Some of you might be surprised to learn that captains of the U.S. Navy’s largest, most complex, and most expensive combat ships — aircraft carriers — are typically not surface warfare officers with extensive experience commanding other navy ships. Instead, they are typically career naval aviators. This is not because a carrier’s only important task is aviation. From navigation to air defense, to the running of a multi-billion-dollar nuclear reactor, the successful functioning of a carrier requires the careful orchestration of a diverse set of expertise. Yet career naval aviators are chosen to lead carriers because launching and recovering aircraft is the most important thing a carrier does. It is the carrier’s raison d’etre.
The analogy extends to U.S. defense modernization, the sprawling bureaucratic apparatus and set of practices that acquire new military programs. Since the end of World War II, contractors for major American weapons platforms have been aerospace and shipbuilding manufacturers of some kind. And, with good reason. From World War II through the end of the Cold War and some years beyond, the most important capability in this enterprise was the bending of metal, followed by the integration of components built by other hardware and (sometimes) software subcontractors. Despite the rapid rise of emerging technology and advanced manufacturing in the commercial sector, the U.S. defense industry has not diversified. Instead, it has consolidated from roughly 75 contractors to five primes, all traditional hardware contractors.
However, because the systems needed to combat higher-tech adversaries will increasingly rely on software-driven digital capabilities — such as data fusion and AI — one should ask: Is the traditional, hardware-centric model of defense acquisition and systems integration always the best model for fielding the most capable military technology that America has to offer?
Given the massive changes wrought by information technology in recent years — and importantly, their corresponding impact on modern warfare — it is worth re-thinking the conventional wisdom that only traditional prime contractors can lead the development and production of all of America’s most consequential defense capabilities. The Defense Department should widen the aperture of what qualifies a company to manage defense acquisition programs and consider including a different kind of leader in the defense industrial enterprise: A prime contractor whose core expertise originates in software, rather than hardware. In other words, a “software prime.”
As a senior executive and a senior advisor at Palantir Technologies, a software company that does a great deal of business with the Defense Department, we have an interest in contracting outcomes. But we also speak to the issue from a place of deep experience — one from the software technology sector, the other as a career naval analyst and defense strategist.
Receptivity to this new approach in the Department of Defense, Congress, and the defense industrial base more broadly is critical because success on a near-future battlefield will largely depend on America’s ability to use its software advantage to connect military assets of all kinds. In the future the United States will need to integrate legacy and next-generation, simple and complex, massive and nano systems seamlessly through land, air, sea, space, and cyberspace. In the war in Ukraine, for example, despite resembling the last century’s defining conflicts in many respects — muddy slogs advanced or repelled by heavy field artillery — the war has shown just how impactful advanced software, AI, and other emerging technologies can be in leveling the playing field between otherwise asymmetrical militaries. While Ukrainian forces have been responsible for innovation on the battlefield, in many instances, the software helping Ukraine make a difference was developed and launched by American technology firms. Many of these technology firms are not the well-known, large defense primes, but, rather small and mid-sized tech companies who have deep software expertise and nimble operating models.
This same type of software-powered warfare will be needed, albeit on a much vaster scale, in the Indo-Pacific, where the United States and China could confront each other in a Cold War-like competition that could eventually turn hot. While American ships, aircraft, submarines, satellites, and weapons are (and will likely remain) among the world’s best as measured in speed, stealth, precision, range, and other traditional metrics, the challenge is that this hardware advantage is likely less determinative of a military advantage than it was a generation ago. Increasingly cheap, capable systems that are enabled by smart software have the ability to disrupt and even overwhelm these platforms. According to the Defense Innovation Board, “Software is leveling the playing field with our rivals, eroding the advantages we have spent many decades accruing.” In this world of fast-changing capabilities, and overwhelming amounts of information and data, the competitive advantage will go to the country whose decision-makers and warfighters are best able to make rapid and well-informed decisions at scale.
The Software Challenge and Advantage in Defense
Despite being one of the U.S. military’s greatest advantages, the challenge of software performance and integration has been with us in various forms for more than a decade. For example, some of the most time-consuming and expensive delays for major weapons systems in recent years — from the aerial refueling tanker to the F-35 Joint Strike Fighter — have been related to hardware-first processes that impede the integration of their most critical component: software. The F-35 was referred to as “a computer that happens to fly” by an Air Force chief of staff. Increasingly, major hardware programs are dependent on software if they are to realize a competitive advantage, including within consumer markets. By one estimate, even the state-of-the-art F-35 uses fewer lines of code than a late-model Mercedes automobile.
Yet in the commercial sector, firms with software expertise are leading in production of high-performing hardware. Consider that the most successful electric vehicle maker to date, Tesla, is essentially a software company that is focused on creating sustainable transportation. The same holds true for large numbers of small, even micro, satellites whose collective power comes from connectivity and software that can be frequently updated and improved.
To a certain extent the challenge stems from how the defense enterprise develops and procures software on its own accord, and not necessarily as part of a major weapons platform. As early as 1987, the Defense Science Board noted the principal problem with military software development was not technical, but managerial. The same report argued the Defense Department could no longer dictate new software technologies and standards given that the commercial market was then some 10 times larger than the military, and that gap would grow exponentially in succeeding years.
Well into the 21st century, most of the same institutional barriers remain. “Software is different than hardware (and not all software is the same),” the Defense Innovation Board felt compelled to restate. Furthermore, “software is never ‘done’ and must be managed as an enduring capability that is treated differently than hardware.” It is therefore unreasonable, if not unfair, to expect entities organized, trained, and incentivized for hardware to essentially invert themselves for major acquisition programs in which software is the defining capability.
Whether in the military or commercial sectors, software can no longer be considered an enabling or “add-on” capability to be cobbled into an existing platform. Nor should it be common for software capabilities to be purchased separately from and, in some cases, long after the hardware has been designed. In short, to fully capitalize — and build on — America’s software advantage we need to put software at the beginning and center of new weapons design and development. In this model, companies whose core expertise is software could be the prime contractor rather than serve as the late-stage add-on.
What Is a Software Prime?
A “software prime” is, at its core, a software company that serves at the lead contractor for a government project. As the prime, the software company would be responsible for building and leading a team of sub-contractors to meet and deliver the project’s requirements. Importantly, that would include the hardware, whether the hardware is a truck, computer screens, antennas, or batteries.
A software company would serve as the prime contractor when it is the software, rather than the hardware, that will determine a project’s success. To be a candidate as the prime contractor, the software company would need to have a software mindset, deep expertise, and a proven track record of having built, tested, and successfully fielded a software architecture that is aligned with the program’s requirements. This capability could have been deployed in the commercial or government sectors but, significantly, there should be tangible past performance evidence. This expertise takes years of investment and experience to build, with the result being software that would be operable and available to the government from the beginning of the program. That would allow the government to rapidly produce the needed capability, rather than take years to develop entirely new software as is so commonly attempted today, generally without success.
The U.S. government is already making a similar shift towards commercial cloud service providers like Google, Amazon Web Services, and Azure. In the past, the government would attempt to build data centers themselves. With the shift to commercial cloud service providers, instead of waiting months to years to procure, rack, power, and secure hardware in a data center, cloud customers can quickly grab a compute engine that fits their needs right away. By leveraging expertise and experience, a software prime could similarly field enterprise-scale, open, and fully modular solutions for defense systems that are commercially proven. As a result, the time to initial delivery is shortened and the platform can be more readily adapted and changed as determined by the mission.
This is not a new idea in the Defense Department. It would build on the “plug-and-play” modular approach to modernization that has been discussed, and even tested, for decades. For example, as early as the mid-1990s, the Navy replaced a “closed” information technology system with a modular, open architecture that made it possible to routinely upgrade — below cost and ahead of schedule — the sonar systems of its attack submarines. But what became known as the modular open systems approach envisioned to include both hardware and software, was not widely adopted across the Navy or other services. Today, modularity is once again recognized as important to making new systems readily upgradable and several new combat platforms are now being developed using modular approaches. Ultimately, it is the software that will make modularity successful, suggesting that software expertise should be in the lead or, at least, a strong partner in the design of new systems.
As we consider the software prime concept, it is important to distinguish this from a traditional systems integrator role that is commonly used by the U.S. government today. Today, systems integrators are asked to cobble together disparate IT and software solutions, both purpose-built and commercial off-the-shelf, into a holistic system, despite the fact that none of these components were designed to work together. The results are generally disappointing, often leading to an unnecessarily complex (and sometimes insecure) architecture and programmatic cost overruns.
Under a model where the importance of software is elevated, software and digital considerations would be the guiding principles of platform development, system acquisition, and contractor team building from the very beginning. The value of a software prime would be evident in the improved interoperability between the hardware and software comprising a platform, like a ground vehicle, or a system, like a command center. Not only would the fusion of software and hardware produce new capabilities more effectively, it would lay the groundwork for those capabilities to be readily upgradable as both software and hardware advances become available, just as was envisioned for the modular open systems approach.
Examples of this future model are few, but some, like the U.S. Army’s Tactical Intelligence Targeting Access Node program, are already here. Although this is a piece of hardware (a truck), its key capabilities come from its software. Last year, the Army downselected to two prime contract bidders, one of which was our company, Palantir.
To be sure, software companies are unlikely to lead the Department of Defense’s largest programs— the kinds with unit costs approaching or in the billions of dollars. Examples of these include sophisticated fighter jets, long-range stealth bombers, attack or ballistic missile submarines, or large surface combatants. In these cases, the hardware requires specific expertise that a software company is unlikely to have. But even for these mega-programs, the software will be critically important to their success. Because of this, software expertise needs to play a more central role than is currently recognized or nurtured by the defense acquisition system. As the importance of software continues to grow, we can see a future where software for these mega-programs is contracted separately from the hardware. This approach would leverage the expertise of a software prime via a contract that is tailored to take advantage of their strengths. The hardware primes for these mega-programs would still be responsible for the majority of the development phase activities, including compliance and testing, leaving the software prime free to stay fast and iterative.
Setting the Stage for Tomorrow’s Software Primes
While the inclusion of a software company as a finalist for a prime contract certainly moves the defense community in the right direction, more progress is needed. And importantly, it is up to both government and commercial actors to help make the change.
First, the recently appointed commission to reform the Planning, Programming, Budget and Execution system — originally conceived in Robert McNamara’s Pentagon of the 1960s — will provide an opportunity to recommend changes that stand a good chance of becoming enshrined in statute. And while it is important to note that the Defense Department already has some of the tools available to change how the Defense Department acquires software and works with software partners (e.g., Other Transaction Authorities, which do not require adherence to defense acquisition regulations), it is now time to get beyond the piecemeal and experimental. The real challenge that the Defense Department faces is cultural, and breaking down such biases is usually the work of a generation. Except in this era of warfare, the United States does not have 20 or even 10 years, roughly the time it took for the Defense Department to move from paper to digitized systems, and then again from the local network to the cloud.
Second, it is not only culture that needs to improve — it is also the culture of collaboration within the commercial sector. Incentives should be generated and, where necessary, top-down pressure from Department of Defense senior leadership should be applied to promote greater intra-industry cooperation (primarily between hardware and software firms). Furthermore, the commercial side needs to show additional effort to normalize everyday cooperation between hardware and software firms, and improved private-to-private partnerships will also set the stage for what good collaboration looks like when working together on government contracts.
Overall, fundamental change will come when the defense acquisition workforce embraces the fact that software is a U.S. advantage and needs to play a more central role in the defense acquisition system. For example, as the military moves towards fielding larger numbers of smaller unmanned or autonomous platforms, there is a recognizable swath of acquisition programs where the idea of a software prime already makes sense. In these areas, Defense Department leadership can begin to view software companies as fully capable of managing and integrating significant combat systems, including those enclosed in an air, sea, ground, or space vehicle.
One does not need to be a futurist or technophile to see why the U.S. defense industrial base needs this shift. If America is going to successfully compete against near-peer powers in the digital age, it will need to lean into its unique software advantage. Software primes can be an important vehicle for the United States to leverage its competitive advantage in the realm of ones, zeros, and algorithms for years to come.
Christine H. Fox served in the Obama administration as acting deputy secretary of defense, and as director of Cost Assessment and Program Evaluation. Most recently, she served as assistant director at Johns Hopkins Applied Physics Lab and is currently serving there as a senior fellow. She is also a member of the Palantir Federal Advisory Board.
Akash Jain currently serves as president and chief technology officer of Palantir USG. He joined Palantir in 2005 as a software engineer, eventually leading the deployment and iteration of Palantir’s platform across the U.S. defense and intelligence communities, as well as alongside critical international partners.
Image: U.S. Navy photo by Rick Naystatt
Comments are closed.