A $40M Laser on a Robot Truck Tests the Cost-Per-Shot Bet

Forty million dollars, a laser that reaches past one kilometre, and a robot truck that can swat a drone out of the air without a human pulling a trigger. Aurelius Systems brought all three to the Association of the U.S. Army annual meeting in Washington this month.

The San Francisco company mounted its Archimedes turret on American Rheinmetall's Ox unmanned ground vehicle — the first public integration since the two firms announced a partnership in July.

The economics flip when the magazine is the grid

A laser does not carry ammunition. It carries a power budget. That distinction is the whole investment case, and it is why Archimedes matters more as a cost curve than as a physics demo.

Aurelius says the turret combines a high-power laser with machine vision and radar, then finds, tracks and burns down Group 1 and Group 2 drones at ranges beyond a kilometre without an operator starting each engagement. The company builds most of the system from commercially available parts and describes it as compact enough to be field-repaired rather than shipped back to a depot.

The contrast is an interceptor. A short-range missile or guided round costs tens of thousands of dollars per shot, and the magazine runs out. A laser spends electricity, so the marginal engagement is cheap until the power or cooling system gives up. The White House's fiscal 2027 budget proposed more than $2 billion for directed-energy development, a signal that the Pentagon sees the exchange ratio as a strategic problem rather than an accounting one.

AeroVironment has already pushed the same logic into contracts. Its LOCUST laser is marketed at under $10 per shot, and the company booked a first international order worth more than $50 million in September, six days after the U.S. Army awarded it a $464.8 million production agreement for the Enduring-High Energy Laser programme.

That is the demand-side proof. The supply side is messier.

Two poles, one procurement decision

Pole one is distributed defence at consumable prices. Put a cheap autonomous turret on every logistics vehicle, every perimeter post and every substation, and let each one defend itself. No central magazine, no reload convoy, no single point of failure. Aurelius chief executive Michael LaFramboise frames the manufacturing ambition in exactly those terms.

We're focused on being the Ford Motor Company in the space, making F-150s, and tens of thousands of them.— Michael LaFramboise, CEO and co-founder, Aurelius Systems

Pole two is the prime integration path. Power, thermal management, certification, sustainment and programme-of-record access are not trade-show problems. They are the reasons the Army's largest laser money still flows through established contractors. A cargo robot that already spends its electrical budget on mobility has less headroom than a dedicated air-defence vehicle, and a laser that draws from a battery, generator or grid needs a supply chain behind every option.

The two poles are not exclusive. Layered defence can absorb both: cheap autonomous effectors at the edge, expensive high-energy systems where power and range justify them. The live question is where the next marginal dollar goes. Aurelius is arguing that the answer should be many small units rather than a few large ones.

What the Ox integration actually proves

Rheinmetall's Ox is one of two vehicles competing in the Army's Ground Optionally Autonomous Transport programme, a cargo contest for machines that haul supplies over broken ground. Adding a laser turns a logistics asset into a self-protecting one. That is a narrower claim than "laser air defence," and a more credible one.

The integration story rests on software as much as hardware. At an Army integration exercise at Fort Carson, Aurelius connected Archimedes to Anduril's Lattice and the Army's Forward Area Air Defense command-and-control system in about 48 hours, according to LaFramboise. The company had already put the turret through T-REX 26-2, a Pentagon experimentation event at Camp Atterbury, where it engaged more than 20 quadcopters and more than five Army-supplied drones it had not seen before.

As we wrote in October, the first ground-autonomy production contract went to Overland AI and its IBEX vehicle. That deal showed the Army will put a robot truck on a programme of record before the autonomy stack is fully settled. Aurelius is trying to attach a weapon to the same wedge.

Why commercial parts are the real wedge

The laser is the visible component. The supply chain is the durable advantage. Aurelius says Archimedes is built largely from commercially available parts, which shortens production cycles and makes field repair plausible. In a category where a single bespoke optical module can set delivery back by months, that choice matters more than a few percentage points of beam power.

The company pushed the logic further in April, announcing a plan to build defense-grade fiber-laser source modules in the United States through a new manufacturing division. Vertical integration is expensive and slow. It also turns a systems integrator into a supplier with something to sell even when a platform bid loses.

That is the part investors tend to underwrite. A turret programme lives or dies with one vehicle competition. A component business can ride several. The $40 million round is small enough that a single production line and a handful of platform wins can change the revenue base, and large enough that the company can afford to own the laser source rather than queue for it.

The bear case is boring and decisive

No Army contract for Archimedes has been announced. The $40 million Series A is real money for a 19-person startup and small money against a $464.8 million prime production award. Aurelius has to convert a partnership, a trade-show display and a successful exercise into a programme office signature before the funding cycle turns.

Power is the physical constraint. A cargo robot that must climb grades and carry a tonne of supplies cannot devote unlimited kilowatts to a laser. Aurelius says Archimedes can draw from a grid, generator or battery, which is flexible on paper and demanding in practice: every source adds weight, heat and a logistics tail.

Then there is the competition. Rheinmetall's Ox is one of two GOAT contenders, not the winner. A laser integration helps the bid, but it does not guarantee the platform reaches volume. The same holds for Aurelius: its turret can be integrated on other vehicles, and the company's open-architecture pitch is designed for exactly that, yet every additional platform is another qualification cycle.

The number to watch is not peak power

Peak power makes a good specification sheet. Cost per engagement makes a business. Aurelius is betting that the second number decides the category, and that a commercially built turret can undercut a bespoke system the way a production truck undercuts a hand-built prototype.

The near-term catalysts are specific. A GOAT down-select would tell us whether the Ox becomes a volume platform. A first Army order for Archimedes would tell us whether the laser is a component or a programme. A manufacturing line for domestic fiber-laser modules, which Aurelius announced in April, would tell us whether the supply chain can scale without a single overseas bottleneck.

Until one of those lands, the honest read is a well-timed demonstration with credible economics and an unproven programme path. That is not a dismissal. It is the stage where defence technology either becomes a supplier or becomes a slide.