Technology / Manufacturing · Commentary

The Pentagon Has Enough Killer Robots. It Needs Builder Ones.

By Jasper Burns and Grace Marchant

Published · 6 min read

The Pentagon is investing billions in autonomous weapons while overlooking the machines that keep those weapons operational.

In the opening hours of a war in the Pacific, long-range missiles strike runways at bases from North Japan to Guam, leaving them cratered and flight operations at a standstill. Many aircraft survive the strikes but cannot take off. A fighter does not have to be shot down to be taken out of the fight. The damaged concrete underneath is enough.

In September 2026, Defense Secretary Pete Hegseth announced plans for a new four-star Autonomous Warfare Command to accelerate the deployment of unmanned and robotic systems across the military.1 It's a welcome recognition that autonomy will shape future warfare. But the Pentagon should look beyond the machines that fight to those that build and repair the infrastructure they depend on.

That infrastructure challenge is growing as the Air Force prepares for such attacks. Under its doctrine of Agile Combat Employment, aircraft disperse from major bases to smaller airfields, remote islands, and even stretches of highway, making them harder to target.2 But dispersing aircraft also multiplies the infrastructure the military must build, maintain, and defend.

The stakes extend beyond manned aircraft. Tomorrow's wars may rely more heavily on drones, mobile missiles, and weapons not yet invented. Large drones like the MQ-9 Reaper still need prepared landing areas, even if unpaved.3 Mobile missile launchers need passable routes to shoot and scoot—relocating after firing to evade being targeted in a counterattack. Even drones that require no runway depend on facilities for storage, maintenance, and resupply. The weapons of tomorrow are uncertain. Their dependence on infrastructure is not.

Unlike investments in particular combat platforms, which may become obsolete as warfare changes, autonomous construction equipment could support virtually any future force structure.

The Pentagon has made autonomous combat systems a priority—it should apply similar urgency to the machines that build and restore the infrastructure supporting them.

The Engineering Bottleneck

Under the U.S. Air Force's Rapid Airfield Damage Recovery (RADR) process, engineers assess damage, clear debris, remove shattered pavement, fill craters, and restore usable surfaces following an attack.4

This is dangerous, skilled, and time-sensitive work. Excavators and bulldozers are indispensable, but their operators may have to sit inside these exposed machines while the threat of another strike persists. Dispersed operations compound the problem: more locations compete for the same limited engineering crews.

Fortunately, construction autonomy is no longer merely a laboratory promise.

In August 2026, San Francisco-based Bedrock Robotics announced that excavators equipped with its technology were working without onboard operators at commercial infrastructure projects in Nevada and Texas. Bedrock adds sensors and computing to conventional machines, avoiding the need for entirely new excavators.5 In October, Bedrock raised $600 million to expand those deployments.

The Pentagon need not develop autonomous construction machinery from scratch or replace its existing fleet. Much of this technology can be retrofitted onto equipment it already owns.

Other firms demonstrate different approaches. In Florida, three haul trucks equipped with technology from Teleo, now owned by defense-autonomy company Havoc, moved earth under the supervision of one operator located 40 miles away. The trucks handle repetitive routes autonomously and request assistance for more complex tasks.6 TerraFirma, an Austin-based startup, is deploying retrofitted excavators, bulldozers, and loaders under remote human supervision.7 AIM Intelligent Machines has developed autonomy systems for bulldozers and excavators, with machines incorporating its technology already operating commercially through a partnership with Komatsu.8

The Pentagon has noticed. In September 2025, the Air Force awarded AIM a nearly $5 million research-and-development contract for automated heavy machinery, with potential applications in airfield construction and repair.9,10 In June 2026, the Army also sought information about commercial autonomous capabilities for bulldozers, excavators, and skid steers.11

These are encouraging steps, but research contracts and market surveys are not equivalent to fielding equipment across operational units. The question is no longer whether the technology merits attention. It is whether the military will make it an ordinary part of its engineering capability.

Building the Future Battlefield

The military applications extend far beyond runways. Autonomous and remotely operated machines could clear damaged roads, grade sites for expeditionary bases, and dig protective berms. A small engineering detachment might eventually supervise several machines instead of supplying a separate operator for each.

In the Indo-Pacific, vast distances and austere conditions make deploying and sustaining skilled engineering crews particularly difficult. Autonomous equipment could reduce the personnel needed at each location, easing the logistical burden of dispersed operations.

Even if an autonomous excavator digs no faster than a human-operated one, it could help sustain around-the-clock construction without requiring operators to work exhausting shifts. Remote operation would also allow engineers to work away from exposed locations. The point is not to eliminate engineers, but to enhance their capabilities: more machines, more hours of operation, and fewer people in harm's way.

Commercial success does not guarantee battlefield readiness. Autonomous machines must operate reliably despite disrupted communications, GPS interference, and unpredictable terrain. These challenges call for military testing.

The opportunity extends beyond defense. Military demand could accelerate America's emerging construction robotics industry, improving technology with applications in civilian infrastructure. The Pentagon should build on existing commercial systems, test them under contested conditions, and begin fielding them across engineering units.

The next generation of military technology should not be limited to the weapons that fight. It should include the machines that make fighting possible.

Conclusion

America's ability to prevail will depend not only on the aircraft it launches or the missiles it fires, but on how quickly it can build, repair, and restore the infrastructure that keeps its forces in the fight. Autonomous construction equipment will not eliminate the need for military engineers, but it could allow fewer personnel to accomplish more while reducing their exposure to danger.

American companies are already demonstrating that this technology works. The Pentagon should move beyond experimentation and make autonomous construction a military priority.

The Pentagon has spent decades teaching machines how to destroy. It's time to teach them how to build.

Sources

  1. U.S. Department of War, “Project Agincourt – Pathway to the Autonomous Warfare Command,” memorandum, Sept. 30, 2026.
  2. U.S. Air Force, "ACE: Forging Multi-Capable Airmen in the Indo-Pacific," Aug. 8, 2025.
  3. U.S. Air Force, "MQ-9 Reaper Completes First Mission Using Dirt Landing Zone," June 21, 2023.
  4. U.S. Air Forces Central, "379th ECES Executes Ninth Air Force's Largest-Ever RADR Exercise," Feb. 6, 2025.
  5. Bedrock Robotics, "Bedrock Robotics Launches First Fully Autonomous Excavator Deployments on Critical U.S. Infrastructure Projects," GlobeNewswire, Aug. 17, 2026.
  6. Teleo, "Building on Innovation: Tomahawk Construction Turns to Teleo to Clear a New Path – From 40 Miles Away," Teleo, Apr. 25, 2024.
  7. J. Coyne, "TerraFirma: Built for Earth, Bound for Mars," Kleiner Perkins, July 15, 2026.
  8. Komatsu, "Komatsu and AIM Intelligent Machines Enter Strategic Partnership for Autonomous Operation of Bulldozers and Hydraulic Excavators," Komatsu, July 31, 2026.
  9. HigherGov, "Contract FA805125C0001: AIM Intelligent Machines," Sept. 25, 2025.
  10. T. Soper, "Air Force Awards $4.9M Contract to Seattle-Area Autonomous Construction Startup AIM," GeekWire, Jan. 22, 2026.
  11. U.S. Department of the Army, "Robotic and Autonomous Construction Equipment (RA-CE)," Request for Information, Solicitation No. PANDTA-26-P-0000RA-CE, June 8, 2026 (updated June 29, 2026).

About the authors

Jasper Burns

Jasper Burns is a former aviator in the U.S. Navy. He flew a variant of the F/A-18 fighter jet and led missions aboard the USS Carl Vinson in the West Pacific before serving in both the intelligence community and the White House. He is a Knight-Hennessy Scholar pursuing a JD/MBA at Stanford University. He also helped organize and served as master of ceremonies for Stanford DEFCON 2026.

Grace Marchant

Grace Marchant is a former U.S. Air Force Civil Engineer Officer who managed a $523 million portfolio of military construction projects in Japan and evaluated contingency airfields across Europe. She is a Knight-Hennessy Scholar pursuing a master's degree in civil and environmental engineering at Stanford University.

The views expressed are those of the authors and do not necessarily reflect the views of the Institute for American Manufacturing & Technology, the U.S. Department of Defense, or any other organization with which the authors are affiliated.

Research and commentary from The Institute for American Manufacturing & Technology.

Back to research
Stay informed

Research from the people building America.

Get Institute research, news, and event updates in your inbox. Confirm by email. Unsubscribe at any time.

Unaffiliated? Enter Independent.
Tell us more (optional)

By subscribing, you request Institute newsletter emails. Optional details help us understand our readers. Privacy policy.