Getting Zucked: What Researchers Should Know Before Starting a Company

If your research works, someone is going to suggest the two of you start a company. Maybe they already have. The field is having its moment, the money has noticed, and the distance between “promising result” and “fundable company” has collapsed to roughly the length of one coffee. The person suggesting it is often charming, often sincere, and often right that there is a company in your work. The question they will never help you ask is the only one that matters: when this company exists, who owns it?

Because here is the thing nobody tells you at that coffee. In most of these pairings, you are not just being recruited. You are being acquired. Your contribution, the ideas, the code, the models, the results, already exists. It is sitting there finished, unregistered, and portable, which in legal terms means it belongs to whoever gets it assigned to the right piece of paper. The operator’s contribution is a promise about the future, and the one thing they produce on day one is the paperwork itself. Look at that asymmetry squarely: your value can be transferred with a signature, and their value is the signature. Every one of these pairings that ends badly for the researcher runs through that gap.

Researchers walk into it for a structural reason, not a stupidity reason. Our entire training points outward. Publish the method, open the code, post the preprint, because the currency of science is credit and credit only exists in public. Nothing in a PhD teaches you to think of your own work as an asset that can be held, which is why the person across the table, whose whole craft is structure, entities, ownership percentages, who signs what, is playing a game you have literally never practiced. That used to be survivable, because research became valuable slowly, over years of citation. Now it becomes valuable the moment it runs, which means the acquisition has to happen early, before you notice what you are holding. Hence the coffee.

The move, step by step#

The move has a standard sequence. I can walk you through it in detail because it was run on me, and the details are the lesson.

Step one: agree on a structure that postpones ownership. In my case the plan was a nonprofit. We would build for the mission, structure “coming soon.” A nonprofit plan is ideal for this move, but “we’ll incorporate after the demo” or “let’s not complicate things yet” work identically. The function is the same: create a period where real work is being produced and no document says who owns it. And notice what mission framing does to the room. Once the shared destination is a cause, asking “who owns what” feels like asking who gets the furniture at a wedding. Whoever raises ownership first becomes the one who made it about money. So nobody raises it. I did not raise it either. Not because I forgot, but because the mission framing worked on me exactly as designed: I liked the cause, and I liked being the one too generous to ask. And the question does not dissolve. It waits for whoever is willing to answer it alone.

Step two: create the entity unilaterally. While our nonprofit stayed a shared intention, the person I was building with registered an LLC. Sole owner: him. Nothing was ever signed between us, so in the only sense that binds we were never co-founders at all, which is exactly the gap the move lives in. He did not steal anything, did not hack anything, did not lie about anything. He simply did, by himself, the one thing we were supposed to do together, and that is the entire move. A plan is a mood; an entity is a fact. Whatever everyone meant, whatever was said at the whiteboard, the state now recognized exactly one owner of exactly one container, and the container was about to be offered as the project’s home.

Step three: ask for the IP, framed as housekeeping. Then came the request: sign the project’s IP over to the LLC. Not presented as a takeover, because it never is. Presented as tidiness. The project needs a legal home, things should be held properly, this just formalizes what we are already doing, it is being held for everyone. Every sentence in this genre contains the word “for,” and every one of them skips the same fact: held by whom. Whoever owns the container owns the contents, whatever the container was said to be for. “For the mission” is not a legal concept. Sole member of the LLC is.

Step four: the document arrives already read. The paperwork shows up drafted, reviewed, decided, needing only your signature, and this is the tell that survives every disguise. The person handing it to you has known its contents for as long as it took to prepare, which means that for exactly that long, the partnership had one informed party and one trusting one. The asymmetry of preparation is the confession. You do not need to prove intent, read motives, or win an argument. The timeline already testified.

I read the document twice, looking for the sentence that would make it innocent. I did not sign, and that outcome is the least interesting part, because by step four the damage is mostly done either way: the working relationship is gone, months of joint work sit in limbo, and your only options are surrender or conflict. The point of knowing the sequence is to break it at step one, where it costs a single awkward conversation, instead of at step four, where it costs the project.

What you already signed#

The week the document arrived, I had to answer a question I had never thought to ask: did I even own the work I was being asked to assign? It is the question that comes before “who will own this company,” and most researchers have never asked it either. In almost every case you answered it years ago, in the onboarding paperwork you signed without reading, and the answer depends entirely on which building you were sitting in.

The university deal. A PhD or a faculty position comes with an IP policy, and the core of it is roughly the same everywhere: the university claims inventions made with its resources, on its time, or under its grants, and if the work was federally funded the claim is not even the university’s choice. You cannot assign that invention to anyone’s LLC without going through the tech transfer office. What you keep is real, though: your name on the paper, the right to publish, a share of licensing revenue, and usually a sanctioned exit where the university licenses the patent to your own spinout for equity, which is how a large fraction of deep-tech startups are born. What you lose is title and control. It is not a bad deal. It is just not the deal most students believe they are on, which is “everything I make is mine.”

The company deal. Industry is sharper. An employment contract almost always assigns your inventions to the employer at the moment they exist: anything within the scope of your job, made with company resources, or touching company confidential information. The copyright in your code was never yours for an instant, and publication needs approval. What you keep: your salary and equity, which is the honest price, and your general skill and knowledge, meaning you keep what you learned but not what you saw. Some places carve out what you build entirely on your own time and equipment; many do not, and there your weekend project is arguable. What you lose is the IP itself, plus a slice of your freedom to build the adjacent thing next.

Read those two deals side by side and the lesson writes itself. When someone slides an assignment across the table, there may be two or three parties in the room you cannot see: a university, a funding agency, a former employer. Signing anyway does not transfer the IP, because you cannot transfer what you do not hold. It transfers the problem, and puts your signature on it.

The rules that break the sequence#

Have the ownership conversation on day one, while it still feels rude. The awkwardness is the test working. Ask directly: who owns what, in what split, and what happens to the work if we part ways? A good-faith partner is not damaged by this question, because good faith has nothing to lose from clarity. The only person the conversation burns is the one whose plan required it never happening. Their reaction is worth more than any reference check.

Match the entity to the promise, and form it together. If the plan is a nonprofit, the first legal object created is the nonprofit, or nothing. The moment a different structure appears, especially with a sole owner, the plan has already changed regardless of what anyone says. Do not evaluate the words; evaluate what got filed, by whom, with whose name on it. If you were not in the room when the company was created, it is not your company. That sentence sounds harsh and is merely descriptive.

Never assign IP to an entity you do not co-own. Assignment is not sharing and it is not licensing. It is a one-way door, and “we’ll sort the equity later” on the far side of it is a negotiation you enter holding nothing. If an entity must hold the work, become an owner first, in writing, in the split you agreed. Later is the most expensive word in this genre.

Divide the labor, never the literacy. “You do the science, I’ll handle the business” is a fine arrangement, plenty of great companies run on it, but the business side is precisely where ownership lives. Delegate the work freely. The day you delegate the understanding, you have handed the other party the pen that signs for both of you.

Check the invisible parties before you promise anything. You now know who else may hold claims on your work. A partner who urges speed over checking them is urging it for a reason, theirs or plain carelessness, and both are disqualifying.

None of this argues for trusting less. Research runs on exactly the openness that makes it takeable, and armoring yourself into paranoia would cost more than any LLC ever could. The argument is narrower: trust should be checkable, and the beginning, when everyone still likes each other and paper feels unnecessary, is the only cheap time to make it so. A partner worth having meets the ownership question with relief, because they wanted it settled too. The one who bristles at daylight has just told you, early and for free, everything the paperwork would have told you later at full price.

I think sometimes about that first coffee. Everything that happened later was already present in the questions we were both too polite to ask at it.

← all writeups