First Trucks in Hand
Palantir handed the U.S. Army its first two AI-powered TITAN (Tactical Intelligence Targeting Access Node) targeting trucks on March 7, 2025, built to compress the spot-track-blast loop under a $178 million deal signed a year earlier. The contract buys 10 prototype nodes — five advanced, five basic. Palantir shipped one of each type this month, proving the verified count sits at two truck-and-trailer sets, not the brigade-scale rollout some headlines implied. The Army Contracting Command at Aberdeen Proving Ground announced the award on March 6, 2024.
The trucks feed the Army's Multi-Domain Task Force. A prototype reached the 1st Multi-Domain Task Force at Joint Base Lewis-McChord, Washington, last summer and entered service by December. The MDTF is the Army’s long-range precision-strike unit, designed to degrade adversary air defenses and stand-off weapons. Its mission depends on collapsing the time from detection to firing solution, exactly what the trucks target. The two new trucks join that unit, though Palantir kept the receiving company's name quiet. The program started in 2021. Raytheon and Palantir built early prototypes; Northrop Grumman built separate ones for the Tactical Exploitation of National Capabilities office and the Defense Innovation Unit. Palantir won the lead role after Project Convergence trials pushed satellite images straight to firing commanders.
The trucks do one job: find targets fast. Each TITAN pairs a truck with a trailer hauling generators and a satellite dish. Calibre Defence mapped the layout: a truck with a rear mission module and a trailer with generators and a satcom link. Palantir's TITAN page calls the truck the Army's next-generation AI ground station. It pulls feeds from satellites, planes, and ground sensors into one screen. It spits out targeting coordinates for long-range guns without the cloud. Palantir said the compute lives in the truck bed and that the goal is to shrink the sensor-to-shooter clock while sparing soldiers mental fatigue.
The two models ride different frames. The advanced node uses the Army's medium trucks and reads space sensors. The basic version bolts onto a Joint Light Tactical Vehicle. Each node pairs a truck with a trailer, so the first two shipments put four vehicles in the field. Akash Jain, Palantir USG's chief technology officer, said the advanced node needs two medium trucks; the basic needs two JLTVs, shipped across five orders.
Palantir leads the build — the first time a software shop has run a major Army hardware buy, CNBC reported. Palantir did not wire the trucks solo. It tapped Northrop Grumman, Anduril, L3Harris, Pacific Defense, Sierra Nevada, Strategic Technology Consulting, and World Wide Technology as partners. Jain called the deal a "leapfrog moment" for the U.S. Army. The goal: fuse sensor feeds with software automation at the edge.
No unit has fired a shot in combat to prove the claim. Army Recognition says TITAN replaces aging stations to speed precision strikes. Calibre Defence noted the MDTF's multi-domain theory hasn't faced live fire — a gap to watch. Palantir said the truck lets commanders pick targets without calling back to headquarters.
Palantir will finish the rest by 2026. The Army then picks full production, eyeing a buy of 100 to 150 units. The first two trucks are real steel now, but the brain that drives the targeting loop lives in the software stack described below.
The Truck's Software Brain
Palantir showed at AUSA 2025 (militaryembedded.com) that its TITAN truck runs a modular software stack ingesting disparate feeds and applying AI inside a rugged chassis. The Army funded Phase 3 of the prototype deal to build the mobile deep-sensing station. Forget the radio truck with bolted-on screens. Code commands the steel.
Software runs the show. Palantir fuses sensor feeds and network links to cut the sensor-to-shooter clock. It pulls signals, imagery, and ground pings into one track. That merge turns a blip into a target grid coordinate.
Inside the cab, Palantir Edge AI works offline or on thin links. Cameras and sensors sweep the ground. Vision models on the Palantir inference platform flag trucks, people, and ships. They tag and map contacts without a rear server. This on-board brain breaks the old rule: the truck needs no fixed command post. A ping hits the model; coordinates hit the gun line.
The steel comes from a team. SNC joined in June 2024 with TRAX software and rack-building help. L3Harris wires the node to Army guns. The other partners handle the remaining build slots. Together they built a scalable ground station.
The truck works while on the move, pushing firing solutions to the squad. Palantir said automation lifts the load off the crew; one track shows the shot. The Army aims to beat peer rivals by shrinking the kill chain.
The contract buys prototypes, but the build pattern changes the game. Palantir leads as a software prime — usually the vehicle maker bosses the coders. Here, the truck serves the code. That flip is why Palantir calls it AI-defined, not just a tuned truck. The software shop now ships the eyes that call the shots.
Palantir has not detailed power and cooling requirements. It won't say the exact ping-to-cue speed. The documented loop is ingest, fuse, infer, fire. That loop is the Army's answer to compressing the kill chain at the edge.
Who Drives the TITAN? Soldiers and Coders
The Army took the trucks, but the soldiers who run the targeting loop sit in a workforce the Pentagon hasn't defined. A 2024 GAO report (gao.gov) found the Pentagon mapped the AI workforce but set no owner or deadline. That hole opens just as TITAN puts AI on four wheels.
The truck needs two crews: soldiers who watch the node, and Palantir coders who feed it. Both show the strain of pushing decisions to the edge.
The crew trades paper range cards for a screen that fuses feeds and calls shots. Line of Departure argues the edge needs boxes that think alone, off-grid. Zequeira wrote in 2024 that edge AI speeds the Military Decision-Making Process. Crews must learn to trust the model, catch its errors, and check its math — not just fire tubes.
The Pentagon's Responsible AI Memo (media.defense.gov) names the AI workforce a core tenet. So do Warfighter Trust and Product Lifecycle. The policy exists, but GAO found no one owns the job definition. The TITAN crew has no standard yet. The first unit can't cite a Military Occupational Specialty for the seat.
Palantir carries the code load. Zero G Talent's job board lists 202 open roles at Palantir, paying a median of $170,000, from $52,000 to $200,000. The board held flat last week, but the posts show where Palantir spends its effort.
| Role | Location | Salary band (USD/year) |
|---|---|---|
| Software Engineer - Defense Applications | New York, NY | 145,000–200,000 |
| Software Engineer - Core Interfaces | New York, NY | 135,000–200,000 |
| Software Engineer - Hosted Model Infrastructure | Palo Alto, CA | 145,000–200,000 |
| Software Engineer - Hosted Model Infrastructure | Washington, D.C. | 145,000–200,000 |
| Software Engineer - Hosted Model Infrastructure | New York, NY | 145,000–200,000 |
| Deal Team - Business Affairs | Palo Alto, CA | 135,000–200,000 |
The salary bands reveal where Palantir invests to keep the trucks effective after delivery. The defense and infrastructure posts build the TITAN tools and keep the models on the truck. Palantir said its code runs the allied militaries' operating system for thousands of daily users. Those jobs feed the same targeting loop the trucks drive.
Palantir also hunts rookies. It posted internships by July 1, 2026, said new grads ship code on day one. The board stays flat, but the pipeline feeds the defense line.
Soldiers need new skills; Palantir needs coders. Both prove the TITAN is a software node needing a human who speaks its tongue at each end. Before the next trucks ship in 2026, the Pentagon must code the operator job — and the Army's top brass is already climbing aboard to see it.
Generals Tour, Raytheon Loses, Hill Stays Quiet
Days after the trucks rolled in, the Army's top brass climbed in. Secretary Dan Driscoll and Gen. Randy George toured them March 10–14, 2025, hitting Lewis-McChord, Irwin, and Huachuca (defence-blog.com). Driscoll's first stateside trip as secretary showed off the AI nodes. The Army skipped the quiet procurement act.
Col. Chris Anderson, the program manager, described the handover as the next phase to deliver revolutionary capabilities to warfighters. Army officials said the truck slashes detection-to-strike time. It fuses land, air, and space feeds to crush the old multi-echelon wait. The Army's contracting office (cpeisw.army.mil) confirms it reads all layers.
Palantir beat RTX (Raytheon's parent) in a three-year contest for the build. The Aberdeen contracting office used an Other Transaction deal — Congress's fast lane for non-traditional vendors. RTX built a rival but lost last March (CNBC). The loss shows the Army now picks code over steel for edge AI.
Old primes still weld the steel. Palantir's team adds Oshkosh Defense to the earlier partners (defence-blog). Northrop and L3Harris took sub roles under the software lead. Anduril, the defense tech startup founded by Palmer Luckey, is part of the team. New and old guard meet on the truck bed.
Capitol Hill stayed mute in our records. No lawmaker praised or panned the fielding. But the OTA law is theirs. The Army's first buy and 2026 vote prove the fast-tech law works.
Bryant Choung, Palantir's defense SVP, told C4ISRNET on March 5 the Army's setup let it field tech fast with soldier input. Primes used to own that loop via support contracts; now the coder does. Choung said soldier touchpoints steered the build.
Working in frontier tech? Zero G Talent tracks the openings: browse frontier tech jobs, openings at Palantir Technologies, and the people building the field.