It Doesn't Have to Work
In 1993, a small team inside the new Clinton administration quietly killed a weapon that had never fired a shot in anger.
It was called Brilliant Pebbles. The idea was almost childishly simple, which is part of why it was so seductive. You would put a constellation of small autonomous interceptors into low orbit — somewhere between seven hundred and a thousand of them, circling the planet about two hundred and fifty miles up. They carried no warhead. No explosive. The plan was to have them detect an enemy missile climbing out of the atmosphere and simply fall on it, hitting it at roughly five miles a second. Metal into metal. Kinetic kill. Physics doing the work that chemistry used to do.
It was one of the most developed concepts to come out of Reagan’s Strategic Defense Initiative — the program a Democratic senator had mocked, on the day after Reagan’s 1983 speech, as “reckless Star Wars schemes.” The name stuck. The hardware did not. Between 1990 and 1992, the kinetic-kill concept was tested three times. All three were judged failures, to varying degrees. When Clinton’s people arrived, they zeroed the line item and redirected the money to interceptors you could bolt to the ground.
By then, the Strategic Defense Initiative had spent something on the order of thirty billion dollars across the 1980s.
I want you to hold that last number in your head, because it is the whole point of this edition, and almost nobody framing the current debate is looking at it.
Thirty billion dollars. On a system that was cancelled. That never worked. That, by the verdict of its own test program, could not be made to work with the technology of its time.
And here is the part that should make you sit up.
The architecture I just described — a proliferated constellation of small kinetic interceptors in low Earth orbit, no explosives, falling on missiles in their early flight — is, almost line for line, what the United States government started buying again this spring.
The thing that already happened
On January 27, 2025, in his first week back in office, Donald Trump signed an executive order for what was briefly called “Iron Dome for America.” Within weeks it was rebranded Golden Dome. On May 20, in the Oval Office, he put a number on it: roughly a hundred and seventy-five billion dollars, operational, he said, by the time he left office in 2029.
That was the press-conference version. The version that matters to an investor is quieter and more recent.
In July 2025, the reconciliation law carried roughly twenty-five billion dollars in first money for the program. On February 3, 2026, the fiscal 2026 defense appropriations added another $13.4 billion tied to space and missile defense for Golden Dome. And in late April, the Space Force confirmed what it had been doing more or less in the dark over the preceding months: twenty contracts, worth up to $3.2 billion, awarded to twelve companies to build prototypes of the one component everyone agrees is the hard part — space-based interceptors. The same kinetic-kill satellites. The orbital killers are supposed to demonstrate an “initial capability” by 2028.
So the catalyst is not a speech. The catalyst is that, this quarter, the program crossed the line that almost every grand defense announcement dies before reaching. It stopped being a slide. The dollars are obligated. The contracts have names on them.
That distinction — slide versus signed contract — is where the money is, and I’ll come back to it.
The man who said the quiet part
The person responsible for turning the order into hardware is a Space Force four-star named Michael Guetlein, confirmed by the Senate in a voice vote on July 17, 2025, with the deliberately unglamorous title of “direct reporting program manager.”
He is not a politician’s pick in the cynical sense. Guetlein is an acquisition man down to the bone — a mechanical and aerospace engineer by training, former head of Space Systems Command, former deputy director of the National Reconnaissance Office, a former program executive at the Missile Defense Agency. If you wanted one person in the United States who understands both the physics of hitting a missile with a satellite and the bureaucracy of paying for it, you would struggle to do better.
Which is exactly why one thing he has said matters more than anything Trump has said.
Guetlein has compared Golden Dome, in scale and ambition, to the Manhattan Project. That is the line that got quoted. It is not the line that counts.
The line that counts is this one, on whether the space-based interceptors can actually be built: “I believe we have proven every element of the physics that we can work. What we have not proven is, first, can I do it economically, and then second, can I do it at scale?”
Read that again, slowly, because the program’s own architect is telling you where the risk lives.
Not in the physics. In the economics, and in the scale.
He is right. And the moment you take him at his word, the entire public debate reveals itself as a conversation about the wrong variable.
The wrong question
Open any serious treatment of Golden Dome and you will find the same argument, conducted with great energy, about whether the thing will work.
The skeptics have the better of it, on the merits. A February 2025 report from the American Physical Society — not a partisan outfit, a body of physicists — laid out the problem that has not changed since the 1960s. To kill a missile in its boost phase, the bright, slow, vulnerable couple of minutes after launch, you have a window of two to four minutes and you need an interceptor close enough to reach it. From orbit, “close enough” means you need a vast number of satellites, because most of them are over the wrong ocean at the moment of launch. The Congressional Budget Office’s costing assumes roughly 7,800 interceptors in orbit, and that constellation alone accounts for more than sixty percent of its price tag.
Then there is the oldest counter-argument in the field. The cost-exchange ratio. If my interceptor costs more than your missile, you win simply by building more missiles than I can afford to shoot down. A saturation attack does not need to be clever. It needs to be cheap. Strategists understood this by the end of the 1960s; it is most of the reason the 1972 Anti-Ballistic Missile Treaty existed at all.
This is the debate. Will it work. Can the physics close. Is it a fantasy.
And for a citizen, or a strategist, or a physicist, it is the right debate to have.
For an investor, it is a category error.
Here is the claim I want to plant in the free portion of this edition, where you can hold me to it: whether Golden Dome ever intercepts a single missile is very nearly irrelevant to the financial outcome. The two questions — does it work, does it pay — have come apart. And the historical record does not merely suggest they have come apart. It proves it.
Go back to the thirty billion dollars.
The Strategic Defense Initiative was, by the verdict of its own tests, a failure. And it was one of the most consequential capital-allocation events in the history of the American defense industrial base. It funded a generation of sensor work, kinetic-kill engineering, and systems integration. It seeded contractors, careers, laboratories, and an entire institutional appetite that never went away — the same Michael Griffin who ran SDI technology later stood up, in the first Trump term, the Space Development Agency now woven into this program. The shield was cancelled. The spending compounded. The money was real whether or not the laser ever fired.
A weapon has to work. An annuity only has to be paid.
Golden Dome, for the names attached to it, is being built as an annuity.
Why this reaches you
You do not need to care about missile defense to have a stake in this. I want to be precise about why.
We are about to watch a multi-decade, politically entrenched spending commitment get written into the base of the federal budget — and the range of plausible totals is not a rounding error. The official figure crept up to a hundred and eighty-five billion dollars in March 2026, by Guetlein’s own admission, after he was asked to fold in “additional space capabilities.” The CBO models a trillion-two over twenty years. The American Enterprise Institute, costing the most expansive architecture, gets to roughly three-point-six trillion. Adjusted for inflation, more.
That is the spread on a single program: from a hundred and eighty-five billion to several trillion, depending on choices not yet locked. Almost nothing else in the public markets right now offers that combination — a near-guaranteed floor of demand, a ceiling measured in trillions, and a political coalition with every incentive to keep the line item alive across administrations.
A capital cycle of that shape, that duration, with that kind of bipartisan inertia behind the spending if not the rhetoric, does not come along often. The last one that rhymed with it — the post-Ukraine European rearmament — is the other subject that has done the most for this newsletter, and it worked for exactly the same reason. The narrative was geopolitical. The trade was the capital cycle underneath it.
The phone in your pocket runs on a defense-seeded technology — GPS came out of the same lineage. You are already, in a sense, a customer of the last shield that “didn’t work.” The question this time is not whether you are exposed. It is whether you are positioned.
What I am watching, and what I am not
I am not watching for a successful intercept test. It would be nice. It is not the thesis.
I am watching the seam between the slide and the signed contract, because that is where the rent is actually allocated — and it is messier than the headline suggests. In April 2026, the trade press was blunt: a year in, the program was, in one outlet’s phrase, “spinning its wheels,” tangled in internal disagreement over the architecture. Lawmakers complained, on the record, that the Pentagon had taken twenty-odd billion dollars and failed to tell Congress in any detail how it would be spent. Guetlein himself has talked about “pathways to pivot” if the interceptor timeline slips.
To most readers, that is a story about dysfunction. To me it is the most important investment signal in the whole file, and it points directly at which of the twelve names actually collects.
Because the contractor list is not a flat field. It runs from the primes you would expect — Lockheed Martin, Northrop Grumman, RTX’s Raytheon, General Dynamics — through the new-space insurgents the headlines love — SpaceX, Anduril, Palantir on the command-and-control “glue” layer — down to small specialist firms most readers have never heard of, the Turions and Quindars and True Anomalies, several of them privately held. They are not the same trade. They do not have the same payoff. And they will not survive the same “pivot.”
That is where the work is. And it is where I have to stop the free portion.
What the rest of this edition does
If you have read this far, you already have the frame, and the frame is the valuable part: Golden Dome is not a bet on a shield. It is a bet on an annuity, and the annuity pays on appropriations, not on intercepts. That single reframe is worth more than most of what you will read about this program this year, because it tells you to stop watching the test range and start watching the budget committee.
What you do not yet have is the part I would actually put money behind.
In the paid section, I lay out which of the twelve contractors are positioned as annuitants and which are decoration — the ones that get renewed across a “pivot” versus the ones that were funded to look like competition. I walk through why the cost-exchange physics, far from killing the thesis, tells you precisely which layer of the architecture gets protected when the budget tightens, and which layer gets quietly cut. I put numbers on the realistic obligated spend over the next thirty-six months, against the milestones that release it — the 2028 demonstration is the one to circle. And I name the scenarios where even the annuity thesis breaks: the specific political and fiscal events that could zero this line the way 1993 zeroed Brilliant Pebbles, and what you would watch for to get out first.
The shield may never work. I am fairly sure parts of it can’t.
The money is already moving. That part is not in doubt, and that part is the trade.