← Back to Fleet

OneWeb (Eutelsat LEO) — Scope & Funding Ledger

The evidence base behind oneweb.json. Reconstructed by reading the Eutelsat earnings calls, trading-statement calls, shareholder calls and the two 2026 investor decks chronologically (2024-08 → 2026-08) — not a figure list — with [MGMT] company-stated vs [ANALYST] flagged and a funded-vs-notional line made explicit. Cross-referenced against dossiers/industry-consolidation.md. Where a figure cannot be sourced to a primary Eutelsat document it is flagged, not adopted.

Last reconstructed: 2026-08-14 · Sources: 13 Eutelsat primary documents in inbox/ (five earnings calls, four trading-statement calls, two shareholder/AGM calls, two investor decks).

Framing caution that governs the whole ledger: Eutelsat reports "LEO revenue," not "OneWeb revenue" (CFO Caudrelier, 2026-08-05: "what we report is not OneWeb revenue. We report LEO revenues … the combination of all LEO revenues"). There is no standalone OneWeb/LEO EBITDA disclosed anywhere.


1. Satellite & generation evolution — reconciled

The credibility-critical distinction: **the 440 is a Gen-1 continuation (follow-on), NOT Gen-2.** IRIS²/Gen-2 is the separate, later system.

Fleet & generation, post-IRIS²-signingsats650+ Gen-1 flying today; 440 continuation committed; IRIS²/Gen-2 LEO revised 264→330 (348 total incl. 18 MEO) on the Aug 7 2026 signing
Gen-1 in orbit
650 sats
Gen-1 continuation (440)
440 sats
IRIS²/Gen-2 LEO (330 of 348 total)
330 sats
replaces the pre-signing 264-sat chart; Hispasat separately reports 342 total (6 fewer than the EC's 348), unreconciled — see dossiers/iris2-spacerise.md OPEN QUESTIONS #7as of 2026-08-07Eutelsat IP deck/earnings calls (Gen-1 figures, unchanged); EC press corner IP/26/1726 (IRIS²/Gen-2 revision, dossiers/iris2-spacerise.md TRACKED SERIES)
CountMeaningPrimary source
>400LEO sats launched at merger (Sept 2023)IP deck 2026-03-24 [MGMT]
+20Spare batch, Oct 2024 (first sats reaching end-of-life)2024-10-29 CFO [MGMT]
650+Full in-orbit Gen-1 fleet now operating2025-08-05 CEO [MGMT]: "a fleet of over 650 satellites … 99%+ availability"
~450–500Minimum for global coverage (1,200 km)2025-02-14 CEO [MGMT]
440Gen-1 CONTINUATION = 100 (by 2027) + 340 (by 2029), AirbusIP deck 2026-03-24 [MGMT]; 100 ordered post-IRIS² signature, 340 ordered Jan 2026
264IRIS²/Gen-2 LEO sats, entry 2030+IP deck 2026-03-24 [MGMT]

"Continuation, not Gen-2" — management's words: 2025-08-05 CEO: "ensuring continuity beyond Gen-1 with the upcoming deployment of 440 follow-on satellites." ⚠️ Late re-labeling: by 2026-08-07 the CEO recast IRIS² itself as "the generation 2 of our constellation." Settled taxonomy: **440 = Gen-1 continuation/bridge; the 528-sat "Next" FCC filing = OneWeb's own filed Gen-2 (WorldVu), now deprioritized; IRIS² (264 LEO sats, LEO/MEO, 5G-NTN) = the effective forward Gen-2** the company is actually building toward, service 2030+. See §7.

⚠️ STALENESS FLAG (2026-08-15, dossier deep-dive routine) — the 264-sat / €10.6bn IRIS² figures above and in §3(b) predate the Aug 7 2026 Rendez-vous 1 signing and are no longer the current disclosed scope.

signingPRIMARY
EC press corner IP/26/1726 (2026-08-07): IRIS² is now 348 satellites total (330 LEO, incl. a new dedicated 66-sat High-LEO defence/security/emergency layer, + 18 MEO) — up from the ~264 LEO baseline this ledger still carries; total programme value now put at €15.6bn (vs. the €10.6bn this ledger's §3(b) carries), per dossiers/iris2-spacerise.md TRACKED SERIES (which also tracks an unreconciled Hispasat-reported 342-vs-348 count).
scopeROUTINE
this is a surfaced divergence, not a rewrite — per docs/specs/LEO-VIABILITY-SPEC.md §10, the evidence narrative above stays operator-owned; oneweb.json's routine-refreshable oneweb:gen2:leoSats claim has been updated to 330 (asOf 2026-08-07) as the mechanically safe half of this refresh. UPDATE 2026-08-28: the "Fleet & generation" chart below has now been refreshed to 330/348 (CHART-oneweb-fleet-generation-refresh, Up-voted 2026-08-22, materialized this run) — the chart is current. The prose/table in §1 above and §3(b)'s €10.6bn capex claim remain left for operator confirmation before rewrite.

2. Capacity

FigureMeaningSource
1.1 TbpsGen-1 sellable capacityIP deck 2026-03-24 [MGMT]
~2 Tbps total / ≥1.5 Tbps sellable to EutelsatIRIS² LEO capacity (incremental)2025-02-14 CEO [MGMT]

Gen-1 envelope: <70 ms latency, 99%+ availability, ~1,200 km.


3. Capex evolution

Gen-1 continuation capex — the figure walked down€BManagement guided the ~440-sat continuation cost down over four disclosures
€2B€2.25B€2.5B2024-102025-022026-03
Funded by €1.5bn equity + ~€1.0bn ECA inside a €5bn refinancing; cost/sat ≈ €2bn ÷ 440as of 2026-08-07Eutelsat earnings calls (MGMT)

(a) OneWeb Gen-1 continuation capex:

OneWeb Gen-1 continuation capex — the figure walked down across four disclosures, funded via €1.5bn equity + ~€1.0bn BPI-backed ECA financing.

2024-10-29MGMT
~€2.5bn
2025-02-14MGMT
€2.0–2.2bn, over FY24-25→FY28-29
2026-03-24 / 2026-08-07MGMT
~€2bn
fundingMGMT
€1.5bn equity + ~€1.0bn BPI-backed ECA financing

(b) IRIS²/Gen-2 cost:

IRIS²/Gen-2: €10.6bn total, ~60% public, Eutelsat share ~€2bn back-end loaded (mostly from FY27-28).

2025-02-14MGMT · CEO
€10.6bn total, ~60% public, Eutelsat share ~€2bn back-end loaded (mostly from FY27-28)

(c) Group gross capex path:

Group gross capex path — guided down, then a ramp — funded by a completed €5bn refinancing.

FY24-25MGMT
€449.8m actual
FY25-26MGMT
€594m actual (below the ~€900m cut guide)
FY26-27MGMT
guide ~€1.2bn
FY26–29MGMT
cumulative ~€4bn (incl. ~€2bn OneWeb replenishment + early IRIS²)
2026-08-07MGMT · CFO
funded by a completed €5bn refinancing

⚠️ The "€5.07B-to-2034" figure does NOT appear in any transcript or the inbox — flagged, not adopted.

gapFLAG
"€5.07B-to-2034" is unsourced, likely a conflation of the €5bn refinancing with a C-band incentive
sourceableMGMT
the ~€4bn FY26–29 cumulative capex and the €5bn refinancing

4. Launch — the market-dependency cautionary tale

OneWeb owns no rockets. It bought launch on the market — and when its provider became geopolitically unavailable, the constellation paid for it in cash and schedule. This is the counterpoint to Starlink's vertical integration.

OneWeb Gen1 — 18 launches to 618 satelliteslaunches13 Soyuz + 3 Falcon 9 + 2 LVM3, completed Mar 2023
Soyuz · Arianespace
13 launches
Falcon 9 · SpaceX
3 launches
LVM3 · ISRO/NSIL
2 launches
Soyuz suspended after Russia's Feb-2022 invasion; the last 5 launches switched to SpaceX + ISROas of 2023-03Press (NASASpaceflight, SpaceNews, CBS); ISRO PRIMARY for LVM3

Owning no launch made OneWeb hostage to geopolitics — Russia's invasion stranded 36 satellites and forced a $229M write-off.

soyuzSECONDARY
Original Arianespace deal: 21 Soyuz launches for $1–2B (2015), "largest commercial rocket buy in history." 13 flew before Russia suspended cooperation in Feb 2022; Arianespace was owed $238M at the 2020 bankruptcy.
write-offPRIMARY
FY2022 statutory accounts: a $229.2M impairment — 36 satellites stranded at Baikonur plus lost launch prepayments. Net loss $389.8M on $9.6M revenue.
scrambleSECONDARY
Re-bought launch at market to finish the fleet: ISRO 2× LVM3 (~$120–130M) + SpaceX 3× Falcon 9 (value undisclosed). 618 satellites complete Mar 2023.
contrastDERIVED
The lesson for the fleet: paying the market for launch is not only costlier ($70–150M/rocket vs Starlink's ~$15M internal) — it is a single point of geopolitical failure a vertically-integrated launcher never carries.

Ground is a real, separately-contracted cost — not a hole to hand-wave.

hughesPRIMARY
~$250M Hughes ground-system contract (2020): gateway electronics + the core module in every user terminal. ~44 gateways planned for the 648-sat network (~22 in development by 2020).

5. LEO revenue trajectory ("LEO," not OneWeb-specific)

LEO revenue trajectory€mReported as blended 'LEO revenue', not standalone OneWeb — no standalone LEO EBITDA disclosed
€187m€242m€297mFY24-25FY25-26
+84% then ~+70% l-f-l; FY26-27 guide >+30%; now 25% of group revenueas of 2026-08-07Eutelsat FY results (MGMT)
PeriodLEO revenueGrowthSource
FY24-25€187m+84.1% l-f-l (15% of group)2025-08-05 [MGMT]
Q1 FY25-26€54m+71%2025-10-21
H1 FY25-26€110.5m+59.7%2026-02-13
Q3 FY25-26€62.2m+65%Q3 deck
FY25-26€297m~+70% (25% of group)2026-08-07 [MGMT]
FY26-27 (guide)>+30%2026-08-07 [MGMT]

Group target: FY28-29 revenue €1.5–1.7bn. Gov anchor: NEXUS (French MOD, €1bn / 10-yr framework, ~€100m/yr); first call-off CENTAURE €350m/8yr (2026-08-07).


6. Design life → 2034 horizon → mid-2032 migration

Design life 5–7.5 yr is the replenishment rationale for the 440.

2024-11-21MGMT · CEO
"a life span of at least 5, but … rather 6 to 7.5 years … The first 2 batches … were launched in 2019 … we'll have to start thinking … about replacement satellites."

Gen-1 is operational to 2034, with migration to IRIS² from mid-2032; IRIS² "secures our future until 2040."

2026-08-07MGMT · CEO
operational to 2034; migration to IRIS² from mid-2032
horizonMGMT
IRIS² "secures our future until 2040"

7. Profitability signals

IRIS² target IRR 10–12% (minimum).

2025-02-14MGMT · CEO
"in line with our typical IRRs of 10% to 12% … That's a minimum one."

LEO margin drag during ramp — LEO is less profitable than the established GEO business.

2026-08-07MGMT · CFO
"the LEO business is … not as profitable as the established GEO one."
marginMGMT
Group adj. EBITDA margin 51.2% FY25-26

No standalone LEO/OneWeb EBITDA is disclosed anywhere.

disclosureGAP
no standalone LEO/OneWeb EBITDA disclosed anywhere

8. FUNDED vs NOTIONAL

FUNDED / committed: the 440 Gen-1 continuation and IRIS²/Gen-2 are both financed.

440MGMT
both batches ordered (Airbus), financed by €1.5bn equity + ~€1.0bn ECA inside a €5bn refinancing ("Gen-1 FY'2026-29 fully funded")
IRIS²MGMT
framework-funded: 12-yr concession signed Dec 2024; €10.6bn / 60% public; Eutelsat ~€2bn from FY27-28; build phase from Aug 2026

NOTIONAL — filed/optional, no committed build:

OneWeb's own Gen-2 — the "Next" 528-satellite FCC filing — is on file but management has deprioritized it in favour of IRIS²; notional, no committed build.

FCCFCC
WorldVu Satellites; 528 sats, 22×24 planes, 1,220 km — OneWeb's own next-gen constellation on file, not a random external placeholder
deprioritizedMGMT
IRIS² positioned to "replace the need for Next-Gen"; "Next" appears in none of the 13 Eutelsat calls, with no committed capex or funding
statusNOTIONAL
OneWeb's filed Gen-2, effectively shelved in favour of IRIS² as the forward capacity — no committed build

Hosted payloads on the 340 continuation satellites — Eutelsat signed its first-ever hosted-payload contract, value confidential; a forward revenue line that may not yet be inside reported LEO revenue.

2026-08-07MGMT
"we have signed the first contract for hosted payload … the pipe is extremely nourished" (value confidential)
watchFLAG
as the first, a forward revenue line that may not yet be inside the reported €297m FY25-26 LEO revenue — one to watch in future earnings; the desk's dossier and briefs have not yet picked this up

Future new GEO video sats — flagged as post-2035, no committed project.

gapNOTIONAL
post-2035, no committed project

9. Figures that cannot be sourced → flagged, not adopted

"€5.07B-to-2034" capex — not in the call or inbox.

gapFLAG
not in the call or inbox
sourceableMGMT
sourceable figures are ~€4bn FY26–29 and the €5bn refinancing

The 528-sat "Next" Gen-2 filing is FCC-sourced (WorldVu) and real, but its scope/timeline/capex are not company-guided — no forward figures adopted.

FCCFCC
FCC-sourced (WorldVu) and real
scopeFLAG
scope / timeline / capex not company-guided; management is leaning to IRIS² instead — so no forward figures are adopted from it

No system-wide IRIS² Tbps — the ~2 Tbps is the LEO layer only; don't extrapolate a system-wide figure.

scaleFLAG
~2 Tbps is the LEO layer only (IRIS² is multi-orbit); don't extrapolate a system-wide figure

Operator read (2026-08-14, pending corroboration) — the ~66-sat OneWeb contribution to IRIS² is likely an accounting construct, not a network.

mechanismFLAG
~66 satellites cannot form a standalone global constellation, so Eutelsat's cited contribution of that order into IRIS² most plausibly functions as a financing/revenue-transfer mechanism — a way to channel IRIS² programme money to Eutelsat — rather than an independent OneWeb sub-constellation
caveatFLAG
operator interpretation, not a sourced figure (the "66" is not in the transcripts); flagged to test against the SpaceRISE cost- and revenue-share disclosures as they emerge

What this maps to in oneweb.json

The fixture models the go-forward continuation IRR: the disclosed ~€2bn continuation capex against the disclosed LEO revenue (€187m → €297m) ramped by a growth multiple × EBITDA-margin slider, benchmarked to Eutelsat's own disclosed 10–12% hurdle. cost/sat derives from €2bn ÷ 440 (both disclosed), so the chain is anchored and the verdict is publishable (base case ~1.5% vs the 10% bar; sliders can clear it). EUR money is FX-normalized to USD via a movable rate.