Back From the Future: Stop Saying the Military Has an Innovation Problem
Senior leaders at the Pentagon love their acronyms and buzzwords. The soup du jour is “innovation,” embodied in various entities such as the Defense Innovation Unit, Defense Innovation Board, Small Business Innovation Research, Rapid Innovation Fund (now defunct), Accelerate the Procurement and Fielding of Innovative Technologies, National Security Innovation Network — the list goes on. All are meant to overcome a perceived problem: The Department of Defense has stopped being good at creating new things and is terrible at innovating.
The Department of Defense is indeed struggling to get the right tools into our military’s hands to fight and win the nation’s wars. Until very recently, its nuclear arsenal ran on floppy disks. Its intelligence systems are so dysfunctional that warfighters prefer to use Google Maps. Experts estimate that our adversaries adopt new technology roughly five to six times faster and, in terms of purchasing power parity, spend one U.S. dollar for every 20 from the American military.
However, describing this as an “innovation challenge” has made things worse. This approach wastes the time of senior leaders with watching innovation theater. It strains relations with a Congress that is rightly skeptical of lavish spending on futuristic projects with little to show for their taxpayer dollars. But worst of all, it has obfuscated the true nature of the problem, enabling the root cause of the U.S. military’s technology woes to go unaddressed.
The Department of Defense does not actually have an innovation challenge. The real culprit is an acquisition system that has failed to keep pace in the digital age. The current process rigidly defines capability requirements in a way that does not account for the iterative nature of modern technology and alienates us from the companies building it. The Pentagon should reimagine its requirements process as a series of defined end states and let the real innovators in America’s private sector figure out how to get there.
Behind the Curtain of Innovation Theater
Once upon a time, the U.S. government truly was a wellspring of innovation. Throughout the Cold War, federal spending on research and development was a dominant force in producing new technology. In the mid-1960s, government funding accounted for over two-thirds of total U.S. expenditures on research and development. What’s more, it reaped serious rewards. The internet, modern rocketry, and powdered orange juice were all crowning achievements of well-managed public sector innovation.
However, something changed at the end of the 20th century. Software became the lynchpin of virtually all technological progress, and the federal government never caught on. Defense acquisition critics have rightly wondered how the military could field six new fighter jets in less than a decade during the Cold War but take more than 20 years to roll out the F-35. Part of the answer is that Cold War-era platforms did not depend on millions of lines of computer code.
Today, the research and development spending of software giants like Google and Amazon dwarfs that of the federal government. High salaries, coupled with an exciting, fast-paced work environment, ensure that the best engineers today will choose SpaceX over NASA. Meanwhile, the state of software development in federal labs lags multiple years behind the private sector. What passes for a commercial defense and aerospace sector is little better. The industry has steadily consolidated into five so-called “Defense Primes” that provide the overwhelming majority of military platforms and effectively operate as an extension of the U.S. government. Their single-digit numbers for independent research and development investment indicate a lack of incentive to take risks by aggressively pursuing new technologies or ideas. When revenue is practically guaranteed, there is little incentive to innovate.
More significant still is the widening delta between public- and private-sector capacity to iterate on technological progress. Powerful new tools like software containerization allow commercial companies to test, deploy, and rapidly refine their products. The U.S. military still approaches the capabilities it acquires as static goods that undergo few changes once the final contract is signed. The cumulative effect is a Department of Defense that has fallen further and further behind on the technology curve. Unfortunately, the military’s leadership has mistaken the symptom for the disease.
Band-Aids Over Bullet Wounds
The response has been something like, “We don’t know what innovation is, but we know we need it now!” A frenzy of activity has ensued, primarily involving press releases on the formation of new “innovation task forces” and innovation funds. Astute observers with longer memories have pointed out that they have seen no shortage of these programs come and go over the years without delivering results. The various band-aids that have been slapped over this bullet wound will fail to stem the bleeding for two reasons.
Firstly, the military will never have the necessary technical expertise to build these tools ourselves. The Department of Defense has currently stood up no fewer than 29 independent software factories in a bid to cultivate organic capability for software development. Whatever talent this disparate collection of efforts generates will rapidly be poached by the private sector long before it delivers any results to the warfighter. Worse still, groups like U.S. Central Command’s new Task Force 99 are attempting to build things like counter-small unmanned aerial systems themselves when a mature commercial market for these products already exists. The military hires and trains world-class warfighters, not world-class engineers. It should stay in its lane.
Secondly, even if these engineering challenges were solved, there is no clear path forward for any of these “grassroots efforts” to scale. Systems that can be put into mass production and delivered to the joint force with the logistical tail to support them are what make a difference in combat. Doing so requires the use of the defense acquisition system, which is the sole prerogative of the military service departments. Even with a litany of new “bridge funds,” these innovation theater troupes have no avenue to see their efforts scaled into programs of record.
The impartial observer might reasonably ask: “What’s the harm in all this innovation buzz?” The answer is that it threatens to keep us perpetually chasing shiny prototypes into the so-called “valley of death.” We are liable to enter the next fight with a litany of half-baked ideas that will never scale into real capabilities for the warfighter. The undersecretary for acquisition and sustainment recently lamented the amount of time the Department of Defense spends obsessing over concepts that will never be put into meaningful levels of production. This trend can be quantified by observing that the ratio of procurement to research and development has fallen from a high of 2.74:1 in the 1980s to nearly one-to-one today. Meanwhile, the root causes of the Department’s technology woes go unaddressed.
The antiquated, byzantine acquisition process of the U.S. military is what truly deprived our warfighters of the incredible benefits of the digital revolution. While some of the efforts to date have recognized this, they give the false impression of an easy fix. The Defense Innovation Unit was founded as a tech-savvy acquisition office with special authorities for contracting to bring Silicon Valley to the military. At the beginning of 2023, they reported $4.9 billion in total contracts awarded since their inception in 2016. While this is an admirable effort, the reality is that this represents just a fraction of a percent of total Department of Defense spending on procurement over the same period. The overwhelming majority of military platforms continue to churn through a process in desperate need of reform.
If You Buy It, They Will Come
At the heart of this reform is reimagining the task of requirements validation in the defense acquisition system. The Pentagon’s existing process for this, the Joint Capabilities Integration and Development System, defines exactly what warfighters will need from each military system in excruciating detail. It was designed with a hardware-first mindset to ensure interoperability across the joint force. However, the military services are starting to wake up to the fact that even in ostensibly traditional hardware platforms, like the Army’s next-generation fighting vehicle, software is the critical ingredient in success. When we say “innovation” today we are primarily talking about software, and if you need software, there is only one place to go: Silicon Valley.
The good news is that much of the tech that programs like the Optionally Manned Fighting Vehicle desperately need — computer vision, mesh networks, autonomous software stacks — already exists in mature forms in the commercial sector. The bad news is that this idea of rigid, exhaustive specifications for products, laid out by acquisition officers on the Joint Staff, is not how Silicon Valley approaches building the future. Startups become the next Google or Amazon by building something that no one has ever thought of before — the exact anthesis of the Joint Staff’s approach. New acquisition vehicles such as the Middle Tier of Acquisition and the Software Acquisitions Pathway are a step in the right direction, but their success is wholly dependent on their parent process for requirements definition.
Conclusion
If the U.S. military is going to ever realize the promise of the digital age, it must change this process to describe “what needs to be done” instead of “how to do it.” Military leaders will find this approach already embodied in the familiar concept of mission command. Leaders describe a desired end state they want to reach and mission parameters to operate within, then empower their subordinates to develop creative ways of getting there. The only difference here is that, instead of subordinate leaders, their vision will be executed by engineers in Silicon Valley.
This shift in thinking requires the right types of leaders with the vision and technical expertise to liaise between the two very different communities — a tough find in the Department of Defense today. The rare, enlightened program manager might leverage existing capabilities in the commercial sector that can be rapidly repurposed to meet the military’s needs. More frequently, a hapless officer, who may or may not have any specialized expertise in the technical aspects of the program they manage, must negotiate complex contracts, evaluate prototypes, develop a data strategy, and act as a systems integrator for hardware-software solutions. The Pentagon needs a different type of team for these projects, with the skills to appropriately spell out technical details like data and intellectual property rights in a way that makes the requirements process viable for private companies to participate.
Acquisition reform is a daunting task, but there is no easy way around it. Putting on more innovation theater for an expedient win gets us no closer to solving the real problems. Whatever the good intentions of these efforts, they distract us from the tough work of treating this disease at the source. Military leaders must embark on real reform to harness the incredible power of America’s private sector and put the best tools possible in the hands of the warfighter.
Erik Johnson is a senior engineer who has worked on numerous software and hardware programs for the U.S. military. He is a former infantry and military intelligence officer in the U.S. Army who served in Iraq and in the Office of the Secretary of Defense. He holds a B.S. from the United States Military Academy and a master’s degree from Georgetown University. The views expressed here are his own and do not necessarily reflect the position of the United States Department of Defense.
Correction: This article has been updated to correct an error in the Defense Innovation Unit’s total value of contracts awarded from 2016 to 2022 as a percentage of procurement spending. The correct number is .499%, not 0.00499% as originally written.”