It’s Grenada, 1983, and American troops are fighting the enemy. The ground troops are getting hit all over; it’s time to call in the reinforcements. So, ground troops pick up their radios and try to contact air support, but as they do so, they realize their radios can’t connect to air support; their frequencies are not the same. Desperate and losing men, they decide to use the only method they could think of: a payphone to call base. The call gets redirected to Fort Bragg, where intermediaries picked up different radios to communicate with the Air Force and the Navy. After a long game of telephone, literally, reinforcements finally arrived, but not before 19 casualties and 89 Americans were wounded. Years later, in 1991, a similar event happened during the Persian Gulf War, where similar communication failures may have contributed to friendly fire.
These failures cannot be dismissed as the work of bad engineers, lazy program managers, or careless leaders. The people involved did what their organizations asked of them, often exceptionally well. That is the harder lesson: the system did not fail because people ignored its rules; it failed because they followed them. When capable people can follow every rule and still produce failure, it’s the system itself that must change.
That is how the Joint Tactical Radio System, or JTRS, was born. Grenada and the Gulf War had given the Joint Staff a clear mandate: enable the services to communicate with one another. At the time, the military operated dozens of incompatible radio systems, almost none of which could connect across services. JTRS promised to unify them, with senior representatives from the Army, Navy, Air Force, and Marine Corps shaping its requirements. Instead, the flagship effort consumed more than 15 years and roughly $6 billion before it was canceled. How did a program designed to solve a systemic failure become one itself?
Go back to 1997, when JTRS was first conceived. Engineers consulted at the outset raised the same warning: the software-defined radio technology needed to run dozens of waveforms in a single box was not yet mature, and packing every service’s requirements into one system would push against hard limits of size, weight, and power. The requirements kept coming anyway. Each service representative, faithful to his or her own branch, added needs that made sense individually and were unworkable collectively. Partway through, LTG Joseph Yakovac, one of the senior Army leaders involved, later reflected that JTRS as structured had become, in his view, an unachievable goal. The program continued for roughly eight more years.
Consider the Ground Mobile Radio, the vehicle-mounted variant that was eventually canceled. By the end, it weighed in excess of 200 pounds and demanded more space and power than the vehicles it was meant to serve could give. The engineers who built it were not failing at their jobs. They were succeeding at an impossible one: executing the accumulated, unreconciled demands of multiple services, multiple contractors, and layers of acquisition offices, none of whom shared a common language for making tradeoffs.
In the military that rank is king: the system and the way it operates, is completely determined by this hierarchy. It makes a lot of sense in high-stakes situations: when the problem is life or death, you can’t wait for a democratized vote to culminate in orders to make a decisions. However, once we are out of the battlefield and into matters of business, this type of deferral can become obstructive. One of the main problems with JTRS was that most people knew the program was not working, but because there was no one senior leader (only a coalition), they had no one to defer to, and therefore, no one to finally say “enough is enough.”
The broader defense ecosystem wrestles with the same structural problem. Picture the four main stakeholders in any major program: the military, focused on the mission; industry, answering to shareholders and margins; the government acquisition system, bound by statute and oversight; and the contractors on the ground, who want to do excellent work but are constrained by direction and funding. Each group functions well internally. Put together, they think, prioritize, and measure success differently, and there is no designated role whose formal job it is to bridge them.
The scale of the cost is documented, and not by critics on the outside. The Government Accountability Office tracks Major Defense Acquisition Programs, the largest and most expensive systems the military buys. In one annual assessment, GAO found that combined cost estimates across just 30 of those programs had grown by $49.3 billion. Not total cost. Cost growth: money beyond what was planned, driven in significant part by misunderstood requirements, poorly specified systems, and builders and users who were not communicating. GAO also found the average time to deliver even an initial capability had stretched to nearly twelve years.
Inside individual programs, the damage compounds. Research published in IEEE Computer found that software projects spend roughly 40 to 50 percent of their total effort on avoidable rework: time spent redoing work because requirements were not understood and handoffs broke down. A 2002 study by the National Institute of Standards and Technology put the national price of poor software quality at $59.5 billion annually, with $22.2 billion of that potentially recoverable through better practices.
JTRS was one program. Dozens of major acquisition programs run simultaneously, and while no two are alike, many carry some version of the same gap and pay some version of the same rework tax.
To understand why this persists, look at how the government side is actually incentivized, and here it is worth being fair. Documentation, process adherence, and risk mitigation are not bureaucratic stubbornness. They are rational, even honorable, responses to a system that punishes failure severely and rewards compliance. Program managers who master that system are doing exactly what the nation asks of them. The gap is that communication quality and cross-team understanding sit outside that incentive structure entirely. There is no contract line item for “can the operator understand this.” There is no acquisition milestone for “does this translate across teams.”
Commercial industry does not face this problem in the same way. In the private sector, clarity equals speed, and poor communication shows up fast in lost sales, failed launches, and customer churn. The feedback loop is short. In defense acquisition, compliance is what keeps a program alive, which means a program can be thoroughly documented, fully reviewed, and still fall short in the field for years, because the process metrics it must hit do not measure the thing that matters most.
JTRS cleared acquisition reviews for fifteen years. It never cleared the field.
That is the pattern underneath every number in this piece. In Grenada and in JTRS alike, people did their jobs correctly and the system still failed, because nobody’s job was to make sure the people building it and the people depending on it were working from the same understanding.
To be clear, the Department has not stood still. The Adaptive Acquisition Framework, middle-tier acquisition, the software acquisition pathway, and the push for soldier touchpoints are all real reforms, and many programs are better for them. But those reforms address speed, tailoring, and iteration. None created a formal translation function: a role whose explicit purpose is reconciling what engineers build, what operators need, and what contractors and government offices can actually agree to. That function still exists only informally, when someone on a program happens to be good at it, and it disappears when that person leaves. It has no line item, no milestone, no title.
None of this is inevitable, and it is not unprecedented. Aviation faced a structurally similar problem decades ago. Investigators kept finding that the deadliest crashes were not caused by mechanical failure but by a copilot who noticed something wrong and could not get a captain to listen in time. The industry’s fix was not better hardware. It was Crew Resource Management: formal training and protocols that made cross-checking and communication a required, measured part of the job, not a trait some crews happened to have.
Defense acquisition has not yet had its equivalent reckoning. But that is a fixable problem, and the people best positioned to fix it are already inside the system. It means treating translation the way aviation came to treat communication: as a formal discipline with its own standards and its own seat in the room. The programs that build this in from the start will not just move faster or spend less. Decades after Grenada, they will finally be the ones that clear the field, not just the review.
Sara Muiños
Defense and Technology Strategist
Sara Muiños has worked across the defense and technology sector for four years and holds master’s degrees in Instructional Design and Technology and Digital Media. Her experience spans strategic communications, digital solutions, marketing, instructional design, and industry events, working alongside technical teams, contractors, business leaders, and government stakeholders.
Her perspective focuses on bridging gaps between disciplines and improving how complex information is communicated, understood, and ultimately put into practice.
Outside of work, she enjoys kitesurfing, sewing, and taking on new creative projects.