The Margin Trap and the Orchestrator Imperative

In my last piece I argued that the defense industrial base’s inability to generate elastic capacity is fundamentally an economic and incentive problem, not a manufacturing one. The monopsony structure of defense procurement has always rewarded stable backlog, predictable cash flows, and long program lifecycles. It doesn’t reward idle surge capacity, cross-program infrastructure, or the kind of capital optionality that agility requires.

The rise of neo-primes (e.g. Anduril, SpaceX, Palantir, and Shield AI) combined with the Department of War’s accelerating use of firm fixed price (FFP) contract vehicles, is creating a competitive wedge that legacy primes can’t ignore and can’t paper over with innovation labs and selective partnerships. The wedge is getting wider, and the financial architecture that made primes dominant for the last thirty years is exactly what is making adaptation so difficult now.

This piece is about the specific nature of that trap, what it would actually take to escape it, and why most primes are currently making the wrong bet (some of them on purpose).

There is a real issue at the heart of the large defense prime.

The financial structure that makes them attractive to institutional capital markets, predictable backlog, long program lives, stable margins, government-backed revenue, is exactly what prevents them from competing credibly in the fastest-growing segments of the defense market. Those segments increasingly reward speed, software integration, automation-driven cost reduction, and price certainty. That isn’t the world legacy primes were built for. And the longer they optimize for the world they know the harder the transition becomes.

To understand why, it helps to look at how public defense primes are actually valued. Their equity stories depend directly on backlog visibility. When Lockheed Martin, RTX, or Northrop Grumman report earnings, the metric that anchors investor confidence isn’t margin expansion or R&D optionality, it’s their backlog depth and funded bookings. Which is completely rational from a capital markets perspective. Long-dated government contracts with cost-plus structures provide the predictability that justifies the multiple.

But that same structure creates a set of deeply embedded behaviors that make transformation expensive and slow.

Capital allocation flows toward backlog, not optionality. When a prime’s valuation is anchored to funded programs, deploying capital toward speculative infrastructure, automation cells designed for cross-program reuse, software platforms not tied to a specific contract, workforce training designed for future reallocation, is hard to justify to a board or an investor. The return isn’t visible in the near term. The risk is nebulous. The payoff is probabilistic. These aren’t the characteristics that public capital markets reward and so the investment doesn’t happen at the scale the market actually needs.

Fixed-price exposure is also dangerous for organizations built around cost-plus contracting. Boeing’s defense segment has made this painfully clear over the last several years. When a firm built around cost-plus takes on fixed-price risk, it isn’t just accepting price discipline, it is accepting a fundamentally different risk model with a cost structure that was never designed to support it. The result isn’t just margin compression, it’s program write-downs, reputational damage, and balance sheet pressure that takes years to recover from. The lesson isn’t that primes shouldn’t do fixed-price work. The lesson is that doing FFP work competitively requires a fundamentally different operating model, not just a different contract type.

And then there is the internal cannibalization problem. If a prime invests heavily in autonomous attritable systems that could eventually substitute for crewed platforms it currently produces, the short-term effect is cannibalization of backlog. But boards don’t reward that kind of strategic clarity. Program managers don’t advocate for it. Business development teams are incentivized to protect existing wins. The result is a pattern of innovation investment that is additive and adjacent rather than disruptive and substitutive, which feels like progress but doesn’t actually change the competitive position.

None of this shows a failure of leadership or strategic vision in isolation. It shows rational behavior within the incentive structure these firms were designed for. The problem is that the market is shifting underneath that structure faster than internal incentives can adapt.

The Department of War’s shift toward firm fixed price contracts over the last few years is widely discussed as a mechanism for improving cost discipline and reducing schedule risk… which is accurate but incomplete. FFP is also a selection mechanism. It is quietly restructuring which firms can compete and on what terms.

Neo-primes were built for this environment. Their economic model is fundamentally different from legacy primes in ways that matter directly for FFP performance.

Software-defined margins change the economics of production in a way that legacy hardware-dominant cost structures can’t easily replicate. Companies like Anduril and Palantir embed software deeply into hardware and systems architectures, and software, once developed, carries near-zero marginal cost to replicate. When the cost structure of a weapons system includes a significant software component that doesn’t scale linearly with unit volume, the economics of fixed-price contracting become much more favorable. The prime that can amortize software development across large unit volumes or across multiple programs has a structural cost advantage that a labor-intensive manufacturer cannot easily close.

Automation does something similar on the production side. Companies building AI-powered, highly automated manufacturing are compressing cycle times, reducing per-unit labor content, and producing at scale with a cost structure that is inherently more compatible with fixed-price risk. When the marginal cost of production is well-understood and largely driven by machine time rather than labor variability, bidding firm fixed price is a manageable risk. When cost is dominated by skilled labor in a tight market, overtime exposure, and program-specific tooling (the hallmarks of a large legacy defense program), FFP becomes a much harder bet.

Private capital tolerates reinvestment in a way public markets don’t and that gap compounds over time. Neo-primes that remain privately held, or that were built on venture capital, can absorb early losses, invest ahead of revenue, and build infrastructure for future programs without quarterly justification. That tolerance for early-stage investment creates a compounding advantage, because each successive program builds on shared infrastructure, shared software, and shared production capability. The cost basis improves over time. The public prime has to justify each capital deployment against a specific program’s funded backlog, which means the infrastructure never gets built at the portfolio level where it would actually generate the flexibility the market now requires.

So what does this mean? It means that neo-primes can bid FFP work at price points that would be financially destructive for a legacy prime operating with a traditional cost structure. This isn’t a temporary phenomenon or a function of neo-primes underpricing to gain market share. It is a structural divergence that will widen as automation matures, software content in defense systems increases, and DoW continues to prefer FFP vehicles for new programs.

FFP isn't really a cost discipline problem, it's an operating model problem, and the reason most primes are struggling with it isn't a talent gap or a motivation gap, it's that the system they built was never designed to win that way.

So what do legacy primes actually do about this?

The honest answer is that most of them are currently doing something that feels strategic but isn’t. They are hedging. They are running innovation programs that don’t threaten core business lines. They are signing partnership agreements with neo-primes that give them market access without requiring transformation. They are acquiring small companies for capability signals rather than operational integration. They are doing enough to have an answer to the question without doing enough to change the answer.

That hedging posture made sense five years ago but it is increasingly untenable as FFP volumes grow, as neo-primes scale, and as DoW acquisition offices gain confidence awarding larger programs to non-traditional vendors.

The way I see it, there are three strategic paths available. Each has a different capital requirement, a different risk profile, and a different end state. What is not available is continuing to operate all three simultaneously without actually choosing.

This is the hardest path and the most valuable one if executed well. It requires primes to shift their core identity (not just their product portfolio) from program-anchored platform owners to operators of configurable industrial networks.

In practical terms that means building or acquiring automation capability designed for redeployment across programs rather than dedicated to individual contracts. It means restructuring supplier relationships so that capacity is contracted at the portfolio level and can be reallocated as priorities shift. It means investing in software platforms that serve as integration layers across multiple hardware programs, allowing cost to be amortized in ways that support FFP bidding. And it means restructuring internal capital allocation so that enterprise-level infrastructure is evaluated on portfolio optionality, not just program-level return.

This path isn’t about becoming asset light. It is about holding a different mix of assets, configurable, software-enabled, cross-program, rather than bespoke, program-dedicated, and labor-intensive. The primes that execute this transition successfully will retain their advantages in systems integration, certification authority, and sovereign assurance while adding the cost structure and agility needed to compete for the programs that matter most over the next decade.

The barrier is primarily internal (surprise). Governance structures, investor expectations, and incentive systems are not currently designed to reward this kind of transformation, and executing it requires sustained leadership commitment across multiple earnings cycles before the financial results become visible. That’s a hard ask in a public company environment, but it is the ask that genuine transformation requires.

There is a window, and it is not unlimited, during which legacy primes can acquire neo-prime firms before those firms either go public, reach sufficient scale to be acquisition-proof, or develop the sovereign assurance and integration credentials to compete directly for programs that are currently prime territory.

Strategic acquisition done well is not about buying innovation theater. It is about integrating operating models. A prime that acquires a SW-defined weapons firm and genuinely integrates its capital structure, production philosophy, and contracting approach into its core business is making a fundamentally different bet than one that acquires the same firm and manages it as a protected subsidiary, insulated from the parent to avoid disruption.

The risk of this path is integration failure. Defense primes have a mixed track record on acquisition integration, and the cultural distance between a venture-backed software-forward defense firm and a publicly traded prime with decades of cost-plus history is substantial (I can attest, I’ve lived both). Acquisitions that don’t produce operating model change don’t produce competitive position change, they produce expensive subsidiaries that eventually (and inevitably) stagnate.

The window also matters because neo-prime valuations are rising and their independence is increasingly strategic. Firms that needed acquisition partners five years ago are increasingly in a position to dictate terms or decline entirely. Ongoing conflicts in Iran are demonstrating how low-cost alternatives are succeeding in the battlespace and the primes are, without a doubt, paying close attention. The time to act on this path is not after the next major FFP award goes to a non-traditional vendor.

This is the path that most primes are implicitly drifting toward without acknowledging it, and there is a version of it that is strategically coherent rather than just a retreat.

If neo-primes increasingly own the elastic, software-defined, FFP-competitive layer of the defense industrial base, the attritable systems, the autonomous platforms, the rapidly iterated software-integrated weapons, then legacy primes might find their most durable competitive position is as the sovereign assurance and sustainment layer that neo-primes can’t yet replicate.

Classified program access, ITAR-controlled technology, long-term sustainment infrastructure, systems integration for the most complex crewed platforms, and the industrial relationships that underpin strategic deterrence are not capabilities that a venture-backed firm scales quickly. Those are genuinely defensible positions, and they represent real strategic value, although neo-primes have been closing that gap faster than most primes would like to admit.

The risk of this path is that it’s a narrowing market. The programs that require this level of sovereign assurance and integration depth represent a stable but not growing share of the defense budget. If DoW continues shifting discretionary dollars toward autonomous systems, software-defined weapons, and attritable platforms (which it is), the sustainment layer becomes a more concentrated and lower-growth business over time.

Owning this path explicitly means investing in the capabilities that make it defensible, classified infrastructure, long-cycle sustainment, integration depth on programs that genuinely require it. It means not pretending to compete in the FFP autonomous systems market when the economics don’t support it. And it means being honest with investors about what kind of company this actually is: a high-certainty, lower-growth sovereign infrastructure provider, rather than maintaining the guise of transformation without the substance behind it.

The strategic error most primes are making is not choosing the wrong path. It’s not choosing at all.

Running transformation, acquisition, and sustainment focus simultaneously without real commitment to any of them produces the worst of all outcomes. The transformation investments are too small to change the operating model. The acquisitions are too isolated to change the cost structure. The sustainment position erodes as DoW diversifies its supplier base. And the quarterly earnings pressure that makes real transformation difficult continues unabated in the background of all of it. Somewhere on an earnings call, an analyst asks about "strategic optionality" and a CFO gives a confident non-answer, and everyone moves on.

The defense market is not going to pause while primes figure this out. DoW acquisition offices are gaining confidence with non-traditional vendors. FFP contract volumes are growing. Neo-primes are scaling. The supply chain infrastructure being built by a new generation of AI-powered, highly automated manufacturers is creating an industrial layer that didn’t exist five years ago and will be substantially more capable five years from now. That matters directly for the prime transformation question because the availability of sophisticated high-throughput manufacturing capability that is not tied to any single prime’s internal production network makes the orchestrator model more feasible. Primes don’t have to build all of the elastic capacity themselves. But they do have to develop the integration and allocation capability to use it, and they have to be willing to restructure contracting, capital, and supplier relationships to actually get there.

The primes that navigate this transition successfully will be the ones that choose a path clearly enough, and early enough, to actually execute it. The ones that don’t will find that the market has made the choice for them, and the options available at that point will be considerably less attractive.

Industrial behavior reflects the system we designed. The system is changing, and the question is whether primes are going to shape that change or react to it. I spent a long time inside that system, and I have a lot of respect for the people trying to move it. But respect doesn't change the math, and the math is moving faster than most earnings calls would suggest.