NLOS-LS: the billion-dollar missile-in-a-box that was 9...

Most weapon programs that get killed are killed because they are years behind, wildly over budget, or never worked at all. The Non-Line-of-Sight Launch System was cancelled for a stranger reason: it was reportedly about 92 percent complete, had already consumed more than a billion dollars over roughly a decade, and simply could not shoot straight when the Army tested it under realistic conditions. That combination, nearly finished but not cost-effective, makes NLOS-LS a rare and instructive case in defense oversight.

This post is part of a continuing series tracing where public money actually goes. The rule is the same each time: pull the primary sources, keep the honest failure critique and the honest mission defense side by side, and let both verdicts stand. NLOS-LS earns a genuine critique on execution. It also earns a genuine defense on purpose, because the capability it was chasing was real, and it did not die with the program.

What NLOS-LS was: a missile in a box

The idea was elegant. Instead of a crewed launcher that has to be driven into position and exposed to return fire, NLOS-LS packaged precision missiles into a self-contained, network-controlled Container Launch Unit. Each CLU held 15 canistered missiles plus a self-locating, networked communications and computing package. You could set it on the ground, bolt it to a truck, or place it aboard a ship. It could locate itself, launch vertically and autonomously, and be fired and controlled remotely over the network, and multiple units could be linked together. The Army designated it XM501.

The concept, sometimes marketed as "missiles in a box," is documented on the XM501 NLOS-LS overview at Wikipedia and corroborated by the GlobalSecurity NetFires program page. Both describe the 15-missile CLU and its networked, remotely fired design.

The system was built by NetFires LLC, a joint venture of Raytheon in Tucson and Lockheed Martin. The division of labor matters and is easy to get wrong. Raytheon built the Precision Attack Missile, known as the PAM. Lockheed Martin built the Container Launch Unit and also developed the second munition, the Loitering Attack Missile, or LAM. The company roles are reported by the Arizona Daily Star and confirmed in the Wikipedia record.

The two munitions were designed for different jobs:

The PAM weighed about 117 pounds, reached roughly 40 kilometres (about 25 miles), and combined GPS and inertial navigation with an imaging-infrared seeker and semi-active laser homing, carrying a multi-mode warhead meant to defeat armor and bunkers.
The LAM was micro-turbojet powered, reached about 72 kilometres (about 45 miles), and could loiter for roughly 30 minutes to hunt targets autonomously.

Missile unit cost was cited at around 500,000 US dollars each. One point to keep straight: the LAM contract was terminated earlier and separately, after cost growth and poor test results, before the 2010 cancellation of the wider system. The PAM was the surviving munition that went into the culminating operational test.

Born inside Future Combat Systems, then spun out

NLOS-LS did not start as a standalone program. It was developed primarily for the Army's Future Combat Systems, the sprawling networked-force modernization effort. When FCS was largely cancelled and restructured in 2009, NLOS-LS was spun out and folded into the Early Infantry Brigade Combat Team, or E-IBCT, Increment 1, where its operational testing continued.

This is the single most important date distinction in the whole story. The Future Combat Systems core was cancelled in 2009. NLOS-LS itself survived that event and was cancelled separately in 2010. Treating the two as the same cancellation is the most common error made about this program. The Pentagon's own operational test office, DOT&E, covered NLOS-LS under the E-IBCT Increment 1 section of its FY2009 Annual Report, which places the program's continued testing squarely in the post-FCS window.

The turning point: White Sands, early 2010

The reckoning came at a Limited User Test at White Sands Missile Range, run from January 26 to February 5, 2010. This was the operational-realism test, the kind meant to show whether a system works the way soldiers would actually use it, not just on an engineer's range under ideal conditions.

It went badly. Of six Precision Attack Missiles fired, four failed to hit their targets. Only two struck. Two of the missiles that missed impacted more than 14 kilometres from the intended target, which is not a near-miss but a fundamental failure of guidance. New navigation software drove six of seven total system aborts. These figures trace to April 16, 2010 congressional testimony by David Duma, the OSD Principal Deputy Director of Operational Test and Evaluation, reported through Military.com / DefenseTech and corroborated across the Wikipedia and defense-press record.

The headline reliability number told the same story. Demonstrated missile reliability in the operational flight test was 61 percent, well below the 85 percent requirement the program was held to. The imaging-infrared seeker, one of the identified root causes of the failures, hit only 5 of 11 times across the 2009 to 2010 testing.

A note on the numbers, because two figures circulate and they are consistent rather than contradictory. As of the September 2009 DOT&E reporting, the cumulative record was about 62 percent, 8 of 13 shots. The 61 percent figure is the culminating value from the April 2010 testimony. Both describe the same trajectory of a missile that could not reliably meet its reliability bar. The 61 percent versus 85 percent comparison is the one to carry.

It is worth being fair about the record before this test. The program had not been an unbroken string of failures. By reported (Raytheon-adjacent) accounts, the PAM had succeeded in roughly 12 of 17 prior test shots, and there were genuine milestone successes, including a November 2008 CLU firing that hit a T-72 tank. That "12 of 17" tally comes from industry-adjacent reporting rather than an audited source, so treat it as reported, not confirmed; the audited-style DOT&E cumulative figure was the 8 of 13 noted above. Either way, the point holds: the collapse was concentrated in the culminating operational test, the one that counted most, not in every shot ever fired.

The cancellation timeline

After the failed test, events moved quickly, and the dates are worth stating precisely:

April 2010: the Army recommended cancellation of NLOS-LS.
May 3, 2010: the system was pulled from the Army website.
May 12, 2010: a House committee moved the program's research-and-development funds toward the Navy budget.
About May 18, 2010: the Defense Department approved the cancellation.
January 6, 2011: Defense Secretary Robert Gates's budget announcement reflected the program's end.

The 2010 cancellation is reported by Defense Systems, and the full timeline including the 2011 budget reference is laid out in the Wikipedia record. Because the formal death is sometimes pinned to the January 6, 2011 Gates announcement and sometimes to the May 2010 DoD approval, the safe statement is simply that NLOS-LS was cancelled in 2010, with the decision reflected in the following year's budget. What it was not is part of the 2009 FCS cancellation. That is a different program and a different year.

The money, kept honest

Here the record requires care, because this is exactly the kind of figure that gets rounded, inflated, or conflated.

NetFires was awarded a development-and-demonstration contract reported at about 1.1 billion US dollars in 2004. That is a contract ceiling for the develop-and-demonstrate phase, not a record of money spent. Press accounts separately describe "more than 1 billion dollars" invested over roughly ten years, and the program was described as about 92 percent complete at cancellation.

Two honest caveats belong on these numbers. First, the roughly 1 billion dollars is a reported, contract-derived figure, not an audited spent-to-date line item from a GAO report or a Selected Acquisition Report. Present it as approximate. Second, keep the two distinct meanings apart: the roughly 1.1 billion dollar 2004 contract award is a ceiling, while "more than 1 billion dollars invested over about a decade" is a reported spend. They point in the same direction, but they are not the same number, and neither is an audited total. The "92 percent complete" figure is program-office framing reported through the press, most fully in the Arizona Daily Star account. Even the Military.com coverage of the decision ran under the blunt headline that the system simply cost too much.

The critique and the defense, side by side

The honest critique: concurrency and a missile that could not hit

On the order of a billion dollars, over roughly a decade, bought a networked precision-fires system that the Army cancelled after its core munition flunked the culminating operational test. Four of six shots missed, two of them landing more than 14 kilometres off target, demonstrated reliability of 61 percent against an 85 percent requirement, and an infrared seeker that hit just 5 of 11 times. On paper the program was about 92 percent complete; in the field it could not reliably put a missile on a target.

That gap between paper progress and demonstrated performance is the textbook signature of concurrency: pushing a system toward low-rate production while its core technology is still immature. NLOS-LS was riding toward fielding inside the FCS and E-IBCT effort while its Precision Attack Missile was still unreliable. This is precisely the "immature and unreliable" pattern the Government Accountability Office repeatedly flagged for the FCS spin-outs. The cost is real money spent on a capability that was never delivered to soldiers.

The honest defense: the mission was legitimate, and it survived

The other verdict is equally true. The underlying goal was sound. Networked, beyond-line-of-sight precision fires, the ability to place an unattended launcher on the battlefield and strike moving armor or a fleeting target without exposing a crewed launcher, is a real and enduring warfighting need. Cancelling a missile that could not meet its reliability requirement is oversight working, not oversight failing. The system was tested honestly, it failed honestly, and it was stopped before it went into full production. That is the process doing its job.

And killing NLOS-LS did not kill the mission. The Navy had planned to use NLOS-LS in the Littoral Combat Ship's surface-to-surface missile module. After the Army cancellation, the Navy assessed alternatives and adopted the Griffin missile in its place for that role, as reported by Defense Media Network. The Army likewise continued pursuing networked precision munitions through other programs. The capability need persisted; only this particular execution of it was abandoned.

The right lesson from NLOS-LS is not "precision fires are a bad idea." It is the oldest lesson in acquisition: test before you buy. A program can be almost finished, backed by two of the country's largest defense contractors, and still be the correct thing to cancel if the culminating test shows it does not work. NLOS-LS is a case where the money was largely spent and the honest answer was still no.

Fact-check notes and sources

Concept, XM501 designation, 15-missile CLU, networked remote firing. Described in the XM501 NLOS-LS overview at Wikipedia and corroborated by the GlobalSecurity NetFires page.
NetFires LLC was a Raytheon and Lockheed Martin joint venture; Raytheon built the PAM, Lockheed built the CLU and the LAM. Company roles reported by the Arizona Daily Star; the Lockheed attribution of both the launcher and the LAM is confirmed in the Wikipedia record. Do not attribute the LAM to Raytheon.
PAM and LAM specifications; roughly 500,000 dollars per missile; LAM contract terminated earlier and separately. Specifications and unit cost from the XM501 NLOS-LS Wikipedia entry.
NLOS-LS originated in Future Combat Systems, then spun out to E-IBCT Increment 1 when FCS was restructured in 2009. Operational-test coverage under E-IBCT Increment 1 appears in the DoD's DOT&E FY2009 Annual Report. FCS was restructured in 2009; NLOS-LS was cancelled separately in 2010.
February 2010 Limited User Test at White Sands (Jan 26 to Feb 5): four of six PAMs missed, two impacted more than 14 km off target, six of seven aborts from new navigation software. Reported by Military.com / DefenseTech, tracing to April 16, 2010 testimony by OSD DOT&E official David Duma.
Demonstrated reliability 61 percent versus an 85 percent requirement; imaging-infrared seeker hit 5 of 11. From the same Military.com / DefenseTech report of the DOT&E testimony. A consistent earlier figure of about 62 percent (8 of 13 shots as of September 2009) appears in DOT&E reporting; 61 percent is the culminating value.
Earlier record: roughly 12 of 17 prior test shots succeeded; November 2008 CLU firing hit a T-72. Reported by the Arizona Daily Star. This is industry-adjacent reporting; treat "12 of 17" as reported, not audited. The DOT&E cumulative figure was 8 of 13 as of September 2009.
Cancellation timeline: Army recommended cancellation April 2010; DoD approved about May 18, 2010; Gates budget announcement January 6, 2011. Reported by Defense Systems and the XM501 NLOS-LS Wikipedia timeline. Distinct from the 2009 FCS cancellation.
Spending: about a 1.1 billion dollar 2004 development contract; "more than 1 billion dollars" reported invested over roughly a decade; program described as about 92 percent complete. Reported by the Arizona Daily Star; the "cost too much" framing from Military.com / DodBuzz. These are reported and program-office figures, not audited GAO or SAR spent-to-date lines; the 1.1 billion dollar contract is a ceiling, not a spend.
Navy adopted the Griffin missile for the Littoral Combat Ship role after cancellation. Reported by Defense Media Network.

NLOS-LS: the billion-dollar missile-in-a-box that was 92 percent finished and still cancelled