The Law the Department of Defense Loves to Break

When the DoD ignores existing commercial solutions that satisfy requirements, the US military is less competitive and warfighters get worse products.

Thirty Years On: The Underwhelming Legacy of Commercial End-Item Preference

It’s been thirty years since Secretary of Defense William Perry spearheaded the most consequential defense acquisition reform in modern history: Since 1994, the Federal Acquisition Streamlining Act (FASA) has required the US government to buy commercial solutions over building custom products. Unfortunately, FASA may be the most violated law on the books. When the Department of Defense ignores existing commercial solutions that satisfy requirements, the US military is less competitive and warfighters get worse products.

I am going to examine how and why things went so wrong. The primary culprit is the government’s aversion to the optics of high corporate profit margins. This has manifested in the government’s perpetuation of “Not Invented Here” syndrome and a misguided use of cost-plus contracts, both of which effectively exclude commercial solutions.

When Perry returned to the Department of Defense almost fifteen years after serving as the Undersecretary of Defense for Research and Engineering it was with the explicit goal of unifying defense and commercial procurement. The military was using too much custom technology, and it needed to take advantage of all the innovation happening in the commercial sector. FASA was supposed to solve this waste of time and money by establishing commercial end-item preference. The law stipulates that the government must use a commercial product if one exists that satisfies the agency’s requirements. But it goes a step further. The government must also use the commercial product if it could be modified to meet the agency’s requirements or could meet the agency’s requirements if those requirements were changed. In other words, the government must exhaust all avenues in ensuring custom development efforts are a last resort.

Yet too often, custom development is the default.

This was what Palantir protested when it sued the Army back in 2016, after the Army excluded commercial companies from bidding on the re-compete for the Distributed Common Ground Systems. Two years later, Palantir won the underdog lawsuit on the basis of the existing FASA legislation. This didn’t mean it was awarded the Distributed Common Ground Systems contract: It merely won the right to compete for the contract, which it then won (full disclosure: I work at Palantir, and for that reason I’m very passionate about this topic).

The defense tech ecosystem often points to this lawsuit as a watershed moment for companies selling commercial technology to the military. In some ways, it was. But in other ways, it was just business as usual. A recent Wall Street Journal article reported that “venture-backed companies have consistently received around 1 percent of total DoD award spending annually going back through 2018.” Now, not every commercial company is venture-backed. However, the venture capital model is a proven engine of innovation and a decent proxy for the Department of Defense’s use of commercial technology.

I’ll keep picking on the Army for consistency, although it is far from the only offender. In March of this year, the Army released its directive for Enabling Modern Software Development and Acquisition Practices. One would hope that a key part of a modern software acquisition strategy would involve embracing the Software-as-a-Service (SaaS) model, which is how businesses have purchased commercial software for the past twenty years. Instead, the directive mandates the use of cost-plus contracts for software development to the “maximum extent practical.”

This is a problem because the SaaS business model exclusively uses firm-fixed price contracts, which enable companies to get operating leverage on their invested capital. With cost-plus contracts, no matter how large a company scales, it will never be able to generate high margins that outrun the company’s fixed costs. That’s why the cost-plus contract is kryptonite to venture-backed companies—for a complete explanation of this dynamic, read Jake Chapman on venture math in defense. Mandating cost-plus is the equivalent of flying a flag that the Army will not be buying commercial software. That’s a violation of FASA, an issue has quite literally already been litigated with the Army.

Fortunately, the Senate Armed Services Committee (SASC) recognized as much in the FY2025 National Defense Authorization Act. The committee noted that the “Directive regarding appropriate contract terms for software development efforts appears to deviate from the current law, section 3453 of title 10.” As a result, the committee is requiring the Army to submit a report by January 15, 2025, justifying its decision to migrate towards cost-plus contracts and explaining how the Army intends to get full industry participation from non-traditional contractors (i.e., commercial firms) who use firm-fixed price contracts.

That wasn’t the only thing the SASC had to say about the Army’s approach to software. An article in Breaking Defense recently exposed the challenges plaguing the Army’s Robotic Combat Vehicle (RCV). The core autonomy software for the RCV is a custom, in-house development effort by the Army called the Robotic Technology Kernel. Twelve years have passed, and the Robotic Technology Kernel is not delivering. A human must always be present in the vehicle, and basic road signs are interpreted as obstacles. Meanwhile, there are many commercial autonomy companies who offer solutions the Army could simply buy. Per the SASC, “The committee is concerned that the US Army continues to fund the Robotic Technology Kernel … The committee believes the US Army should reexamine its funding decisions and consider further engagement with providers in the ground autonomy industrial base.”

The preference for custom, in-house solutions over commercial technology is a systemic issue not just for the Department of Defense but also the Intelligence Community. In 2023, Percipient AI accused the National Geospatial-Intelligence Agency for violating—you guessed it—FASA. The agency awarded systems integrator CACI a prime contract for obtaining and storing visual intelligence data. Percipient AI hoped CACI would issue it a subcontract for its commercial platform for computer vision. Instead, CACI decided to develop a computer vision platform in-house. In June 2024, the Federal Circuit reversed the Court of Federal Claims’ dismissal of Percipient AI’s complaint. The new ruling provides potential subcontractors the ability to challenge an agency’s violation of procurement law, which hopefully makes it harder for a prime contractor to avoid bringing on a commercial solution to fulfill a requirement. Companies should not have to sue the government for FASA to be enforced. That’s the nuclear option, and it comes with a huge opportunity cost.

A more palatable option available to disgruntled companies is to write a report. The National Reconnaissance Office (NRO) Industry Advisory Working Group, a consortium of commercial software companies, published a report earlier this year documenting the systemic language in NRO solicitations that discriminated against commercial software. For example, companies were strongly encouraged to use only “free and open-source software” in their proposed solutions. This precluded commercial companies with a paid, SaaS product from bidding. What this means in practice is that the NRO is gifting systems integrators a monopoly on software procurement by paying them to build a bespoke solution stitching together “free” software. It’s expensive custom development thinly disguised as the economical approach.

It’s not just software. Unit X, a new book about the origin story of the Defense Innovation Unit (DIU), documents the enduring pathology of “Not Invented Here.” In 2017, DIU wanted to award a $15 million contract to Capella, a commercial company building a constellation of low-cost, Synthetic Aperture Radar satellites. There was immediate opposition from the intelligence agencies who viewed Capella as directly competitive to the government’s ongoing but foundering custom, multi-billion-dollar Synthetic Aperture Radar satellite. The DIU eventually succeeded in getting Capella a contract, and the rest is history. As for the intelligence community’s custom satellite? Per Unit X, “By 2023, the firm that had won the contract still hadn’t delivered a working satellite, and the Pentagon was soliciting new bids from other prime contractors. Because it’s a ‘black’ program, we’ll never know exactly how much money has been wasted, or what went wrong, or who screwed up—that’s all classified.”

The Government’s Refusal to Price Value

What’s at the root of the government’s widespread violation of commercial end-item preference? I would argue it’s the government’s inability (or refusal) to price value. The everyday American buys products by considering price and quality. They don’t care about the profit margins of the company making their car or vacuum. The government views procurement decisions in the inverse. They examine everything from a cost basis (how much did it cost Company X to make Product Y), scrutinize the profit margins of the company, and then decide if they are getting a good deal. This is emotional, not rational. It relates to FASA violations because it favors systems integrators building low-margin, custom projects billed as developer hours over commercial companies with a high-margin, existing product.

As the recent Report of the Commission on the National Defense Strategy made abundantly clear, the United States is in a crisis. Emotional decisions are a peacetime luxury, and an inability to price value leads to dangerous logical fallacies. In 2023, the government announced the Better Contracting Initiative (BCI), an effort to save the government billions annually in its purchases from contractors. A key justification of the BCI was that “With many of the largest Federal contractors operating at historically high margins, resulting in costs to taxpayers that are simply too high, agencies must act as an organized enterprise now more than ever.” The BCI then cites a report that finds contractor profit margins have grown to 15 to 17 percent, above historical norms of 6 to 10 percent.

This is very concerning, but not for the reason the BCI purports. It is incorrect to blindly equate higher margins with higher taxpayer costs. To spell it out: Companies could be selling products at high margins but lower overall cost. In fact, this is what you would expect from innovative commercial companies amortizing self-funded research and development across a large number of customers. The best-known example is, of course, SpaceX. The company lowered launch costs and humiliated the United Launch Alliance’s technology and business model in the process. Who cares if SpaceX’s margins are higher? Perhaps the US government, but certainly not China or Russia. Those countries just see that the United States is home to the firm that launched 87 percent of all tonnage to orbit in Q1 2024.

The US government chooses to obsess about profit margins—rather than value delivered—at its own peril. This mindset is the best way to ensure repeated FASA violations, but more importantly, it will doom the United States to failure in the next war it fights.

---

Madeline Zimmerman works on defense at Palantir and supports the company’s First Breakfast initiative. She also writes Kinetic Reviews, a Substack about the history and future of defense technology. The thoughts expressed here are her own.