The Pentagon wants a constellation of satellites that can shoot down enemy missiles moments after launch, before they ever leave the boost phase. Two MIT physicists just ran the numbers, and they say the arithmetic doesn't survive contact with orbital mechanics.

In an analysis published September 2, 2026, in the Bulletin of the Atomic Scientists, Lisbeth Gronlund and David Wright of MIT's Laboratory for Nuclear Security and Policy took a hard look at "Brilliant Swarms," a Booz Allen Hamilton-developed concept for the space-based interceptor layer of the Golden Dome missile shield. The proposal calls for roughly 2,000 satellites weighing 40 to 80 kilograms apiece. Gronlund and Wright's calculations put the real number closer to 1,100 kilograms per satellite — nearly 14 times heavier than the upper end of Booz Allen's own advertised range, and as much as 27 times heavier than the lower end.

Where the Extra Mass Comes From

The gap starts with the kill vehicle itself. Gronlund and Wright calculate that a kill vehicle capable of closing on and destroying a boosting missile needs roughly 95 kilograms of mass on its own — propellant, sensors, guidance electronics, and structure. That's already more than the entire satellite budget in the Brilliant Swarms concept. Add the rest of the interceptor — the booster stage needed to launch the kill vehicle off the satellite bus with enough velocity and the bus itself, with power, communications, and station-keeping fuel — and the total climbs past 1,100 kilograms.

The math compounds from there. Because boost phase is short — Gronlund and Wright note that U.S. and Russian long-range, solid-fuel missiles burn out in about three minutes, with China's believed to do the same, while North Korea's newer Hwasong-18 burns out in as little as 170 seconds — the constellation has to be large enough that a satellite carrying an interceptor is already positioned near any missile launch site on Earth at the moment of launch. Gronlund and Wright estimate that intercepting a single missile in boost phase would require roughly 4,200 satellites in orbit. Defending against a five-missile salvo — a far more realistic threat scenario — pushes that to about 21,000 interceptor satellites, representing some 24,000 tons of hardware in orbit.

The price tag scales accordingly. Booz Allen has publicly put initial deployment cost at roughly $25 billion. Gronlund and Wright's own recalculation — accounting for the mass overrun and for the fact that satellites in low Earth orbit degrade and need replacement every four to five years, meaning the constellation isn't a one-time purchase but a permanent, recurring expense — puts total costs at roughly $85 billion over 20 years.

And even fully built, Gronlund and Wright argue the system has a built-in vulnerability: current missiles can already maneuver during boost phase, which shifts the intercept point the defense has to hit, and an adversary could also field cheap decoy boosters that force the constellation to expend real interceptors on fake targets, multiplying the effective cost of the offense-defense exchange in the attacker's favor. The two physicists note the Pentagon has itself studied a further countermeasure: designing future long-range missiles with boosters that burn out in under a minute, shrinking the intercept window even further.

Testing Guetlein's Own Red Line

What makes the analysis pointed rather than just academic is that it directly targets a threshold the Pentagon's own Golden Dome director set for himself. In April 15, 2026 testimony before the House Armed Services Committee's Strategic Forces Subcommittee, Gen. Michael Guetlein — director of Golden Dome for America — told lawmakers that "if boost-phase intercept from space is not affordable and scalable, we will not produce it, because we have other options to get after it," according to Breaking Defense's account of the hearing. In his prepared written statement for that same hearing, Guetlein framed the stakes in similar terms, describing a strategic environment in which "our Nation faces a new generation of missile threats that are faster, more maneuverable, and more numerous than ever before."

At that same hearing, Guetlein testified alongside Missile Defense Agency director Lt. Gen. Heath Collins, Army Space and Missile Defense Command's Lt. Gen. Sean Gainey, and Assistant Secretary of Defense for Space Policy Marc Berkowitz. In his written statement, Guetlein described the Space-Based Interceptor Program as using a "prize-based acquisition strategy" designed to reward industry performance and open the competition to non-traditional defense companies, framed within a broader Golden Dome strategy he said "remains centered on affordable and scalable capabilities."

Gronlund and Wright's analysis is, in effect, an attempt to test that affordability-and-scalability standard before the Pentagon reaches its own verdict. If their figures hold up, Brilliant Swarms fails Guetlein's own test on both counts: the mass overrun would balloon the constellation size — and therefore the cost — far beyond what Booz Allen has advertised, and Gronlund and Wright's $85 billion, 20-year estimate dwarfs the $25 billion Booz Allen has publicly cited.

Why It Matters

Golden Dome is being pitched as a generational missile-defense system, and its space-based boost-phase layer is the piece that would, in theory, catch missiles at their most vulnerable moment — right after launch, before decoys and countermeasures deploy and before multiple warheads separate. That's a real technical advantage over intercepting missiles in mid-course or terminal phase. But the reason no country has ever fielded such a system is precisely the physics Gronlund and Wright lay out: boost phase is brief, Earth is big, and a constellation has to be enormous to guarantee a satellite is in the right place at the right time.

This isn't the first time outside physicists have clashed with Pentagon mass estimates for space-based missile defense — similar debates dogged the Reagan-era Strategic Defense Initiative decades ago. What's different now is that the program's own director has publicly committed to an affordability and scalability test as the gating condition for whether space-based interceptors get built at all. That makes independent mass and cost estimates like this one directly relevant to a decision still pending inside the Pentagon, rather than a purely academic dispute. If Golden Dome's own program office is weighing that decision using satellite mass figures more than an order of magnitude lower than what outside analysis suggests, the budget numbers lawmakers are working from — and the timeline for when, or whether, a workable system reaches orbit — may need a serious second look.

Sources