Rocket Lab Values Iridium's Spectrum at $3.32B—More Than the Entire Constellation.

Rocket Lab ($RKLB) has told the market what it thinks Iridium's spectrum is worth. It did so on 13 August, in the pro forma section of the registration statement for the merger — where it has sat unquoted through two weeks of coverage of the collar, the bridge and the exchange ratio. We missed it too: our own piece eight days later said no such number existed.

The number is $3,315,500,000. It is the preliminary fair value Rocket Lab's own purchase accounting assigns to a line item called simply "Spectrum," and it is the largest single asset in the deal.

Two comparisons make it land. Rocket Lab marks Iridium's entire operational LEO constellation and its ground segment at $1,927,018,000 — so the licence is worth 1.72 times the hardware. And on 1 April 2026, the last trading day before the Financial Times reported the Amazon–Globalstar transaction and Iridium's shares began to move, the whole company's stock was worth $3.02 billion at $28.52 across its 105,956,272 shares. The spectrum mark alone is 110% of what the equity market said the entire business was worth five months ago.

The allocation, which is the article

The card above the article carries it in full.

Five things fall out of it, and the first is the only one that has travelled.

• Spectrum is 43.7% of the price — of the $7,589,872,000 of total consideration, and about 41% of the $8.0 billion enterprise value the two companies quote.
• It is indefinite-lived, so it never amortises. Only the $1.01 billion of technology and customer relationships runs through the income statement, at $67.6 million a year. The largest asset Rocket Lab is buying is the one that will never appear as an expense — but it does carry an annual impairment test, and at this size an impairment would be a very large one.
• The constellation is valued at replacement cost. Rocket Lab says it used "a replacement cost analysis" for the property, plant and equipment, i.e. what it would cost to build the thing — and Rocket Lab is one of a handful of companies on earth that could actually build it. That it still comes out below the licence is the point.
• The $1.125 billion deferred tax liability is the mechanical consequence of writing intangibles up by $4.247 billion over their carrying value. It is not a cash cost, but it is a real measure of how much of this price is a step-up rather than a going concern.
• Goodwill is $2.202 billion, 29.0% of the price — the part Rocket Lab could not attach to any identifiable asset at all.

Where the number came from, which is SpaceX

The background section of the same filing supplies the provenance, and it is the second half of the story.

Iridium's board did not go looking for a buyer. On 23 September 2025 — two weeks after SpaceX ($SPCX) announced its purchase of EchoStar's spectrum licences — it met and discussed, in the filing's own words, "the prospect that the SpaceX-EchoStar Transaction and the valuation of the spectrum licences of EchoStar Corporation implied by the SpaceX-EchoStar Transaction could lead to interest from third parties with respect to the acquisition of Iridium or its spectrum assets at valuations that Iridium and its stockholders might find attractive." On that basis the board authorised management to entertain unsolicited approaches.

This deal exists because of a print SpaceX set, by the target board's own account. We priced that print in August: roughly $19.6 billion for about 65 MHz of nationwide US mid-band, against AT&T's $23 billion for about 50 MHz from the same seller.

This is where most analysis would divide one by the other. We are not going to. The filing discloses no megahertz figure for Iridium, and the two assets are not the same kind of thing: EchoStar's licences are technology-neutral, flexible-use, nationwide US mid-band that a terrestrial carrier could deploy tomorrow; Iridium's is a narrow, globally harmonised mobile-satellite allocation coordinated through the ITU and usable in every jurisdiction on earth. A dollars-per-megahertz ratio between them would be a number, but it would not be a measurement, and anyone quoting one has invented it. What can be said without inventing anything is the total: Rocket Lab's own accountants marked Iridium's spectrum at about 17% of what SpaceX paid EchoStar. That is a comparison of two cheques, not of two assets, and it is the only honest form the comparison takes.

The two things this corrects in our own coverage

We published a piece on this deal on 21 August and it got two things wrong, both because we read the merger agreement and the announcement rather than the pro forma notes filed eight days earlier.

First, we refused to value the spectrum, on the grounds that "it has no disclosed valuation, no comparable transaction inside this agreement, and any figure quoted for it today is invented." The refusal was right in principle and wrong on the facts: a disclosed valuation existed, prepared by the buyer, and it is the figure above.

Second, we said Iridium's $1.775 billion term loan looked likely to stay in place. The filed pro forma assumes the opposite. "Pay-off of Iridium's existing debt" of $1,774,721,000 is a component of the purchase consideration, and it comes straight out of cash at closing. The filing does keep the door open — it notes amendments to Iridium's existing debt agreements "which, if completed, could reduce borrowings under the Bridge Facility" — but the numbers as filed assume the loan is repaid.

That second correction changes the size of the deal's cash leg materially.

The cash requirement is $4.88 billion, not $2.86 billion

We framed the cash half of this deal as the $2.861 billion paid to Iridium's shareholders at $27.00 a share. The pro forma adds three more outflows at closing:

Against that: a committed $3.6 billion bridge facility, and $2.39 billion of cash, cash equivalents and marketable securities on Rocket Lab's balance sheet at 30 June. The second chart above puts the three bars side by side. The two funding sources together cover it, and leave the combined company with essentially nothing.

The bridge is now priced. The filing states an effective interest rate of approximately 8.0%, producing $285.7 million of pro forma interest expense for a full year and $141.4 million for the June half, with a disclosed sensitivity of $4.55 million per 12.5 basis points. Issuance costs are $25.7 million.

What the combined company looks like

This is the part the pro forma statements are actually for, and it is not what the deal's framing suggests.

Revenue roughly doubles. Gross margin goes up three points, because Iridium's service revenue carries better margins than launch. The operating loss barely moves — Iridium's operating income very nearly offsets the amortisation and stock-compensation adjustments the acquisition creates.

And then the net loss widens by $60.4 million in a single half year, entirely on financing.

Put the two numbers next to each other. Iridium earned $31.3 million of net income in that half. The bridge costs $141.4 million over the same period. Rocket Lab is acquiring a profitable business and the combination is less profitable than the buyer was alone, because the way it is being paid for costs four and a half times what the business earns.

That is not a criticism of the deal — a 364-day bridge is scaffolding, not the building, and the whole point of the equity distribution agreement Rocket Lab opened on 13 August is to take it down. But it is the shape of the company for as long as the scaffolding is up, and the pro forma is the first document to state it in dollars.

Every valuation of Iridium sits below the price

The filing also publishes the work Iridium's financial adviser did. Four independent methods, all struck against management's own forecasts, and not one of them reaches the $54.00 a share the merger consideration is worth:

The discounted cash flow used a 9.0–10.0% cost of capital and a 2.5–3.5% perpetuity growth rate, on a ten-year forecast. Its top is $8.46 below the price.

The gap between the highest standalone valuation and what Rocket Lab is paying is the spectrum, the strategic fit, and whatever competitive tension the process produced. The adviser did not price the spectrum separately either. The only party that has put a number on it is the buyer, in its purchase accounting, after the fact — which is exactly why that number is worth reporting and exactly why it should not be treated as a market price.

For completeness, the same adviser ran Rocket Lab: discounting a 2031 enterprise-value-to-revenue range of 20–40x back at a 15–18% cost of equity produced $46.23 to $102.45 a share, against the $84.54 close on 26 June. Our own model bases at $33.23. We are well below the bottom of that range and comfortable there; the difference is almost entirely the exit multiple, and 20x is not a number we would defend on a business we think converges toward space hardware economics.

Reconciling against our model

Our Rocket Lab model carries Iridium as a separate vertical that earns nothing until the acquisition closes, built from the bottom up: 2.627 million billable subscribers at 30 June, $20.47 of monthly revenue each, 36,000 net additions a quarter ramping toward a ceiling, plus subscriber equipment and engineering services. All of that is ours — Iridium has never published a forecast.

Until now. The filing discloses management's own standalone projections through 2035, and this is the first time anyone outside the deal has seen them:

Our revenue build is close to theirs, and that is a real check. These figures include Iridium's Aireon transaction; a second forecast in the same filing, prepared before Aireon, puts 2027 revenue at $896 million against $1,039 million — a $143 million difference. Our model projects $967 million for 2028. Holding that Aireon delta flat gives roughly $956 million on a like-for-like basis, so two builds constructed with no knowledge of each other land within about 1%. Ours is 8–12% below the Aireon-inclusive figures across 2028–2030, which is what it should be, because our model does not include Aireon.

Our margins were too low, and we have raised them. Management's forecast implies an EBITDA margin of 55.5% in 2027 rising to 64.2% by 2035; our model ran 45.0% gliding to 50.0% because nothing better was public. It now carries management's own figures, which is worth +$0.49 a share. Against that we have taken $247 million off net cash for the transaction and issuance costs paid at closing, which the model had omitted — −$0.35. Two changes on the best disclosure available and the base moves $33.09 to $33.23.

That is the honest scale of what a hundred pages of pro forma detail does to a valuation, and it is worth sitting with. Doubling the Iridium vertical's capital intensity would cost −$0.28. The assumptions we most wanted to revise move the number by pennies.

The finding is what the model cannot hold. The Iridium vertical is worth $9.16 of the $33.23 base — 27.6% of it — and dropping it entirely takes fair value to $24.07. That $9.16 is the discounted value of subscriber cash flows and nothing else. It contains no spectrum asset, because a revenue-times-margin model has nowhere to put one.

Our model assumes the combined company carries 705.0 million shares. On that count Rocket Lab is paying $10.77 a share for Iridium, and we value the resulting operating business at $9.16. The gap is not a verdict on the price — it is the part of the price that has no home in a model built from revenue, and 43.7% of it is the spectrum, worth $4.70 a share on the same count.

So the filing moves our number by fourteen cents and changes what we think the deal is about: the entire case for this acquisition sits in the one line item that cannot be modelled from a revenue forecast. Anyone who tells you Rocket Lab is overpaying on nine times sales, or underpaying because L-band is priceless, is arguing about the $3.32 billion — and until Rocket Lab says what it intends to do with the licence, nobody outside the company can settle it.

One more consequence, in the model's own valuation section. Its exit multiple is 9.0x revenue, and its stated reason was that this is what Rocket Lab itself agreed to pay for a mature space-comms business — the cleanest kind of comparable, because it is a transaction rather than a screen. The purchase accounting shows that 41% of that $8.0 billion is a licence mark, so the 9.0x is not a clean read on operating economics. We have kept the multiple and corrected the reason: the licence is not severable from the revenue it produces, so stripping it out and re-dividing would double-count. But every 1.0x on that exit multiple is worth $3.77 a share — twelve times what any Iridium operating assumption in this file moves — and it is now anchored to a number that is partly an accounting estimate.

One further note on our own construction. Our share count and net cash are pro forma for the equity-financed path: 705.0 million shares, assuming the $1.94 billion equity programme is fully drawn, and net cash of roughly negative $580 million once the closing costs above are taken out. The filing's pro forma takes the other path — the bridge stays as debt at closing and no equity is drawn. On our own arithmetic that swap is worth about $1.50 a share against us, from a smaller share count offset by $2.2 billion more net debt. Which path Rocket Lab actually takes is the largest open question in this file, and it is unanswered.

What to watch

1. The 24 September 2026 special meeting. Iridium's stockholders vote at 8:30 a.m. Eastern, virtually, on a record date of 21 August. A fairness opinion whose every method tops out below the price is the standard raw material for opposition, and none has surfaced yet.
2. What replaces the $3.6 billion bridge, and when. Permanent debt at anything near 8.0% costs $285.7 million a year against a company that has never run leveraged; equity does not, but dilutes. The filing assumes debt. The company's own actions since 13 August suggest equity. Both cannot be right.
3. Whether Iridium's $1.775 billion term loan is amended to survive. The pro forma repays it. If the amendments succeed instead, $1.775 billion of the cash requirement disappears and the bridge shrinks with it.
4. The 2031 free-cash-flow trough in management's own forecast. Unlevered free cash flow goes to $(21) million in 2031 on constellation replacement capital expenditure and does not regain its 2030 level anywhere in the forecast. Note that the ten-year version of this forecast was not provided to Rocket Lab — only the run through 2030 was.
5. Any statement from Rocket Lab about what the L-band licence is for. The purchase accounting says it is 43.7% of the price. Nothing public says what it earns.

The purchase price allocation, the merger consideration and its $1,774,721,000 debt pay-off, the $4,882,776,000 of closing cash outflows, the bridge facility's approximately 8.0% effective rate and its interest and sensitivity figures, the pro forma income statements, Iridium management's 2026–2035 forecasts, the financial adviser's valuation ranges and the 23 September 2025 board discussion of the SpaceX–EchoStar transaction are all disclosed in the merger registration statement filed 13 August 2026 and the amendment of 24 August 2026 that set the 24 September meeting. Iridium's 105,956,272 shares and the $28.52 close of 1 April come from the same filing; Rocket Lab's $2.39 billion of cash and securities at 30 June is from its own second-quarter results. Every percentage, per-share conversion and ratio here is R40 arithmetic on those figures. The fair value of $33.23, the $9.16 Iridium contribution, the $967 million 2028 projection, the sensitivities and the $1.50 financing-path estimate are our model's — assumptions, not disclosures, and the model was revised on the day of publication to take in the disclosed margins and closing costs. The SpaceX–EchoStar consideration of roughly $19.6 billion is press-reported across a $17–19.6 billion range.