Rebellion Defense had the ingredients: an AI platform for mission planning and decision support, a credentialed team, and serious VC backing. They won early DoD prototype contracts. Then the procurement timelines stretched — months became years.Rebellion Defense won early DoD prototype contracts but burned through funding as production timelines stretched. The company lost co-founders and shuttered UK operations before stabilizing.
The company burned through funding chasing contracts that didn't materialize on schedule, lost co-founders, and eventually shuttered its UK operations. Rebellion survived, but barely.
Figuring out what to do between prototype and production contract is a big part of surviving the defense tech valley of death. A GAO report on the Defense Innovation Unit found a 51% cumulative prototype-to-production transition rate from FY2016 through FY2023, but FY2023 alone saw only 10 transitions, down from 17 the prior year.GAO report on DIU: 51% cumulative prototype-to-production transition rate from FY2016–2023. FY2023 saw only 10 transitions, down from 17 the prior year. The best fast-track pathway the DoD has built only converts half the projects that enter it, and it is slowing down.
Defense tech founders and their investors often struggle to define a viable path from prototype to production contract. The reasons are structural: there is little industry training on how to work through a procurement system built for established primes, it is hard to compete against defense contractors who are specifically selected and trained in compliance-first proposal writing, and the DoD often expects the prime industrial base to own production-scale delivery, making startup ownership of that transition genuinely ambiguous.
In reality, good acquisition strategy is often a very simple and boring review of small amounts of data: contract vehicle types, transition rates, funding timelines, program-of-record dependencies.
When Reform Programs Inherit The Problem §
The classic first attempt at fixing the valley of death is to ask program offices what reforms the acquisition system should pursue.
This usually ends in sadness, because: without time to think, people give bad answers; people over-index on opinions of people in positions of authority; and people offer solutions, not problems to solve.
You get a list like: create an Other Transaction Authority pathway, expand SBIR Phase III transition incentives, stand up a Rapid Defense Experimentation Reserve. The implicit "we should do this because it solves X problem" gets lost.
The main fault with this approach is that it does not capture and solve the biggest problems.
When you prioritize solutions, your job is done once everyone agrees to do your idea. When you prioritize problems, you prioritize fixing the issues you actually have. You cannot prioritize problems for years, execute effectively, and not solve your problems.
The RDER is the clearest example. Senate appropriators concluded that it "has not resulted in accelerated fielding outcomes," a reform mechanism built specifically to bridge the valley of death that inherited the system's inertia instead.Senate appropriators on RDER: the program "has not resulted in accelerated fielding outcomes" despite being designed specifically to bridge the prototype-to-production gap.
The NPS Acquisition Research Program found that Air Force SBIR Phase II-to-Phase III transition rates ran 8.8% between 2015 and 2018. Meanwhile, a separate GAO assessment puts the average time from program start to initial operating capability at 12 years for major defense acquisition programs.NPS: Air Force SBIR Phase II-to-III transition rate was 8.8% (2015–2018). GAO: average time from program start to initial operating capability is 12 years for major defense acquisition programs.
The solutions have been executing for years. The problems remain.
Where The Acquisition System Actually Breaks §
The a-ha moment is this: everything you're doing should be solving a problem. Align defense tech founders, their investors, and program offices on the biggest structural problems in the transition pipeline. Derive your reform strategy based on solving those problems. Periodically revisit your problems list to make sure it is still accurate.
Most defense acquisitions have problems that live in the following places: the OTA-to-FAR handoff where prototype winners face compliance requirements designed for primes; requirements documents that can take 18–36 months to finalize after a prototype succeeds; ITAR/security certification timelines that assume a large contractor's compliance infrastructure; the funding gap between a prototype award ending and a program-of-record beginning; congressional budget cycles that don't map to startup runway; and the absence of a bridge financial instrument between development and production.
These transition failure points are auditable and you can directly build a prioritized list of problems based off of them.
Take the OTA-to-FAR handoff. A startup wins a prototype under Other Transaction Authority on technical merit. The evaluation criteria reward capability. The production contract defaults to FAR-based procurement, where evaluation criteria reward compliance documentation, past performance at volume, and financial reserves that primes have accumulated over decades.
The DoD's November 2025 Acquisition Transformation Strategy redesignates the entire system as the "Warfighting Acquisition System" with speed as the principal measure of effectiveness. Whether renaming translates to restructuring is the open question, but the OTA-to-FAR cliff remains.
When The Technology Wins And The Startup Doesn't §
Production contract prioritization is notoriously fraught in the defense industry. One major contributor is that defense tech startups are pretty bad at articulating why their path to production matters more than a prime's.
Another contributor is that prime contractors often require FAR-compliant proposal infrastructure to compete at production volume, which structurally disadvantages startups. Program offices then resolve the tension through percentage-based work share arrangements with established primes rather than full capability transitions.
Terran Orbital built small satellites for the Space Development Agency's Transport Layer, the mesh network of LEO satellites the Pentagon needs for resilient communications and targeting. Breaking Defense reported that the satellites worked. Terran won early SDA contracts and demonstrated smallsat manufacturing capability.Terran Orbital demonstrated smallsat manufacturing capability for SDA's Transport Layer but couldn't sustain independent operations. Lockheed Martin acquired the company for $450M in October 2024.
Then came the balance sheet problem: sustaining independent operations against Northrop Grumman and Boeing required capital reserves that SDA contract payments alone couldn't provide. Lockheed Martin acquired Terran Orbital for $450M in October 2024. The technology won. The startup didn't — Lockheed absorbed the capability into its satellite division.
The contrast is Anduril, which Business Insider reported won a $642M Marines counter-drone contract competing against nine other bidders. Anduril broke through to major production by building acquisition infrastructure (legal, compliance, business development) as a core competency alongside engineering.Anduril won a $642M Marines counter-drone contract against nine competitors — a breakthrough achieved by treating acquisition infrastructure as a core competency, not an afterthought.
Pretty quickly you can get to defense tech principles about burn-rate-to-contract-timeline ratios, compliance infrastructure investment levels, and acquisition team headcount if you frame things as solving problems and believe it can be done.
What Founders And Investors Should Do Differently §
If you are a defense tech founder working on acquisition strategy, look at where your pipeline problems actually live. Do not stare at one contract opportunity and lament the lack of a clear evaluation framework. That is where a single problem lives.
Instead, look at the structural patterns: OTA vehicle type, program office budget cycle, prime teaming relationships, compliance certification lead times. Those patterns justify or foreclose your path to production. Saronic's prototype-to-production timeline on the Navy autonomous surface vessel program ran under a year and yielded a $392M contract. That is an outlier, but it is an auditable one: the Navy was the motivated buyer, the requirements were already defined, and the contract vehicle was designed for speed.Saronic's Navy autonomous surface vessel contract: prototype-to-production in under a year, $392M award. An outlier enabled by a motivated buyer, pre-defined requirements, and a speed-optimized contract vehicle.
If you are a defense-focused VC that wants your portfolio companies to survive the valley, expose them to the transition data early. If you only ever show founders term sheets and portfolio milestones, do not be surprised when you have a portfolio full of acquisition-naive teams.
Explicitly structure bridge capital for the 12-to-24-month gap between prototype award and production contract award. This is not speculative; it is the documented timeline.
The UK's DASA offers Defence Innovation Loans designed specifically to bridge development to commercialization. The US system has no equivalent instrument.
Whoever you are, always ground your organization in the why of what you are doing. Once you lose sight of the problem, you've definitely lost sight of the solution.
Where The Transformation Strategy Stalls §
One irony is that the DoD can pursue acquisition transformation to a fault. The data on what startups need is not hard to find, and sometimes the competitive pressure from adversaries makes the required reforms obvious without extensive research. The November 2025 "Warfighting Acquisition System" rebrand risks becoming exactly the kind of solution-first move this piece has been describing: a reform that executes effectively for years and still leaves the OTA-to-FAR cliff intact.
The startups that survive will be the ones that treat acquisition as a core competency from day one — which is a reasonable adaptation to an unreasonable constraint.
But that's the part worth sitting with. Everything described here is basically ersatz program management that startups should never have needed to develop in the first place.