On 15 September Rocket Lab told the SEC that the $3.6 billion 364-day bridge commitment behind its Iridium acquisition "was formally terminated in full." The same report says the equity programme finished at approximately $1.944 billion of gross proceeds on approximately 29.3 million shares, and Iridium's $1,774.7 million term loan now stays outstanding after closing. Four days earlier we priced that bridge at $285,669 thousand of interest a year. There is no bridge now, so there is no subtraction.
A bridge at 8.0% on $3.6bn became 25 to 75 basis points on $1.8bn
Start with what Iridium already pays. Its term loan bears SOFR plus 2.25% with a 0.75% floor, and Iridium's own quarterly report expects interest on it of approximately $90.0 million over the twelve months from 30 June. Consent and Amendment No. 4 raises that margin only at closing, and only to a range — SOFR plus 2.50% to 3.00%, on Rocket Lab's credit ratings. The coupon is not printable yet. The step is, on the $1,774.7 million outstanding at 30 June:
| Margin step at closing | Extra interest a year | Post-close interest |
|---|---|---|
| +25 bp, to SOFR + 2.50% | $4.4M | $94.4M |
| +50 bp | $8.9M | $98.9M |
| +75 bp, to SOFR + 3.00% | $13.3M | $103.3M |
$1,774.7m × 0.25% = $4.4m; × 0.75% = $13.3m. The ratings grid is the only assumption doing work, so all three are printed; the top still lands an order of magnitude below the bridge. As extra interest over what Iridium pays now, August cost $285.7m − $90.0m = $195.7 million a year. September costs $4.4m to $13.3m.
The residual stops being tens of millions
Iridium generated $185.8 million of operating cash flow in the first half of 2026 against $51.8 million of capital expenditure: $185.8m − $51.8m = $134.0 million, after $41.9 million of its own interest. On 11 September we added that interest back, because the pro forma extinguished the debt behind it, and set $172.6m against $141.4m of bridge interest: $31.2 million.
The add-back is now wrong. Iridium's debt survives, so its reported cash flow is already net of that interest, and the only deduction left is half a year of the step: $134.0m − $2.2m to $6.7m = $127.3 million to $131.8 million. Against the roughly $300 million the Raymond James note was reported to credit Iridium with, a full year of the step leaves $286.7m to $295.6m — within 5% of where it started.
The equity cost 4.90% of the company, which is permanent
29.3m ÷ 598,350,482 outstanding on the 5 August cover of Rocket Lab's quarterly report = 4.90%; $1,944m ÷ 29.3m = about $66.35 a share. Both are the filing's own "approximately", and net proceeds after commissions are disclosed nowhere, so the cash is smaller than it reads.
The counter-case is that nothing got cheaper, only differently priced. Take the disclosed $4,882.8 million of closing cash outflows, drop the $1,774.7m pay-off, add back the $183.4 million Aireon seller loan and $154.7 million of Aireon term loans the pro forma excludes and expects Rocket Lab to repay at closing: about $3,446.2 million. Gross proceeds cover 56% of that, and the remaining $1.5 billion comes off a balance sheet holding $2.39 billion of cash and securities at 30 June. A 364-day bridge would have come down; 4.90% of the company does not.
Our 25 August piece was superseded, not misread
It reported that the pro forma treats Iridium's term loan as repaid, taking the cash requirement to $4.88 billion. The exhibit said that on 13 August and says it still; the deal changed underneath it. The 11 September piece is the one whose conclusion dies, because its $31.2 million prices a borrowing that will never be drawn. Its closing line held: "If Rocket Lab prices notes or a term loan well inside 8.0%, or funds more of the price with equity, the arithmetic above improves." Rocket Lab did both.
Our model assumed the equity path and the company took it
In August we called the financing route the largest open question in the file, and put the filing's debt path at about $1.50 a share against our model, which assumed equity. It closed our way. Restruck, the pro forma count is 598.35m + 29.3m + 41.26m of stock leg + 40.9m of preferred = 709.8 million against the model's 705.0 million, 0.68% high. Net cash is untouched: assuming $1,774.7m of loan and repaying it cost a balance sheet the same. The gap is the $338.1 million of Aireon debt the model does not carry, or $0.48 a share — maintenance, not an article edit.
What overturns the finding is Rocket Lab deciding it does not want Iridium's loan: the amendment prices keeping it at 25 to 75 basis points and prices leaving at 1.00% twice, a repricing premium and a post-anniversary exit fee of about $17.7 million each. Iridium's stockholders vote on 24 September, the FCC transfer is outstanding, and closing is still guided to mid-2027.
Rocket Lab's current report of 15 September 2026 and its press release carry the bridge termination, the completed equity programme and its share count, Consent and Amendment No. 4's carve-out, ratings grid, exit fee and repricing premium, the unsecured guarantee from Rocket Lab USA and the sufficiency statement. The whole August structure — the bridge draw and its rate, the pro forma interest, the term-loan pay-off, the closing cash outflows, the Aireon balances — is the pro forma exhibit of 13 August 2026. Iridium's margin, expected term-loan interest, half-year cash flow and capital expenditure are its own second-quarter report, which also withdrew guidance; Rocket Lab's share count and liquidity are its. The Raymond James figure is press-reported, not a note we hold. Every step and conversion is ours and gross of undisclosed commissions; the pro forma share count and fair value are our model's assumptions.
