The IRAD Trap: Why Independent Research Funding From Defense Primes Is Rarely as Independent as It Looks

Independent Research and Development funding sounds like a gift. A prime contractor writes an internal budget line, does work it believes will matter strategically, and the government reimburses a chunk of the cost through overhead rates on existing contracts. No deliverables, no program office breathing down your neck, no RFP. Pure exploration.

Photo by Tima Miroshnichenko on Pexels.

Except it rarely stays that way.

For defense tech startups, IRAD creates a specific trap that plays out in at least three different forms. None of them are obvious at term sheet time, and all of them are expensive to unwind.

What IRAD Actually Is (And What It Creates)

Primes can charge up to roughly 3% of their total contract costs as IRAD expenses, recoverable through overhead on cost-type government contracts. The Defense Contract Audit Agency reviews these annually. The notional purpose is to fund long-horizon technical work that keeps the prime's capabilities ahead of potential contract requirements.

In practice, primes use IRAD budgets to pre-position for future competitions. They scout technology, prototype adjacent capabilities, and increasingly: make informal arrangements with startups that have something interesting on the bench.

Those arrangements come in several flavors. Sometimes it's a paid pilot disguised as a study contract. Sometimes it's a data-sharing agreement with a right of first refusal buried in section 12. Sometimes it's a joint development agreement where the work is funded by the prime's IRAD pool and the IP ownership language would make any competent patent attorney wince.

Founders see revenue and validation. Investors see a credible anchor partner. What both parties often miss is what the prime is building toward.

The Three Versions of the IRAD Trap

graph TD
    A[Prime Approaches Startup] --> B{Arrangement Type}
    B --> C[Paid Pilot / Study]
    B --> D[Joint Development Agreement]
    B --> E[Data Sharing + ROFR]
    C --> F(Background IP Contamination)
    D --> G(Foreground IP Dispute)
    E --> H(Acquisition Leverage)

Background IP contamination. In a joint development arrangement, primes almost always insist on license rights to background IP that the startup brings to the table. The justification sounds reasonable: they need rights to the pre-existing technology to use whatever gets built together. But those license terms, if poorly scoped, can effectively give the prime a royalty-free, sublicensable right to your core technology stack. By the time a Series B investor's counsel flags it, the prime has already used that IP in a proposal for a program you've been trying to compete for directly.

Foreground IP disputes at commercialization. Work funded (even partially) by a prime's IRAD budget can create joint ownership claims on whatever gets developed. Joint ownership of a patent in the U.S. means either party can license the technology to third parties without consent and without sharing proceeds. That's a structurally fatal problem for any startup trying to build a defensible market position. Investors who discover this mid-diligence either reprice hard or walk.

Acquisition leverage. This one is subtler. A prime that has done IRAD-funded work with your startup has already written internal memos positioning your technology as an extension of their roadmap. When they approach you about an acquisition, they come with a manufactured narrative about integration and they come with information about your technical maturity that you gave them willingly. The information asymmetry is entirely in their favor. Right of first refusal provisions compound this: a ROFR on an acquisition means every other potential buyer knows they may be negotiating just to set a floor price for the prime.

What Diligence Should Actually Look Like

Venture investors doing diligence on defense tech companies need to treat every prime-adjacent commercial arrangement as a potential IP encumbrance. The questions that matter:

Who funded the work, even partially? IRAD dollars that flow through a subcontract or a service agreement still carry the prime's IP expectations. Ask for the source of funds behind every commercial agreement signed in the prior three years.

What rights did the prime receive on background IP? Get the actual license language, not the founder's characterization of it. Founders consistently underestimate how broadly defined "background IP" becomes in these agreements.

Is there any ROFR, right of first negotiation, or consent right on a change of control? These provisions are common, they are often not disclosed proactively, and they are deal-killers in competitive acquisition processes.

Has the company disclosed its technology to the prime in any form that could constitute prior art or reduce the novelty of pending patents?

What Founders Should Do Before Signing

Not all prime relationships are traps. Some produce genuine customer traction and technical validation that helps a startup navigate the valley of death faster. The difference is almost entirely in the IP terms negotiated at the outset.

Scope background IP licenses narrowly and tie them to a specific use case. Never grant sublicensable rights. Push for field-of-use restrictions that preserve your ability to compete in adjacent commercial markets.

Negotiate sunset provisions on ROFRs, or refuse them entirely. A right of first negotiation is less damaging than a right of first refusal; know the difference before you sign.

Get independent IP counsel before any joint development agreement is executed. Not your general corporate counsel. Someone who specifically works on government contractor IP disputes and understands DFARS rights clauses, because those will eventually govern how the government itself can use what you build.

The prime's IRAD budget exists to serve the prime's competitive positioning. That objective is not aligned with yours. Treat every dollar that comes attached to a prime's internal research program as a contingent liability until counsel tells you otherwise.