US C-band spectrum & clearing economics

THESIS

The upper C-band proceeding is primarily a balance-sheet event for SES (clearing costs vs incentive-payment proceeds vs transponder capacity loss), and only secondarily a D2D fight. The CVR mechanic from the Intelsat acquisition sharpens this: SES splits net proceeds from the first 100 MHz 57.5/42.5 with legacy Intelsat shareholders but keeps 100% of anything cleared above 100 MHz — a pure incentive-capture logic would push SES to maximize clearing (every extra MHz above 100 is fully SES's), yet SES capped its own ask at 160 MHz, not 180. That gap is the real tell: capacity preservation outweighs even 100%-owned incremental incentive revenue past some threshold. Falsifier: SES publicly advocating a larger clearing block, or accepting one without escalating compensation demands.

STATE OF PLAY

FCC's July 1 draft — a 143-page combined Report and Order, Order of Proposed Modification, and Order on Reconsideration — proposes clearing 160 MHz (3.98–4.14 GHz), matching SES's June 18 ex parte ask exactly and undercutting the earlier-floated 180 MHz ceiling. Combined with the already-cleared lower C-band (3.7–3.98 GHz), this assembles a contiguous 440 MHz "super-band" for terrestrial 5G/6G use. Congress (One Big Beautiful Bill Act) mandates at least 100 MHz auctioned by July 2027 regardless of the clearing-size fight — the 160-vs-180 argument is about the margin above a floor that auctions regardless. Draft service-start dates: top-75 US markets by December 2030, remaining markets by July 2031 (reported, draft-stage). The draft explicitly defers all satellite-ops carve-out requests: "Due to the complexity of the Upper C-band transition, at this time we defer consideration of proposals to add further advanced satellite operations to the C-band" — confirmed as covering SpaceX's D2D ask (tied to spectrum SpaceX is separately acquiring from EchoStar: $17B for AWS-4 + $2.6B for AWS-3, both 2 GHz-band deals, not C-band itself — SpaceX wants C-band as an additional stacked asset). A second, independently-bylined outlet (Light Reading, Jeff Baumgartner) now also names SES alongside SpaceX and OQ Technology as having sought "advanced satellite operations" consideration — a materially stronger claim than the single uncorroborated Techtimes piece flagged 2026-07-04. Direct ECFS pull of all 95 SES filings in GN 25-59 by title still shows nothing D2D-labeled (cost/reimbursement/allocation-methodology/service-quality ex partes only) — the tension is unresolved, not disproven; see OPEN Q3. Compensation structure (secondary-sourced, ses.com IR FAQ + advanced- television + satnews): SES issued Intelsat's former shareholders CVRs for 42.5% of net proceeds from the first 100 MHz monetized; SES keeps 57.5% of that tranche and 100% of anything above 100 MHz. First-100-MHz incentive pool implied by the reported "$1.3B floor = 42.5%" figure is ≈$3.06B — a per-MHz rate (~$30.6M/MHz) startlingly close to round 1's realized rate (~$32.3M/MHz; see TRACKED SERIES arithmetic). SES+Intelsat's own disclosed clearing/transition cost is ~$3.6B + $150M contingency for 160 MHz (~$22.5M/MHz cost) — 4.2x round 1's realized per-MHz clearing cost (~$5.33M/MHz) even though round 1's incentive-payment rate and round 2's implied rate track closely. That cost/incentive divergence is a live open question, not yet explained (see OPEN Q4).

STRUCTURAL DYNAMICS

POLITICAL & REGULATORY

IMPLICATIONS FOR SES

Thread: "US C-band spectrum auction / FCC Part 100 licensing overhaul" (as named in priors.md). Balance-sheet relevance rivals IRIS²: net SES capture = (57.5% × first-100MHz incentive pool) + (100% × remaining-60MHz incentive pool) − ~$3.6B clearing cost − CVR payout. Given the first-100MHz pool alone is estimated ≈$3.06B (of which SES keeps 57.5% ≈ $1.76B) and the remaining 60 MHz is 100% SES's at a presumably similar per-MHz rate (~$1.8B, if the rate holds — unconfirmed), gross SES-side proceeds plausibly run $3.5-4B+ against a ~$3.75B all-in cost estimate — i.e. this reads close to breakeven on cost/proceeds alone before accounting for the multi-year capacity loss, which is the part SES visibly weights more heavily (see THESIS). This is a tighter and more skeptical read than "clearing proceeds vs $3.6B costs vs capacity loss determines whether this is net-positive" — the arithmetic above suggests net-positive is not obvious even before capacity loss is priced in.

ANALYTIC STANCE

Trust the mirrored ECFS filing METADATA (filer, date, doc title, filing type) directly over press summaries of "who asked for what" — but filenames alone can't disprove a content-level claim, only fail to corroborate it, as this run's SES/D2D tension shows. When direct PDF/article fetch 403-blocks (as it did again this cycle for every fcc.gov, lightreading.com, and satellitetoday.com URL attempted), WebSearch's synthesized snippets are a usable but lower-confidence substitute — good enough to surface a number or a named claim, not good enough to resolve a direct textual dispute (see OPEN Q3). Notable and worth logging at next retro: the FCC's public ECFS API (publicapi.fcc.gov) returned clean 200s this cycle even while the direct fcc.gov/docs.fcc.gov document-serving paths and most trade-press domains 403-blocked — treat the API and the document server as separately-gated egress paths, not one blanket FCC block. Known blind spot closed this run: the $3.6B clearing-cost estimate now has a comparison point (round 1's realized ~$5.33M/MHz vs round 2's implied ~$22.5M/MHz) — but the gap is flagged, not explained (OPEN Q4). New blind spot opened: Eutelsat's round-2 economic position is untouched by this dossier and by the daily brief so far.

OPEN QUESTIONS

SUPERSEDED

(none yet — dossier created 2026-07-05; this run refined rather than reversed the standing thesis)

TRACKED SERIES

REVISION HISTORY