A missing signature or ambiguous clause in IP assignment agreements can derail deals worth millions in minutes.
I was on a final diligence call for a Series A investment. The tech was brilliant, the team was sharp, the market was wide open. It felt like a sure thing. Then, the investor's counsel asked a simple question: "Can you send over the executed IP assignment from the contractor who built your MVP?" Silence. The founder stammered, "We paid his invoices... isn't that enough?" In that moment, I watched a $20 million valuation evaporate. The deal didn't just get a haircut; it died on the spot.
This isn't a rare occurrence. It's a landmine that many investors and founders don't see until it's too late. An Intellectual Property (IP) assignment agreement isn't just a piece of legal boilerplate; it's the deed to the core asset you're investing in. And in my 25 years of building and auditing tech companies, I've learned that the state of these documents is the most reliable, unforgiving mirror of a founding team's discipline and their ability to execute. It tells you whether you're investing in bedrock or quicksand.
The Unseen Foundation: Why IP Assignment Isn't Paperwork, It's the Deed to Your Investment#
For the vast majority of technology startups, the intellectual property is the business. The code, the algorithms, the brand, the data models—these aren't just assets; they are the engine of value and the primary barrier to entry for competitors. Without clear, undisputed ownership of this IP, the company effectively owns nothing of value. This is where the legal concept of the "chain of title" becomes paramount.
Think of it like the deed to a house. A clean chain of title is an unbroken paper trail of contracts, signatures, and legal filings that proves, without a shadow of a doubt, that every single contribution—from the first line of code written by a founder in their garage to the logo designed by a freelancer—is legally owned by the company entity. This paper trail is composed of IP assignment agreements, which are the legal instruments that transfer ownership from the individual creator to the company.
For an investor, verifying this chain of title isn't optional; it's a fundamental part of due diligence. Investors are paid to take on market risk, product risk, and execution risk—but they are not paid to absorb fundamental legal ownership risk. A flaw in the chain of title is an unforced error, a sign of sloppiness that can render an entire investment worthless before it even begins.
The Investor’s X-Ray: Reading the Chain of Title as a Founder Competency Audit#
I don't look at an IP audit as a legal checklist. I see it as a founder competency audit. The technology, and the legal paperwork that surrounds it, is a direct reflection of the leadership's discipline, foresight, and systemic thinking. A messy IP chain isn't just a legal problem; it's a symptom of a chaotic, reactive, and undisciplined team—the very definition of execution risk.
This reframes the entire purpose of IP diligence. While most investors view it as a defensive, risk-mitigation step, it's actually one of the most powerful offensive diagnostic tools available. A two-hour review of a startup's IP assignment practices can reveal more about a team's ability to execute than a two-hour product demo. It's a cheap, high-signal method to diagnose the core competency and discipline of the founding team before committing capital.
Specific gaps in the chain of title are direct indicators of deeper operational weaknesses:
- A missing founder assignment for pre-incorporation work? This signals a lack of strategic foresight and an inability to handle the most foundational basics of company formation.
- No agreements with early contractors? This reveals a naive belief that paying an invoice equals ownership—a critical and costly business misunderstanding that suggests a lack of sophistication.
- Inconsistent documentation and record-keeping? This points to systemic "process debt" and a lack of operational rigor that will inevitably plague product development, hiring, and scalability down the road.
The following table translates these legal red flags into the founder competency signals they truly represent.
| IP Red Flag | What It Looks Like | What It Really Means (The Founder Competency Signal) | Potential Valuation Impact |
|---|---|---|---|
| Missing Founder Assignment | A founder created the core algorithm before the company was incorporated, with no document transferring it to the new entity. | Lack of Foundational Alignment & Foresight: The founders failed to address the most basic step of securing their core asset. This predicts future chaos in governance and strategic planning. | Critical / Deal-Killer |
| Vague Contractor Agreement | A freelancer's agreement states they "will assign" IP upon final payment, but lacks the present-tense language "hereby assigns". | Operational Naivete & Legal Sloppiness: The team doesn't understand fundamental contract law and relies on assumptions. This signals they will likely make similar costly mistakes with customer or partnership contracts. | High / Significant Valuation Haircut & Delays |
| Unmanaged AI-Generated Code | The codebase is full of snippets from GitHub Copilot, with no policy or tracking of its use or the licenses of its training data. | Blindness to Emerging Risks: The team is chasing productivity gains without understanding the complex, emerging IP risks. This indicates a lack of strategic depth and an inability to manage a modern tech stack. | Medium-High / Requires Deep Audit & Potential Rework |
The Ghosts in the Machine: Three Classic Skeletons in a Startup's Closet#
Over the years, I've seen the same preventable mistakes kill deals time and again. I think of them as the ghosts that haunt a startup's cap table, often lying dormant until the high-stakes pressure of a funding round or acquisition brings them to light.
The Ghost of the Departed Co-Founder#
I saw a deal where a co-founder left on bad terms after six months. A year later, with a term sheet on the table, he re-emerged. Because his initial IP assignment was never signed, he claimed 50% ownership of the core technology. The deal collapsed, and the company was held hostage by its own creator. This is a classic scenario that every investor fears. Without clear, upfront assignment, a departing founder can walk away with a critical piece of the company's value, potentially even becoming a direct competitor using the very technology the startup was built on.
The Ghost of the "Helpful" Contractor#
Under U.S. law, there's a dangerous misconception that paying a freelancer means you own their work. In reality, without a written agreement stating otherwise, an independent contractor retains ownership of the IP they create. The "work for hire" doctrine is narrow and often doesn't apply automatically to software development.
I witnessed this firsthand during an acquisition diligence. A startup had paid a freelancer $5,000 to build their initial app. Two years later, the buyer discovered there was no IP assignment in place. The freelancer, now aware of the company's multi-million-dollar valuation, demanded $500,000 to sign the rights over. The acquirer, unwilling to take on the risk and delay, walked away from the deal. A small, early oversight created a massive, deal-killing liability.
The Ghost of a Single Word: "Hereby Assigns" vs. "Will Assign"#
This is the most subtle, yet one of the most lethal, ghosts. In a landmark case, Stanford v. Roche, the U.S. Supreme Court drew a critical distinction between two phrases in assignment agreements.
- "hereby assigns" is a present-tense transfer of ownership. The moment the agreement is signed, the IP belongs to the company.
- "will assign" is merely a promise to transfer ownership in the future. This implies a second action is needed, and if the creator refuses, the company is left with a breach of contract claim, not the IP itself.
It's the difference between a seller handing you the keys to your new house versus them saying, "I promise I'll give you the keys later." One gives you ownership now; the other gives you a potential lawsuit. An agreement with the wrong phrasing is a ticking time bomb in the cap table, creating a fatal flaw in the chain of title that a sharp-eyed legal team will spot in minutes.
The New Deal-Killers: Navigating IP Minefields in the Age of AI#
The classic ghosts are bad enough. But the rise of generative AI has summoned a new generation of deal-killers—ones that most diligence checklists haven't even caught up to yet. This is where a simple legal review fails and a deep, systemic understanding of modern technology becomes critical.
The Phantom Author: Who Owns AI-Generated Code?#
Ask yourself this question: if your portfolio company's developers use GitHub Copilot to write 40% of their code, who owns that 40%? The answer is terrifyingly ambiguous. Under current U.S. Copyright law, a work must have "human authorship" to be protectable. In the case of
Thaler v. Perlmutter, courts have affirmed that work generated "autonomously" by an AI system is not copyrightable.
This means a significant portion of a company's core asset—its source code—may not be ownable at all, potentially placing it in the public domain. While GitHub's terms of service state that the user owns the output, these terms cannot override federal law. This creates a massive, uncharted ambiguity in the chain of title. A competitor could potentially copy large swaths of AI-generated code without legal consequence, rendering the startup's supposed "proprietary" technology completely indefensible.
The Shadow AI Leak: Your IP in ChatGPT's Brain#
Consider this scenario: your portfolio company's engineer is stuck on a complex, proprietary algorithm. To be more productive, they paste the source code into a public LLM like ChatGPT or Gemini and ask for help debugging it. This isn't just a query. That proprietary data can be absorbed into the model's training set, becoming a permanent, irreversible leak of trade secrets.
This phenomenon, which I call "Shadow AI," represents a new vector for IP exfiltration that bypasses all traditional security perimeters. The company's most valuable secrets are not being stolen by a hacker, but leaked by a well-intentioned employee with a web browser. This blurs the line between IP protection and cybersecurity. A company's IP strategy is now inextricably linked to its internal data governance. During diligence, the question is no longer just "Can I see your IP assignments?" It must now be, "What is your policy on employee use of public generative AI tools?" The absence of a robust policy is a critical new red flag.
The Trojan Horse in the Code: Open Source Contamination via AI#
The final AI-era ghost is perhaps the most insidious. AI code assistants are trained on vast public codebases, including projects with highly restrictive "copyleft" licenses like AGPL or GPL. These licenses have a "viral" effect: any software that incorporates code under such a license must itself be made open source.
If an AI tool suggests a code snippet derived from an AGPL-licensed project, and a developer unthinkingly incorporates it into the company's proprietary product, the company may now be legally obligated to make its entire source code publicly available. This is an existential threat. It's the equivalent of a software virus that forces a company to give away its most valuable asset for free, utterly destroying its commercial value. This risk is far more dangerous than a simple bug and requires sophisticated code scanning and developer policies to prevent.
Are You Investing in Bedrock or Quicksand?#
A startup's approach to IP assignment is more than a legal detail—it's a fundamental indicator of its operational maturity, strategic foresight, and ultimately, its resilience. The classic pitfalls can sink a deal, but the new AI-era risks can render a company's core assets worthless without anyone even realizing it.
A traditional, checklist-driven diligence process will catch some of the ghosts of the past. But it takes a founder's perspective—a systemic understanding of how technology, people, and processes truly interact—to spot the invisible threats of today and tomorrow.
These aren't just legal technicalities; they are fundamental questions about the durability of your investment and the quality of the team you're backing. If you're wondering what other systemic risks might be hiding in your deal flow, let's have a conversation.