The deal died over a patent assignment.
The patents had moved from the inventor to the university. They had never moved from the inventor and the university to the biotech whose program was about to be licensed. There were gaps at three links: inventor to biotech, inventor to university, university to biotech. Swiss cheese, except that in the data room it did not look like Swiss cheese. It did not look like anything. Everyone stayed caught up in the excitement of a deal that size, straight through the preliminary discussions, and nobody went back to read the assignments.
The company had raised about $40 million from venture investors. A large pharmaceutical company was ready to put in $15 million, at a price that valued that one program at $196 million. Everyone was already doing the next sum. With a big partner's money and name behind it, the whole company might be worth something like $750 million by the time the partner came back to buy it, likely within two years. That number was a hope, not a price. It never got tested. The deal did not close.
Everyone priced the upside, and nobody set a standard for the footnote.
I have seen a version of this at least half a dozen times. I leave the companies and compounds out on purpose. The pattern is the useful part.
Why biotech deals die this way
A license is the first half of a purchase.
Most biotech acquisitions begin as a license: a specific compound, for a specific use, in a specific territory and time, with milestone payments tied to clinical data. If the larger company likes what it licensed and helped develop, it bids for the asset, or it buys the whole biotech and shelves the programs it does not need.
So whatever is wrong with the licensor's rights travels the entire path from license to sale. A broken chain of title does not stay in a footnote. It surfaces on the day the buyer's lawyers read it, which is the worst possible day.
Two lenses
Counsel reads the document and asks what could be contested later. An investor reads the market and asks what the asset is worth. Each question is incomplete alone. The $196 million partnership had the second in abundance.
I have watched structure decide outcomes for twenty-five years. One merger closed fast because the diligence was tight and the records were spotless. Another had perfect science and disciplined investors, but early agreements carried restrictions nobody anticipated, and governance fights consumed the energy that should have built product. The difference was whether anyone had built a foundation that could hold weight.
Early in my career I sat in one of those rooms. The company had two cancer programs. One was a chemotherapy drug wrapped in a three-part peptide coat, built to shed one piece near the tumor, one at the cell surface and the last inside the cell, where it released the drug. Taking it into larger trials meant making far more of it. The head of manufacturing held up a Pringles can to explain why that was hard: this is not dropping three tennis balls into a tube.
I asked where it was made. A contract lab in Europe. Then I asked who owned the process. He didn't know. The lab did. We had to license the know-how to make our own drug, the program slid behind the company's lead product, and a two-program company became, in practice, a one-program company. Nobody had asked who owned what.
Set the standard before the debate
The fix is plain and it works. Decide in writing how much proof each level of commitment requires, before anyone is excited.
Legal reasoning already has the ladder. Probable cause justifies an investigation and no check. A preponderance of the evidence justifies a sizable allocation. Proof beyond a reasonable doubt, with the downside bounded, justifies a concentrated bet.
Chain of title sits on the second rung. A clean assignment from every inventor and every institution to the company is the thing being bought. If it is not in hand, the case stays at probable cause, whatever number is on the term sheet. Set the thresholds in advance and the goalposts cannot slide when the room falls in love with the deal.
Assertions without artifacts are expensive opinions.
For each claim I want the signed document: the assignment, the cap table snapshot, the regulatory correspondence. Founder assurances and verbal commitments do not count. Memories fade. Documents do not, though lawyers misread them often enough.
Four things every case must prove
Every legal claim needs its elements, and so does an investment.
Market. Will customers pay at the scale required? Usage numbers and expressed interest are not evidence. Signed contracts, pilot results and retention data are. In life sciences this is the reimbursement question.
Technology. Does it work at the cost and scale the business needs? That takes expert assessment. Sit with the scientists and learn what the data shows, not what the slide deck says it shows.
Execution. Will this team hold up under the challenges this company will actually face? Look at how leadership handled earlier setbacks and whether the board has the independence to step in. In enterprise software, execution is mostly a sales question. In regulated life sciences it is clinical protocols, regulators and a founding scientist whose identity is fused with the science.
Economics. Do the returns survive the capital required to reach them? Plenty of companies work and still fail their investors here.
Weak evidence on any one element should end the case. Close the gap or do not invest.
The call
The $196 million partnership had a strong team, sound science and a real market. None of that was in the footnote, and the footnote is where the deal died. Reading the documents is slower than believing the story. It is also the only reading that survives a buyer's diligence.
Protect the IP from day one. Keep the cap table clean. Ask how every protective provision will read under stress a decade from now.
Before the next exciting meeting, ask one question out loud: who owns what? Then ask who has read the chain of title, and who will put their name on it.