Dr. Alex Miller is the Chief Technology Officer of the United States Army.
Greg Little is a senior counselor at Palantir.
Cost-plus contracts were first used in 1907, formally authorized in the 1916 National Defense Authorization Act, and widely adopted during the world wars to encourage production in radically uncertain fiscal and geopolitical environments. The bargain was simple. Companies would be good partners to the government and produce war materiel as efficiently as possible. In return for taking this risk, the government would cover companies’ costs plus a fixed percentage of profit (thus the term, “cost plus”).
The wars ended, but cost plus remained. Today, the U.S. government continues to rely on cost-plus contracts as a default way to get things done. At first glance, it still looks like a sensible arrangement: reimburse vendors for allowable costs, add a “reasonable” profit margin, and reduce risk for complex or uncertain work. But the logic only holds if you assume the world is static; if the first derivative is zero, nothing gets better. If you believe the world is dynamic, then you need an arrangement that encourages improvement. You want investment, efficiency, and innovation. To get those things, you need to hold out the prospect of substantial future profit.
Cost plus doesn’t encourage investment, efficiency, and innovation. Quite the opposite. If you believe technology makes things better, faster, and cheaper, then cost plus locks you into a model that is inherently short technology. The more time and resources a vendor uses, the more they get paid. Efficiency is punished, inefficiency is rewarded, and accountability—well, that gets lost in the paperwork.
In fiscal year 2024, the federal government spent over $230.2 billion on cost-reimbursement contracts—many of which rewarded activity, not results. Under cost plus, the act of executing the contract becomes the product. Entire programs are staffed with contractors billing “reasonable” or even cheap hourly rates with unreasonably little to show for it: reports no one reads, systems no one uses, or processes no one questions. A vendor can fail to deliver anything meaningful and still get paid in full (at a profit!), so long as the timesheets are accurate. That’s not just inefficient—it’s fundamentally misaligned with the public interest.
The government has even acknowledged the problem, creating cost-plus-incentive-fee (CPIF) contracts that adjust a vendor’s fees based on how well they meet performance targets, often tied to cost. But these tweaks don’t change the underlying dynamic because they still preserve the core structure of cost plus. The government remains on the hook for reimbursing every allowable cost, no matter how high, up to the ceiling. Vendors are never forced to internalize the downside of inefficiency, only to earn slightly more or less on the margin depending on whether they “beat” a target. At best, it’s a bonus system layered on top of a fundamentally misaligned model.
The reality is that CPIF contracts measure success against pre-negotiated cost baselines, which quickly become outdated in fast-moving technology environments. If a project should cost half as much because of a new tool or process, the incentive doesn’t force a vendor to exploit that advance—it just nudges them to come in a little below their last estimate. That makes sense only in a world where innovation is stagnant. But in practice, it cements the same perverse incentives: run long, bill high, and count on the government to pick up the tab.
Now imagine flipping the model on its head. What if, instead of getting paid more when things cost more, a vendor got paid more when the government spent less? What if a contractor’s compensation came not from racking up hours or expanding staff, but from the value they created: measurable, targeted, provable outcomes? That’s the premise behind a cost-minus contract.
In a cost-minus model, the government shares a portion of the value saved, recovered, or created with the contractor that delivers it. Suddenly, contracting would be about rewards for results instead of billing for effort. This is different from the existing Fixed-Price Incentive model, which sets a price ceiling (but no floor) and adjusts profits based on lower actual costs. While a good concept, the government still accepts risk up to the ceiling. Cost minus creates an incentive to actively lower actual expenditures—reducing the government’s total outlay—while giving the vendor a portion of the savings as profit.
Picture this: a fraud analytics firm partners with a state unemployment agency during the pandemic. Using anomaly detection and pattern analysis, they prevent $200 million in bogus claims. Under a cost-plus contract, they might earn a flat $1-2 million for their time and have no incentive to prevent bogus claims. Under a cost-minus model, they could earn 5-10% of the bogus claims they prevented—$10 to $20 million. The state would still come out ahead by $180 million.
We know this model works in the real world. Contingency-fee law firms only get paid if they win, whistleblowers under the False Claims Act are rewarded a share of what they recover, and cybersecurity “bug bounty” programs pay only when someone actually finds a flaw. No bug? No bounty. No savings? No invoice.
Cost minus works. It scales. It builds a cadre of vendors with track records of measurable performance. Most importantly, it fully aligns incentives for taxpayers, the government, and the warfighter.
And contrary to what you might expect, cost minus also aligns with shareholder incentives. Shareholders don’t just seek raw, guaranteed profit. They seek valuation. Profit and revenue are messy proxies for share price. What really matters is whether a company has the confidence and competitiveness to grow in a system that rewards value. Good companies—those with the will and conviction to compete—will thrive in a cost-minus world. The only companies that will hate it are the ones that don’t think they can hack it.
Where else could cost minus work? Just about anywhere you can track a “before” and “after.” Start with improper payments. In 2024, the federal government reported over $162 billion in improper payments—whether through fraud, error, or lack of documentation. That’s not a rounding error. It’s almost 10% of discretionary spending. To put that in perspective, it’s more than the combined budgets of the Departments of Justice, Homeland Security, and State. Yet year after year, we pay vendors to analyze, audit, and report on the problem, not to stop it. A cost-minus contract could flip the model: give a firm a cut of the improper payments it prevents.
Then there’s tax enforcement. Treasury and the IRS have long known that tax underreporting and evasion costs the federal government over $1 trillion per year. Software that helps close that gap should be paid based on what it recovers, not the number of analysts it requires. Or take cloud optimization, where agency bills balloon due to idle compute, orphaned storage, and bad provisioning. If a vendor helps reduce that bill by $100 million, give them 10%. Taxpayers still keep $90 million they were losing anyway.
For the Department of Defense, the most important use cases for cost minus are in programs that are both highly complex and high stakes—manufacturing and production efforts where uncertainty is great but desired outcomes are clear. Munitions, energetics, and vehicle manufacturing are prime examples. These are exactly the domains where sharpened incentives could accelerate delivery and lower costs, yet design inertia has locked the system into contracting models from a different era. Because the mission is “no fail,” the reflex has been to preserve current contracts at all costs, due to the fear that any disruption risks failing to deliver the quantities we need. In reality, this has saddled the department with both technical and fiscal debt, making it a bad customer in an effective monopsony that is incentivized to perform according to last century’s metrics.
Critics will say the cost-minus model invites gaming. What if vendors manipulate numbers? What if they take credit for savings they didn’t generate? These are valid concerns, but they have valid solutions: independent baselining, automated audit trails, third-party validation, and clawbacks. The answer to potential abuse isn’t to stick with broken incentives—it’s to design better ones. Past performance can also be a control mechanism: vendors with a history of delivering under firm-fixed-price contracts pose less risk, and vice versa.
The truth is, we know why cost minus hasn’t happened at scale. It’s not about legality or policy—there’s no law preventing a cost-minus contract. Agencies could launch pilots tomorrow under Commercial Solutions Openings, Other Transaction Authorities, or simple pay-for-success agreements. The problem is cultural. The acquisition system was built to control risk, not reward performance. It’s “safe” to spend $100 million on a failed software project under a cost-plus vehicle, but “risky” to give a startup 10% of the $200 million they save the government. That’s not oversight. That’s fear.
Cost-minus contracts also require a fundamental mindset shift in how the government manages requirements. Instead of locking vendors into rigid, pre-determined deliverables, the government must give industry flexibility to find more efficient and effective paths to the same outcome. That means greater risk acceptance on the government side, but also greater accountability on the vendor side—measured in results, not compliance checkboxes. The focus should shift from oversight of inputs to the velocity of value delivery: how quickly solutions can be fielded, tested, improved, and scaled.
This framing isn’t new. It echoes DoD’s own reform language around “speed to delivery,” “buy before build,” and “commercial first”. If the department is serious about leveraging commercial innovation, it must move beyond requirements processes that assume static solutions and instead embrace iterative, outcome-driven contracting. This is especially critical for diversifying and commercializing supply chains, deliberately investing in resilience, and fully harnessing dual-use technology. In a world where private-sector innovation cycles run in weeks, not years, the government cannot afford acquisition models that lock it into the past.
It’s also about power. If your business model depends on inefficiency, you’re not going to like cost-minus contracts. But that’s exactly the point. The firms that say they want to help the government modernize should be the first to volunteer for a cost-minus pilot. If they don’t, you’ll know why.
This is a call to every program executive, contracting officer, inspector general, and Hill staffer: if you believe in performance, pilot a cost-minus contract in your domain. If it works, scale it. If it doesn’t, you’ll still have learned more than from another white paper, oversight report, or RFP cycle. Build on the momentum already underway with the termination of JCIDS, the FAR rewrite, and the Army’s Open Solicitation—steps designed to get the right firms aligned against the right problems.
Cost-minus contracts would move the department toward the budget flexibility it needs to leverage new technology at speed. They would also put pressure on the defense industrial base to stop mirroring the bureaucracy of the DoD itself, whose structures and processes too often protect incumbency rather than deliver capability. For decades, the result has been predictable: systems that are late, over budget, and misaligned with operational timelines.
Don’t let the perfect be the enemy of the possible. Don’t pay for effort, compliance, or the hours it took. Pay for what it’s worth.
 | A guest post by
|