Who Funds the Military Roadmap

Why the Prime-Integrator Fix Leaves the Army With Software That Stops Improving

In my last post Big Budget, Small Market, I argued that most DoW software categories are too small to fund a venture-scale company on their own, and that in human performance no single buyer owns the soldier, so no DoW-only startup can raise enough to build the integrated bundle the Army wants. A fair reader's response: fine, let a prime integrator assemble it. Take the best commercial point solutions, wrap them in one contract, handle the compliance and hand the Army a single front end. That model exists, and it's how some bids for the Army's Holistic Health and Fitness Management System (H2FMS) are being put together. My question for anyone who likes it: who pays for the next five years of innovation inside that bundle?

I don't think anyone does.

Credit where it's due

The integrator brings skills the Army needs. The Army's H2FMS call for solutions in September 2025 asked for a unified digital platform built on a modular systems architecture, awarded through OTAs, covering the physical, mental, spiritual, nutritional and sleep domains (link). Turning that into a compliant, authorized, supportable system is hard work. When GAO looked at FedRAMP in 2024, it found that agency and vendor estimates of authorization costs "ranged anywhere from tens of thousands to millions of dollars," and DoW impact levels add more on top (link). Plenty of good small companies can't carry that alone. Contracting skill matters too. Primes know how to win and run programs, and the Army needs someone who does.

My critique is more about structure. A well-run integrator can still produce a bad outcome if the incentives point the wrong way.

What happens to the subs

Start with revenue. For a small strength-training or nutrition software company, an Army program the service says will reach 129 active-duty brigades by fiscal 2029 is bigger than a contract (link). It becomes the company. DoW revenue turns into the largest line on the P&L almost overnight, and that money flows through someone else's contract. The Congressional Research Service spells out the legal reality: because subcontractors "do not have a direct contractual relationship with federal purchasing agencies the way that prime contractors do, they lack some of the legal protections that benefit prime contractors" (link). Your biggest customer isn't even your customer.

Then the authorization boundary. Under DoW's Cloud Computing Security Requirements Guide, the mission owner's ATO covers the application, which inherits controls from the authorized cloud offerings beneath it, and "a CSO with a DoD PA does not eliminate the requirement for a given application using the CSO to have an ATO" (link). When the integrator builds and owns that boundary, the sub's component is authorized as part of the prime's system. The sub doesn't walk away with an authorization it can take to another command. It deploys where the prime deploys, and nowhere else.

Then the user. The prime owns the relationship with the program office, the commanders and, in practice, the practitioners. This is the cost I care about most, because I've lived it.

As I wrote in our first piece in this series (link), Teamworks worked as a subcontractor under Booz Allen on SOCOM's Preservation of the Force and Family program. We had no direct path to the end customer. We couldn't sit with the strength coaches, dietitians and psychologists using our software and ask what was broken. Requests, complaints and ideas reached us filtered, late or all too often not at all; we found out about problems the way you find out about a surprise party: last and from someone else. Not being able to talk directly to the customer was an enormous blow to the value we could deliver. It was bad for us, and worse for the operators the program existed to serve.

Software gets better because the people building it hear from the people using it, constantly. The Army knows this. After two days at Fort Benning evaluating candidate H2F app designs, Jonathan Flinn, H2F program director with the 199th Infantry Brigade, told Common Defense, "I need the soldier to want to use this management system," and said earlier efforts, including at the 75th Ranger Regiment, failed because they didn't secure participation from end users (link). You can't earn that participation through two layers of contract.

The captive exit

Now look at the sub from an investor's chair. Its largest customer sits behind a prime. Its authorization lives inside the prime's boundary. Its user relationship runs through the prime. Who buys this company? Realistically, one party: the prime. That makes the exit captive, and a captive exit is priced by its only bidder, who knows it. Defense tech already skews this way. PitchBook's September 2025 analyst note "Liquidity Horizons in Defense Tech" found that M&A accounts for roughly 72% of defense tech exits (exit counts peaked at 74 in 2021 and came back to 71 in 2024), while IPOs remain rare, "reserved for scaled dual-use firms like Palantir, Rocket Lab, and Voyager" (link). A human performance point solution living inside one prime's contract won't be the exception.

Outside capital sees that ceiling and stays home. Venture investors underwrite the chance of a large, contested outcome. Remove the competition for the asset and you remove the reason to fund it. The innovation money won't come from capital markets.

Could it come from the prime? In principle. Primes do run venture arms, and some are growing fast: Lockheed Martin raised its fund's capacity from $400 million to $1 billion in April 2026 (link). But that money chases quantum computing, autonomy, AI, directed energy and microelectronics, the technologies that feed big hardware programs. I have yet to meet a corporate venture team with a thesis on brigade dietitian workflows. And an integrator on a software program is paid for integration and labor, not for R&D inside its subcontractors. In H2F, the human performance staffing is a separate contract altogether; the Army awarded one H2F services contract worth up to $247 million in January 2025 (link). Unless the software integrator writes meaningful equity checks into the companies whose products it bundles, nobody funds their roadmaps. The sub gets a shot in the arm from DoW revenue, then spends the next five years serving one master on a flat budget.

The exception that proves the rule

Oura is the counterexample, and it makes the point for me. Oura has worked with the military since 2019 and has deployed tens of thousands of rings across the services, by its own count; its defense contracts have totaled more than $20 million, according to HigherGov data cited by the Dallas Morning News (link). The company calls DoW its largest enterprise customer and is building manufacturing in Fort Worth to support that defense business. In 2024 the Defense Health Agency announced a $96 million award for Oura rings and services; after bid protests, DHA cancelled the solicitation in March 2025, saying it was "no longer considered a priority" (link, link).

None of that was existential for Oura. Its S-1 reports $1.21 billion in revenue for the nine months ended June 30, 2026, up 74%, with 5.0 million paid members and 3.6 million rings sold over the trailing twelve months (link). It postponed its Nasdaq IPO on September 29, citing uncertainty in the IPO market, while saying paid members had reached 5.7 million and that the business is profitable (link). Consumers fund Oura's research. The military gets the benefit of that research at a marginal price. If a DoW contract shifts, the roadmap keeps moving.

Palantir taught the same lesson in piece #1. When a market bigger than DoW pays for the platform, DoW revenue is incremental, and the government gets a vendor that keeps improving whether or not the next option year is exercised.

What the Army actually gets

Follow the integrator model to its end state. The Army gets one login (maybe) and one dashboard (again, maybe) on top of components that stopped improving the year the contract was signed. Soldier data sits split across vendors, each holding a slice of the training, nutrition, sleep and injury history. When a sub fails, gets acquired or is swapped out at recompete, the unit loses that history, or inherits it in a format nobody maintains.

H2F's own economics make this expensive. The Army says data currently being peer-reviewed shows $8.15 in cost avoidance and readiness gains for every dollar spent on H2F (link). That return (or at least the ability to support it with data) depends on longitudinal data and on software that keeps getting better at turning it into decisions for commanders and practitioners. When the vendor stalls, the buyer pays.

Extend the Army's own logic

The good news is that the Army has already named the middleman problem. When it announced the Palantir Enterprise Agreement at the end of July 2025, consolidating 75 contracts under a ceiling of up to $10 billion over ten years, the Army said the deal was designed to cut out "contract and re-seller pass-through fees" (link). Army CIO Leonel Garciga said it would be "one of many enterprise licensing agreements," and the Army has signed enterprise-wide software deals with Salesforce and Appian as well (link, link, link). When Secretary Hegseth made the Software Acquisition Pathway the preferred route for software in a March 2025 memo, a defense official summed it up: "Software companies make software, we're going to buy software from software companies" (link, link). The FY2027 budget request goes further, proposing "$13.3 billion in contract services reductions to streamline external support" (link). It's a proposal, not an appropriation, but the direction is clear.

Apply that logic to H2F. Buy platforms whose core is funded by a market bigger than DoW, so the roadmap doesn't depend on one program's option years. Structure contracts so the government has a direct relationship with the companies building the technology: direct licenses, direct authorization paths, direct access to users. Pay integrators for integration where it's needed, and don't let the integration layer become the only door into the building.

That's why Teamworks built for sport first and brought the military a platform that the sports market had already funded and proven. Our commitment to this mission doesn't hinge on any single contract.

The soldier in a brigade at Fort Drum doesn't care who holds the prime. She cares whether her strength coach can see last month's training load, whether her dietitian's plan reflects how she's sleeping, and whether the tool she opens next year is better than the one she opened this year. Buy for that.