George Gilder: Trump in Space
In some respects, the President's Golden Dome missile shield makes him even more Reagan's true heir than his tax policy.
This essay is free, but with Premium Membership you get MORE. Join today.
NOTE: My friend George Gilder was Ronald Reagan’s most-quoted living author and one of the principal fathers of the Supply-Side Economics that powered the Reagan boom and ended — for almost half a century now — the postwar every-two-and-a-half-year cycle of recession and “recovery”. His Information Theory of Economics is transformative. We published his must-read “Trump is Reagan’s True Heir” last week.
But George is also one of the genuine prophets of the Information Age, and he was also present at the creation of Reagan’s Strategic Defense Initiative (SDI, or “Star Wars”). So this essay is particularly significant. — RDM
by George Gilder
July 29, 2025
Ronald Reagan redefined economics and ushered in 40 years of prosperity, low inflation, and low interest rates by his supply-side tax cutting program. President Trump is continuing and enhancing that legacy.
Even more revolutionary, and certainly further ahead of its time than Reagan’s tax reforms, was President Reagan’s Strategic Defense Initiative. Initially intended to render a first strike nuclear attack implausible by blowing many of the incoming missiles out of the sky, SDI was the first fruit of the rise of the integrated circuit and its overthrow of the previous supremacy of offensive over defensive weapons systems.
As I explained to Reagan in his office in the early-1980s, the integrated circuit would replace the “balance of terror” with what is now dubbed AI (artificial intelligence) in defense. To make my point, I handed President Reagan a 16K DRAM chip from Micron Technologies of Idaho that I soon celebrated in The Spirit of Enterprise before it went public, and that almost 50 years later remains a core holding in our portfolio.
Potatoes on the Moon
I thus modestly claim that I at least partially won the debate with Peter Drucker, who declared at the time, as I recall, that “making memory chips in Idaho was like growing coconuts in South Dakota,” or, I might add, potatoes on the moon. I say I only partially won the debate because his larger point that microchips are intrinsically a global attainment turns out to be true today when we still depend on China, Taiwan, Israel, and Europe for key advances. Even in the United States, our industry relies on engineers from China who comprise close to half of the total U.S. endowment of microchip engineers.
But just as the U.S. presidents between Reagan and Trump did little to advance Reagan’s economic policy — though they benefited from leaving it largely undisturbed — SDI was mostly abandoned by subsequent administrations.
There was some excuse for this. Four decades ago, the underlying technologies were far from ready. Typical microchips in the early 1980s sported 100,000 transistors. Today, the Apple M4 chip contains 28 billion. In the 1980s, fiber optics was still in early stages and the United States was not far along the laser learning curve. Targeting was primitive. Even if the lasers had been ready, hitting a target perhaps thousands of miles away — an ICBM in boost phase — would require accuracy within 10 -5 of a degree. Daunting was the prospect of central control of a counterattack against hundreds of missiles — or thousands after the MIRVs had separated from their booster — although as I told Reagan, microchips would enable distributed control.
Still, as the Israelis demonstrated with Iron Dome, more could have been done with determined leadership. Yet even Iron Dome and the U.S. Navy’s ever more awesome Aegis system for defending carriers fall well short of Reagan’s dream, which was to shoot down ICBMs in their boost phase.
Since the MIRVS coming off an ICBM are directed at various targets, taking down ICBMs in boost would make it impossible for the enemy to be confident that every U.S. missile site would be targeted by multiple incoming missiles, the standard for a successful first strike. That’s the standard because a first strike is insane if the enemy retains the capacity to shoot back. Even a partly successful missile defense system, striking in the boost phase, takes a first strike off the board. Finally, that dream is possible. Thanks to the progress of artificial intelligence we have the technology, and Trump clearly has the will.
A Brilliant Swarm
Among the systems being considered for elements of Golden Dome is something called Brilliant Swarm, a crowd of small kinetic satellites launched from space aimed at enemy missiles in the boost phase. Equipped with artificial intelligence and sensors of a precision impossible even a decade ago, the satellites are autonomous and able to adjust their own flights to track moving targets. Communicating across the swarm, different clusters opt for the targets they can most easily reach and destroy. There is no need for centralized control.
Just as in the past when we have seen massive increases in computational power appear at first in some centralized location—the telephone network, main frame computers, server farms—the Law of the Microcosm inevitably asserts itself: intelligence migrates to the edge of the network, enabled by the drastic shrinkage of foot print and power consumption, and demanded by the limits of lightspeed.
In Reagan’s day the dream was to use lasers in space rather than kinetic weapons. Golden Dome will do just the opposite.
While the Brilliant Swarm seeks out incoming missiles in boost phase, additional defense layers will defend against weapons at lower altitude. Kinetics will be used, but anti-missile missiles are expensive, often a multiple of the cost of the bogie they destroy. Here lasers — ground based — will come into their own. Israel has already used “Iron Beam” lasers as part of Iron Dome; the Navy’s Aegis system now includes laser weapons that have successfully shot down ballistic missiles at the edge of the atmosphere.
Ground-based lasers have few power constraints. They can power through the atmosphere, especially with AI making instantaneous corrections. Costing somewhere between a few cents and a few dollars per shot, they are orders of magnitude cheaper than any incoming weapon.
Trump’s Edge
In addition to AI and dramatically better lasers, Trump has one more advantage Reagan never enjoyed: A private space industry, so much more efficient than NASA that private companies, including SpaceX, now handle most launches for both NASA and the U.S. Space Force.
Space Force itself is another ratification of Reagan’s vision. Taking over responsibility for missile warning, satellite protection and operations, space launches and astronaut missions, the 15,000-person force institutionalizes space defense as a bureaucratic interest.
Because AI is the enabler of almost everything the Space Force does and Golden Dome will do, the contribution of private firms to the effort goes far beyond space launches. In addition to SpaceX, names like Northrop Grumman and Lockheed Martin are no surprise. Palantir and Anduril are deep in the mix, but so are firms such as Booz Allen Hamilton, which christened Brilliant Swarm, and Shield AI. These and dozens more ventures, many less than a decade old, are all fruits of defense contractors founded with the mission of using AI to protect U.S. military forces.
Early critics of Reagan’s SDI howled that the system could never stop every incoming nuke, exposing millions to fiery death. That missed the key point of SDI: deterring a first strike. In the age of AI, however, there is a real prospect of a defense that catches every incoming missile, rendering the ballistic missile obsolete.
So, both Drucker and I were right after all. The microchips that exalt defense over offense are ultimately dependent on the new world order that they make possible and that Israel epitomizes in the maelstrom of the Middle East.
Interesting, and I do hope the system works as promised. Just one question, part facetious, part serious: How do the system developers plan to deal with Murphy's Law? You know, "Anything that can go wrong will, at the worst possible moment."
Anyone who has worked in tech, or even with tech has had to handle that problem. Perhaps overwhelm it with massive redundancy?