Listen to the article
A response to the Financial Times
By André Pienaar
The Financial Times has published a careful and well-sourced case against Golden Dome, President Trump’s comprehensive space-based air defence system designed to protect the US homeland. In making its case against the feasibility and merits of Golden Dome, the FT marshals the standard objections: the engineering is formidable, the price is understated and the politics will not hold. Although these are serious points for consideration, the FT’s analysis is focused on the wrong question.
The implicit test the article applies is the one inherited from President Reagan’s Strategic Defense Initiative (SDI) debates of the 1980s: can this system reliably defeat a full-scale, dedicated salvo from a peer nuclear power?
But that has never been the design objective, and it is the wrong question. We do not judge air defence by whether it makes a country completely invulnerable. We judge it by whether it complicates an adversary’s planning, raises the number of weapons required for a given objective, denies cheap coercive options and buys leaders decision time in a crisis. Against those criteria, the calculation looks quite different and the case for Golden Dome is compelling.
The threat that is actually being defended against
Strip away the ICBM framing and look at what has changed in the homeland threat picture over the past decade.
The FT article’s observation that no ballistic missile has ever been launched at the continental United States is presented as evidence of low risk. It is more accurately evidence that deterrence has worked so far, under conditions that are now rapidly changing.
The United States today has effectively no continental defence against cruise missiles launched from submarines or converted merchant hulls, none against hypersonic glide vehicles, and a ground-based interceptor fleet that is sized for a limited rogue-state launch rather than for a coercive strike below the threshold of general nuclear war. Russia has fielded and used intermediate-range systems in Ukraine specifically to demonstrate reach. China is expanding its warhead stockpile at a pace that will materially change the strategic arithmetic before 2030. North Korea’s ICBM programme is no longer aspirational and has intercontinental reach.
The scenario that should concern planners is not Armageddon. It is a limited, deniable or coercive strike used to extract a political concession during a crisis over Taiwan or the Baltic. The most tiered combined air attack that Russia launched on Ukraine on 13–14 May involved 1,567 strike drones and 56 ballistic and cruise missiles. The largest single integrated air raid in the first war with Iran remains 13–14 April 2024, when Iran fired 120 ballistic missiles, 170 drones and 30 cruise missiles – a combined total of circa 320 airborne weapons – at Israel. A system that cannot stop thousands of warheads can very plausibly stop such a limited strike, and it is the ability to stop a limited strike that removes the coercive option from the table.
The cost comparison is not like for like
The gap between the $185bn programme figure and the Congressional Budget Office’s (CBO) $1.2tn is real, but it is not a gap created by spin.
The CBO modelled a notional full architecture, including a mature space-based interceptor layer, priced over 20 years of development, procurement and sustainment. The $185bn programme budget describes only an initial capability. These are different phases of development, and quoting them side by side without that qualification does the reader a disservice.
Even taking the higher number at face value, $1.2tn over two decades is roughly $60bn a year – under five per cent of current annual US defence spending. It is comparable in scale to the lifecycle cost of a single tactical aircraft programme. Whether that is good value depends entirely on what one believes the value of the US homeland as a safe sanctuary for the American people to be.
The physics argument has aged badly
The most quoted objection to space-based interception is the absentee ratio: because satellites orbit and launch sites do not, only a small fraction of any constellation is over the right place at the right moment, so the constellation must be enormous. The mathematics is correct. What has changed is what “enormous” costs.
The scepticism dates from an era when launch ran at roughly $20,000 per kilogram to orbit. A constellation of several thousand satellites was considered a thought experiment. Reusable heavy lift has cut that figure by more than an order of magnitude and is on track to cut it again.
A single commercial operator now manufactures and flies a constellation larger than every other satellite ever launched combined. Serial production of small, cheap, mass-produced spacecraft – the precise industrial capability the absentee-ratio problem demands – has moved from impossible to routine within a decade.
This does not make space-based interception easy. It moves the constraint from orbital mechanics to manufacturing throughput and launch cadence, which are engineering and capital problems rather than physical ones. That is a meaningful distinction, and the article does not make it.
Defence does not start an arms race that is already running
The stability objection holds that building defences will provoke adversaries into building more offence. It assumes a baseline of restraint that no longer exists. Our adversaries’ hostile intent, combined with new capabilities, has been in evidence for years. China’s warhead expansion, Russia’s nuclear-powered cruise missile and underwater delivery systems, and North Korea’s ICBM testing all predate Golden Dome by several years. One cannot destabilise a competition that is already fully under way. The relevant question is whether the United States meets that competition with offence alone, or with a mix that includes denial and defence and restores the safety of the US homeland.
There is a related point the article overlooks. Japan and Canada are in discussions about participation. Allied integration into a shared sensing and command architecture is arguably the most strategically consequential element of the programme, and the least discussed. It potentially binds allied territory into a common early-warning picture in a way that no bilateral arrangement currently achieves.
Regardless of its sceptics, Golden Dome is moving forward at pace. Consider the sequence. An executive order in January 2025 established the programme office. A public commitment followed in May 2025 with a three-year fielding target. By January 2026, the Missile Defense Agency had approved 2,440 vendors out of 2,463 applicants under the SHIELD multiple-award vehicle, having added 340 companies in the third tranche alone. By April 2026, the Space Force had placed 20 contracts worth up to $3.2bn with 12 companies to develop space-based interceptor capability. By July 2026, the Space Development Agency had signed deals worth up to $1.75bn for 36 missile-tracking satellites, scaling proliferated missile-warning coverage faster than its original timeline suggested. Congress had appropriated roughly $38bn across reconciliation and FY26 before any of it.
Eighteen months from executive order to a qualified industrial base of two and a half thousand firms, with money moving to prime and non-traditional suppliers alike, is not how the American defence enterprise normally behaves. The comparison that matters is not Iron Dome or SDI but the F-35, where the interval between requirement and initial operating capability ran to 15 years, or the Ground-Based Midcourse Defense programme, which took a decade to field an interceptor that has never been produced at scale.
The mechanism deserves credit. The Pentagon has moved away from cost-plus contracting and the practice of waiting for technologies to mature before committing. The SHIELD structure pre-qualifies a very wide field and then competes task orders inside it, which means a capability requirement can be converted into a funded contract in weeks rather than in budget cycles. That the qualified base spans traditional primes, commercial space companies and start-ups is not incidental. It creates competitive pressure inside a programme rather than between programmes, and it gives the government multiple technical paths to the same effect rather than a single vendor’s roadmap and a single vendor’s schedule risk.
The breadth cuts both ways. Spreading the work across a wide supplier base is a strength because it draws on proven capacity. The risk is that integrating dozens of programmes into one seamless system is precisely the sort of challenge that drives cost and schedule overruns. But it is a better problem than the one it replaced. A programme that fails at integration can be fixed by hard systems engineering. A programme that fails because it took 12 years to award a contract cannot be fixed at all – it will be overtaken by the reality of the threat.
Whatever happens to the interceptor layer, the United States now has a validated, pre-competed contracting pathway for space and missile defence work, a much larger set of firms cleared to bid on it, and a tracking constellation being built ahead of schedule. Those assets outlive any single administration. If Golden Dome’s legacy turns out to be that the Pentagon learned how to buy, build and innovate quickly again to meet a near and present danger, that alone would justify a considerable share of what has been spent.
