Something fundamental has changed in defense. For decades, procurement followed predictable rules. Long programs, slow cycles, heavy documentation. Investors could ignore the sector because nothing moved fast enough to matter.
Then Ghost Shark happened.
In May 2022, Australia gave Anduril a classified brief, a mandate, and a timeline. Three years later, they delivered a new class of autonomous underwater vehicle from a blank sheet using private capital, software velocity, and a development model that looked more like modern tech than legacy defense. They delivered three prototypes on budget and ahead of schedule, unlocking a program of record worth US$1.1 billion (1).
Defense is now “eating the world” (2), precisely the way software once did. But most investors continue to apply software logic to a system operating under sovereign constraints.
Software scales by finding users. Defense scales by surviving procurement gatekeeping and sovereign capability requirements. These are fundamentally different problems. For this guest article, Pierre-Marie Derouin joined us to break down how defense has shifted from slow procurement to software-speed execution.
Pierre-Marie Derouin. Photo Credit: Pierre-Marie Derouin
The Software Playbook Does Not Work Here
Classic venture capital assumes markets are open. A founder builds a product, finds product-market fit, scales to users, and raises capital based on growth signals.
This model collapses in defense. Not because defense is slow, but because defense is gated.
In software, a demo shows product potential. A pilot shows market validation. In defense, the same signals mean something entirely different. A demo shows technical potential, but nothing about procurement viability. A pilot is a test, not validation. A grant tells you the government is interested, not that it will buy. Most investors treat these as equivalent to software traction. They are not.
Ghost Shark had demos and prototypes. But what Ghost Shark actually had was something software companies almost never achieve: sovereign mandate. The Australian Department of Defence specified a requirement, allocated a timeline, and committed capital upfront. That is not a test but an acceptance.
Most defense investors miss this distinction. They see a company with defense contracts and assume traction. They see pilots and think product-market fit. Then, inexplicably, the company stalls. It remains in the prototype stage for five years. It consumes capital without converting pilots to production. The investors blame procurement cycles.
The real problem is structural. They invested in a company that had solved the technology problem without solving the procurement problem. In software, these are the same thing. In defense, they are entirely different gates.
Why Sovereign Constraints Rewrite the Rules
Defense does not work like a level playing field. Defense works inside sovereign systems. This changes everything.
Every defense transaction is a sovereign decision. The buyer optimises for security, strategic autonomy, and institutional control. That has three immediate consequences:
- Procurement gates do not open easily for outsiders. You cannot accelerate them through better fundraising or faster execution. You must survive institutional vetting. That is political, not commercial.
- Capability requirements are non-negotiable. A software company can pivot based on market feedback. A defense company operating under a sovereign mandate cannot. The specification is fixed. The timeline is fixed. Your job is execution fidelity, not innovation flexibility.
- Foreign direct investment (FDI) scrutiny has increased dramatically. Defense sector FDI cases now face the highest screening levels in Europe. Capital origin matters. A venture fund from Singapore faces different regulatory friction from a European fund. This friction is not temporary. It is structural.
Ghost Shark succeeded because Anduril was North American, and Australia is a Five Eyes ally. The sovereignty equation was aligned. Take the same company, same technology, different capital source or buyer nation, and the entire transaction becomes friction-laden.
Classic venture models do not account for this. They assume capital is fungible. It is not in defense. They assume procurement friction is temporary. It is not.
What Gate 0 Actually Is
Gate 0 is the point at which a sovereign buyer decides a company is structurally eligible to be trusted with sensitive work at scale. Before Gate 0, everything is exploration. After Gate 0, the system treats the company as a serious candidate for delivering real capability.
Institutional buyers screen five interdependent criteria. All five must be passed. Failure in any one usually blocks access to the most valuable opportunities:
- IP ownership and encumbrance status. Many founders discover far too late that research funding or academic arrangements have given others veto rights over commercialisation. For an institutional buyer, that is an unacceptable risk.
- Export control and ITAR or dual-use classification. A technology subject to restrictive export control is harder to field and share. For a European buyer, finding out that a promising system is effectively subject to U.S. export controls can render the entire proposition non-viable.
- Security clearance pathway feasibility. Can the people who build and support the system be cleared to work on sensitive programs? Teams that discover mid-sales cycle that key members cannot be cleared create bottlenecks that delay contract execution and erode buyer confidence.
- Sovereignty and data residency alignment. For European buyers, it matters where data is stored, where models are trained, and who controls the code. Ownership structures dominated by non-European investors and opaque data-handling policies create friction.
- Security incident and litigation history. Undisclosed IP litigation or security breaches discovered during diligence are treated as signals of founder inexperience or deliberate opacity. The default answer becomes no.
Gate 0 is not a dimension to be weighed. It is a gating requirement that must be satisfied before any other evaluation proceeds.
The New Temporality After Gate 0
Once Gate 0 is passed, the system behaves very differently.
Ghost Shark shows the pattern. From classified brief to prototype delivery was three years. From prototype acceptance to program of record was months. From program of record to production scaling was another 18 months (3). That is software velocity inside a defense wrapper.
This acceleration only happens when two conditions are met. First, the sovereign buyer has accepted the operational concept and is willing to take technical risk early. That is Gate 0. It is a strategic decision to adopt a new capability model.
Second, the company is structurally compatible with that acceleration. It has a clean IP, an understood export status, a realistic clearance pathway, sovereign-aligned ownership, and no hidden litigation or incident history. Most defense companies are not built this way. They are configured for 10-year programs. Put them on a three-year timeline with fixed specifications and a sovereign mandate, and they break.
Anduril could execute because it combined software style autonomy with people who understood how military systems and security constraints work. That combination is still rare.
What This Means For Investors
The New Defense model requires investors to think like operators, not like generalist venture capitalists. There are three new landmarks they must navigate:
- Traction looks different. Demos, pilots, and grants are not forward motion. They are tests that precede Gate 0. The real question is whether a company is close to passing Gate 0 screening and securing a sovereign mandate.
- Capital structure matters more than capital efficiency. Mismatched capital creates friction that even excellent execution cannot remove. European capital has advantages in European procurement. Allied capital has advantages in NATO contexts.
- Team composition must include institutional operators. People who have lived inside defense ministries, procurement agencies, or classified programs understand how Gate 0 works in practice.
The companies that thrive in the New Defense understand this is a different game with different gates, different constraints, and different notions of traction. They design for Gate 0 from day one and then use the acceleration it enables.
The Ghost Shark program will not be an outlier. It will be the template. But that template only works for companies and investors who understand why Gate 0 exists and build around it.
Sources
- Reuters — “Australia to spend $1.1 billion on Anduril undersea drone fleet”, 9 Sept 2025
- Resilience Media — “State of Defence Tech 2025”, 28 Sept 2025
- VoxelMatters — “How LFAM enabled Anduril’s Ghost Shark Program of Record”, 15 Oct 2025
