IP Due Diligence
Diligence investors actually rely on.
The IP schedule in a data room tells you what was filed. It does not tell you what holds up. For frontier technology investments, the difference between the two is where deals are made, delayed, or lost.
We are retained by venture capital firms, growth equity investors, and strategic acquirers to evaluate what the schedule does not say: whether claims are defensible against the prior art the seller did not disclose, whether the chain of title traces cleanly to enforceable ownership, whether the portfolio actually tracks the product the buyer is paying for, and whether the trade secrets and data assets that complete the picture will still be assets the day after closing.
The substantive questions depend on technical understanding that cannot be outsourced to a checklist. We have conducted diligence on portfolios later pledged as collateral in IP-backed financing, on foundational cryptographic patents that were subsequently acquired, and on companies through the full arc from formation to initial public offering. The diligence we produce is calibrated for counterparties who will scrutinize it.
Scoping
Our engagements are scoped to the stage of the transaction. Investors at different stages are asking different questions, and the diligence they need reflects that.
Seed IP Review. At seed and pre-seed, the question is binary: are there fundamental IP defects that would make a later exit impossible? We confirm that the founders own what they claim to own, that prior employment obligations have been addressed, that any patents are defensible at the fundamentals, and that no obvious red flag disqualifies the opportunity. The deliverable is a concise executive memo identifying material issues and recommended next steps. This engagement is appropriate when deal velocity is high and the decision is whether to proceed, not whether to price a premium.
Series A Assessment. At Series A, the diligence informs valuation and deal structuring. We conduct a full inventory of the IP assets and the contracts that govern them: patents, pending applications, trade secrets, proprietary data, and inbound and outbound licenses. We trace the assignment chain from original creators through the company. We evaluate patent claim strength and prior art exposure with attention to the prior art the examiner did not reach. We review training data licensing and provenance. We assess trade secret identification and the practices that actually preserve that treatment. Deliverables include a detailed report, executive summary, risk matrix with materiality ratings, and a remediation roadmap.
Growth Stage and M&A. For Series B and later investments, growth equity transactions, and acquisitions, our diligence extends to the full scope of what a sophisticated acquirer or investor expects. In addition to the Series A work, we conduct freedom-to-operate analysis, evaluate litigation risk, review existing license agreements for change-of-control and other material provisions, conduct management interviews on IP strategy and protection practices, and position the portfolio relative to the competitive landscape. The deliverable is a comprehensive report suitable for investment committee presentation, with follow-up support through closing.
What we examine
Patent portfolio quality. Patent counts are the least useful measure of a portfolio's value. What matters is whether the claims cover what the product actually does, whether they are defensible against the prior art that matters (not only the prior art the examiner found), and whether the portfolio is structured for what the company needs to do next. For AI inventions, this includes assessment of Section 101 eligibility risk that in-house teams often underestimate until a portfolio is challenged. For cryptography, it includes the specific 101 and 112 posture of implementations of mathematical constructions, an area where generalist analysis produces misleading results.
IP ownership chain. A portfolio is worth what its owner can actually enforce. We trace ownership from original invention through every assignment, employment agreement, and corporate transaction that touches the asset. Defects that surface here (missing executions, assignments to entities that no longer exist, inventions made before assignment obligations attached) can render otherwise valuable patents unenforceable, unlicensable, or uninsurable. For companies that grew through acquisition or through multiple corporate reorganizations, this trace frequently produces surprises that change the structure of the deal.
Training data and proprietary data. For AI companies, the most valuable asset is often the data, and the hardest question is whether the company has the rights it needs to the data it has. We review acquisition practices, licensing documentation, terms-of-service compliance, and the exposure to copyright, privacy, and contractual claims that attach to data of uncertain provenance. Where documentation is thin, we assess the nature and magnitude of the risk in terms an investment committee can act on rather than abstractions that require further interpretation.
Trade secret protection. In frontier technology, trade secrets protect what patents cannot: model architectures, training methodologies, unpublished algorithms, and security processes. We assess whether the company has adequately identified what it treats as a trade secret, whether it has implemented the practices required to preserve that treatment, and whether the protection will survive a change in ownership or a departing employee. A trade secret that was never documented as a trade secret is rarely a trade secret at all.
Why this work is different in our domains
IP diligence looks different in frontier technology than in mature industries. The risks that matter, the documents that contain them, and the judgment required to separate signal from noise vary across the domains where we work.
Artificial intelligence. Training data rights and provenance are the single largest source of undisclosed risk, and the one most commonly missed by diligence that was not designed for AI. Model weights and derivative-model licensing raise questions that most diligence checklists do not contain. Section 101 eligibility is a persistent question for machine learning inventions, and a portfolio that has not been stress-tested against it may not be ready for the scrutiny of an acquirer.
Post-quantum cryptography. Patent positioning around standardization is the defining question, with material implications for both valuation and freedom to commercialize. Claims on mathematical constructions and claims on their implementations carry different risks and different enforceability profiles. Export controls complicate both transfer and use. The publication histories of the inventors often matter more than the public record suggests.
Cybersecurity. The tension between trade-secret protection and patent disclosure is substantive. A cybersecurity company's most valuable asset may be the one it has never described in a patent or a public filing, and the diligence question is whether that asset is protected, documented, and transferable to a new owner without losing its value.
Data-driven systems. Data licensing as distinct from technology licensing, with its own vocabulary of field restrictions, derivative-data rights, and aggregation rights. Cross-border data transfer obligations baked into the underlying contracts. The allocation of liability for data-quality and data-lineage claims that, if not understood before closing, produce disputes after it.