The IAEA believes a little over 200 kilograms of Iran’s 60% enriched uranium sits in a tunnel complex under Isfahan. The bombers hit Isfahan in June 2025. They collapsed the tunnel entrances. They didn’t reach the uranium. The inspectors haven’t been back since.
Everyone knows how to settle it. Send people in, dig it out, fly it home. Former NATO commander James Stavridis told the Wall Street Journal it could be the largest special operations mission in history. It would be under drone and missile fire, near a major city. Asked about ground troops, Trump said the US wouldn’t do it now, “maybe later.”
The uranium is still there. The problem wasn’t finding the target. It was the boots.
That problem is going away.
This isn’t about how to fight China. It’s about how to deal with badly behaving populations of 50 million or more when there’s no appetite for 200,000 boots on the ground.
There are many badly behaving countries in the world with populations large enough that we can’t send boots in all the time to police them. Even a few million people is hard to control.
The Real Constraint
The US didn’t lose Iraq or Afghanistan on the battlefield. It took Baghdad in three weeks. It lost the will to stay. Nearly 4,500 Americans died in Iraq and more than 2,400 in Afghanistan.
After that, every American threat carried an asterisk. Air strikes, yes. Ground troops, no. Iran has 90 million people. Nobody in Washington proposes to occupy it, and Tehran knows it.
Winning by air is difficult. Air power can destroy a building. It can’t hold a town, search a tunnel, or make sure the thing it hit stays dead.
If we wanted to control bad behavior in Afghanistan, we didn’t need nation building. With a robot army, we could have policed the camps and left the country to itself.
The Robots Are Already Here
Ukraine is building the answer now, at scale. Its ground robots ran about 112,000 frontline missions in the first eight months of this year, more than 25,000 of them in August alone. Ukraine has contracted more than 22,000 ground robots this year. The Defense Ministry’s stated goal is for robots to handle 100% of frontline logistics.
They haul ammunition, food and water. They evacuate the wounded. Every mission is a soldier who didn’t walk that road.
A robot costs $2 million when you build a thousand. Those are special order. When you build 10 million you are getting consumer product prices.
Ukraine is building tens of thousands. In the first four months of this year it spent about $330 million through its frontline ordering system alone on more than 181,000 drones, ground robots and electronic-warfare systems, under $2,000 apiece.
The supply line can be robotic too. From 2011 to 2014 the Marines flew unmanned K-MAX cargo helicopters in Afghanistan. They lifted more than 4.5 million pounds and flew thousands of missions without a pilot on board. That was over a decade ago.
Robots that break can be picked up by robots and flown back to be fixed. Extend that far enough and you get robotic aircraft carriers with their own robotic support. The tail that used to need thousands of people doesn’t.
Making a robot army is not an overnight project.
The Threat That Works
In 1963 John F. Kennedy warned that the President in the 1970s might face 15, 20 or 25 nations with nuclear weapons. Sixty years later there are nine.
Part of the reason is force. Israel destroyed Iraq’s reactor in 1981 and Syria’s in 2007. The US and Israel struck Iran’s program in June 2025 and again this year.
Part of it is timing. The 2007 National Intelligence Estimate concluded Iran halted its nuclear weapons program in the fall of 2003. American troops had reached Baghdad that April. In December 2003 Libya agreed to give up its weapons programs.
The threat was credible then. It stopped being credible when the public stopped accepting casualties.
Iran wouldn’t have grown to where it is if this technology had been available. Diplomacy that avoided conflict could have been a simpler message: a threat to go in with a robot army.
North Korea is the same story. There was no appetite to go in, and sanctions don’t work well on large populations. It tested its first nuclear weapon in 2006.
Robots put the credibility back. A President who can send a robotic force into Isfahan, dig out the uranium and bring it home, without a single flag-draped coffin, doesn’t have to say “maybe later.” And the other side knows he doesn’t.
We don’t even need robots to dig out the uranium. We can monitor the land over it. If the Iranians try to dig tunnels to reach it, we kill them. There are lots of ways to guard things. The robots become a physical lock on the site: a kill zone over the way in, not an occupying army.
Look at What China Is Doing
China is building real robot armies now, and putting them on display.
At its Victory Day parade in Beijing in September 2025, the People’s Liberation Army rolled out armed unmanned ground vehicles, a robotic mine-clearing vehicle and “robot wolves,” four-legged combat robots. In January, state television released footage of armed robot dogs in training. PLA exercises have used robot wolves in urban-warfare drills and beach landings, sending them in first to take the initial fire from coastal defenses.
Some are built on commercial four-legged robots like Unitree’s Go2, fitted with rifles or rockets.
China may be ahead of us in that thinking.
The US Army halted its Robotic Combat Vehicle program in 2025 and reopened it to industry. Developers told National Defense magazine this year that large-scale autonomous ground operations are at least a decade away.
In Summary
This isn’t how to fight China. That needs army level firepower.
This is for worldwide police work without boots on the ground.