Monetizing a Weapon You Don't Build: How Palantir and Anduril Play the Drone Defense Game via Epirus Leonidas

The opening question is blunt: how can Palantir and Anduril get paid for a weapon they don't build , and what does one example reveal about their larger opportunity? David Carbutt looks for the answer not in the hardware itself but in the software layers that make hardware useful. The case in point is Epirus ; last year the company said its Leonidas system brought down 49 drones in a single high-power microwave pulse , with two different drone types flying together in that swarm. It was a controlled company demonstration , on its own range, yet the promise is clear: instead of firing a separate interceptor for every target, you sweep the sky with directed energy.

Leonidas is attractive because it flips the firing logic . The system sends microwave energy meant to disrupt electronics and can steer that energy across an area; the story that both drone types in the swarm went down sits at the center of the pitch. Put yourself on a base: the first wave is not the last, and missiles are scarce. If Leonidas can keep firing on electricity and cooling alone , you can leave some targets to that channel and husband expensive interceptors for harder threats. The metric therefore shifts from how many fell to what the defender still has left to use .

The cost asymmetry frames the economics: the video pairs a roughly $1 million missile against a $500 drone with Epirus talking about pennies per engagement . That line is about the variable cost of a pulse, not the full price of buying and operating the system. Still, if the effect proves dependable, a commander gains a reason to route suitable drones to the microwave channel and preserve missiles for other tasks , which in turn creates a reason to change the existing defense and give the software that ties it together more to do.

How does a new weapon plug into the base? Lattice's job

The practical knot begins here: how does a new weapon work with what the base already has? Anduril's Lattice is presented as an answer — software that connects sensors and fielded equipment as a network . Epirus announced a Leonidas–Lattice integration for the Marine Corps a few years ago, giving the microwave weapon a path to receive cueing through Anduril software . If you know where the incoming drone is, you can bring the right effector to bear; if that information arrives late, even the best weapon fires at empty sky.

The idea that a better weapon from someone else can make existing software more valuable grows from that link. The customer has already bought gear to detect and to defeat ; Epirus now offers a different way to handle a swarm , and before buying, the customer wants to know how it fits the existing infrastructure . The Lattice link begins to answer that without requiring Anduril to build the microwave weapon itself . Carbutt stresses not a single-device tool but a layer that remains useful as hardware changes .

Palantir approaches the same problem through information . Its Maven Smart System fuses imagery, sensor returns and reporting into a coherent view that helps staff prioritize where to send scarce assets . The pair said in late 2024 the platforms would be interlinked to accelerate AI for national security — Lattice operating next to sensors and shooters , Maven structuring the broader operational picture for commanders. The video stresses the nuance: these pacts do not field a turnkey Palantir–Leonidas weapon and do not grant independent release authority ; they show why interoperation is strategically valuable .

Is anyone actually paying? The $87 million anchor

The next test is whether anyone will pay for software on its own . Anduril already has an answer: the initial award within the Army's March 2026 Anduril deal directed roughly $87 million toward Lattice in support of the Pentagon's counter-UAS task force . The wider arrangement carries a $20 billion ceiling through March 2036 , a contract vehicle that permits orders up to that level rather than recognized revenue; Breaking Defense on March 16, 2026 quoted Anduril's president saying “there is no money attached yet, this is just a vehicle but it removes friction.” In other words, a paying software buyer for counter-drone work exists , distinct from the appeal of any single weapon.

Why that order makes sense is told through inventory and factory logic : if Lattice brings different sensors and effectors onto one network , a commander can allocate scarce resources more accurately even as inventory turns over, so an outside improvement can still be an inside opportunity — yet counting every Leonidas sale as software revenue would be a mistake. In March 2025 Epirus said it was joining Palantir's Warp Speed customer program in Cohort 2 to use that software for manufacturing workflows and supply chain management , the factory-side of the same thesis: when a single missing component blocks shipment, you must find the bottleneck, ask the supplier to deliver sooner or resequence work to give a credible delivery date , and Warp Speed is framed as the software for managing that information.

Two different customer problems split here. Maven is for the staff who need to understand what is happening and decide where to send limited defenses ; Warp Speed is for the manufacturer who needs to understand what can be built and when . In the Epirus case Palantir's role starts before Leonidas leaves the factory , a revenue lane separate from battlefield decision support. The announcement does not disclose what Epirus pays Palantir nor how much output has improved, but it makes tangible a problem outside the battlefield : a defense builder must turn plans into deliverable products , and software can earn its keep there as well.

The broader prize is what the video generalizes: whenever a new technology arrives, who gets paid to make it useful alongside what the customer already owns , and who helps the companies behind that technology actually build and deliver it . A firm can compete by building the best piece of equipment , or by supplying software that helps a customer bring different pieces together . The Leonidas–Lattice link shows they are not mutually exclusive. Epirus supplies the weapon, Lattice connects it to the wider defense net ; an improvement made outside Anduril can still create an opportunity for Anduril's software . For the buyer, the goal is dependable protection , not a single device, and connections matter as much as components.

The final check remains operational, not demonstrative : the technology must deliver under the conditions where the customer wants to use it . The video notes Epirus hardware has been tested with troops in the Middle East and the Pacific , including reported trials in the Philippines in April 2025 with military personnel , and that such experience helps define what works and what needs fixing . Further purchases would be another demand signal, but software suppliers must still prove their own business . For Anduril, the counter-drone software order covers a wider role than one Leonidas integration ; the fate of a single weapon does not explain Lattice demand. For Palantir, Epirus is one relationship to track among many , not proof of the whole thesis, which is why the video argues an investment case needs meaningful financial performance at the right price , not just an exciting demo.