our model had EUR 2bn. The incremental call versus our assumption is therefore approximately EUR 1.4bn, all back-end loaded into 2027-2034.

3. Gen-1 extension beyond the 440 already contracted: An additional 229 satellites on top of the 440 GEN1 already procured, at a cost of approximately EUR 1bn, to be deployed through 2034 to maintain service continuity into the IRIS2 era. We had modelled a Gen-1 follow-on of approximately 340 satellites at an estimated EUR 1bn; the announced figure of 229 satellites is smaller in count but the EUR 1bn cost is consistent with our estimate.

Total satellite count

So the Eutelsat-operated LEO fleet at full build-out (mid-2030s) is 648 + 440 + 229 + 330 = 1,647 satellites, plus a 330-satellite IRIS2 constellation taking over from the Gen-1 core. In practice, some of the 648 existing Gen-1 will have been decommissioned by then, so the operational fleet at IRIS2 maturity is approximately 330 (IRIS2) plus whatever Gen-1 remains operational through 2034.

What it costs and when

The company says capex through FY2029 is unchanged — consistent with the IRIS2 step-up being entirely post-2029.

The credit read

The announcement is positive for the equity and long-dated, and essentially irrelevant to our near-term credit positioning — which is where we are (sidelined, watching the rights issue). The revenue promise of over EUR 10bn between 2032 and 2040 is the same figure as before; what has changed is the scope is larger, the capacity is greater, and the commercial Ku constellation is not fully operational until mid-2032 versus the 2030 we had. So the wait is longer.

For the bonds, none of this changes the 2027 wall or the liquidity picture. The EUR 3.39bn IRIS2 commitment is phased over 2027-2034 with the bulk post-2029, and the company explicitly says capex to FY2029 is unchanged. The financing of the 2027-2034 tranche is undefined — the EU funds 60% of the total EUR 10.6bn programme, mostly upfront, which means the public funding is substantially front-loaded into R&D and early construction; the Eutelsat private contribution is back-loaded and will need a funding solution that does not yet exist. That is the question the 2028 and 2029 bondholders should be asking: what is the funding mechanism for EUR 3.4bn of back-end capex, when the GEO book is in secular decline and the Gen-1 revenues have consistently missed?

The "First Rendez-Vous" milestone marks transition to industrial execution — manufacturing and launches will now begin. This is operationally meaningful, but from a credit perspective it simply confirms the capex commitment rather than resolving the funding question.

answered in 53s

Eutelsat - comment