We just don't know whether it would succeed.
The system, called Ground-based Midcourse Defense (GMD), is a work in progress. It has failed to destroy dummy warheads in six of 10 tests since becoming operational in 2004, but the two most recents tests succeeded. Here's how it works.
Satellites and radar detect and track it.
Infrared sensors on satellites detect the launch and the missile's path by its heat signature until its rockets burn out. Powerful X-band radar on U.S. ships and additional ground radar track the missile from below. If its target appears to be the United States, the GMD is activated.
An ICBM launched from North Korea would take about half an hour to reach the U.S. mainland, and about 20 minutes of that would be spent in the stratosphere — the "midcourse phase." This is where the GMD would try to intercept the warhead.
The missile releases the warhead — and decoys.
In the stratosphere, burned-out boosters fall away, releasing the warhead, a cloud of debris and any decoys meant to confuse interceptors. Simple Mylar balloons could be decoys, for instance, because in space, everything is equally weightless. These decoys are called "countermeasures."
After the most recent successful test, in May, the Pentagon upgraded its assessment of the GMD in a memo, saying the system "has demonstrated capability to defend the U.S. homeland from a small number of intermediate-range or intercontinental missile threats with simple countermeasures."
GMD launches interceptors
Military commanders would almost certainly launch two or four interceptors at each incoming ICBM to have a better chance at hitting the warhead.
The entire U.S. stockpile comprises 44 interceptors, housed at two GMD installations in Alaska and California. Interceptors are much more expensive to make than bombs and decoys, said Laura Grego, a physicist who has extensively researched the GMD for the Union of Concerned Scientists and wrote a 2016 paper critical of the system. Grego is among the experts are skeptical of the GMD, in part because successful tests have been conducted under controlled circumstances with relatively simple decoys — "scripted for success," she said.
"We're not asking it to defend us in realistic, real-world scenarios that it would be expected to be used in," Grego said, "so we don't even have the information to judge how well it would do."
The interceptor releases a 'kill vehicle''
Booster rockets separate as they burn out, eventually leaving the five-foot-long "kill vehicle" to hunt down the warhead in the stratosphere.
That kill vehicle, a new version that debuted with the May test, is one reason other experts are more optimistic about the GMD's chance of success.
Tom Karako, director of the Missile Defense Project for the Center for Strategic and International Studies, said the latest test was closer to reality than many people think. He said the checkered test record shows the natural progression of an evolving — and improving — technology. "The systems deployed for GMD today are different than what failed in 2010," Karako said.
The kill vehicle picks out the warhead.
The kill vehicle uses its sensors in addition to guidance from satellites and ultrasensitive ground radar to distinguish the warhead from the decoys and debris and to lock onto the missile's path.
Plans for a web of space-based sensors that could provide a closer view of the warhead and decoys have bounced around the Pentagon and Congress but have not yet been implemented, Karako said.
The kill vehicle slams into the warhead and destroys it.
On-board thrusters steer the kill vehicle into the warhead, shattering it on impact before the nuclear bomb has a chance to detonate. In the May test, everything went perfectly. But the consequences of less-than-perfect could be catastrophic.
"If North Korea sent six ICBM warheads at the United States and we got five of them, you'd say, 'Hey, five out of six, not bad,' " said Bruce W. MacDonald, former assistant director for national security at the White House Office of Science and Technology Policy. "But if you ended up losing Seattle . . . you'd still feel pretty bad even though it was 80 percent effective. That's what you're dealing with."
The bottom line is that relying on the tech is very risky.
"I don't think missile defense is going to save our bacon," Grego said. "We're going to have to solve [North Korea] diplomatically, and that's tough to hear. People don't like to hear that we're vulnerable."
Defense Department Missile Defense Agency; Center for Strategic and International Studies Missile Defense Project; Union of Concerned Scientists; 38 North; Lockheed Martin; Boeing; Navy News Service; "Nuclear Dynamics in a Multipolar Strategic Ballistic Missile Defense World," by Charles D. Ferguson and Bruce W. MacDonald, published by the Federation of American Scientists. Note: Animations are schematic.