Venture investors sign NDAs late in a deal, for a defined disclosure, and almost never to see a pitch deck; asking for one at that stage marks you as new to the process. The instinct behind the ask, protecting what you have built, is correct. The mechanism is the part that is wrong.
The economics behind the norm
The maths makes NDAs unworkable from the investor's chair. Writing on 14 February 2006, Brad Feld set out why most VCs do not sign NDAs, quoting Guy Kawasaki's point that a firm may be looking at three or four similar deals at the same time. Fred Wilson took up the same question the next day, and his readers added the blunter reason: the side with the chequebook sets the terms. Signing an NDA per deck would create a legal exposure map no fund could operate under, because any future investment in the category could be framed as a breach.
An investor who steals ideas has a bigger problem than any contract, because reputation is the enforcement mechanism in venture. Word travels fast in a small industry, and dealflow dries up for funds founders do not trust. That is a weaker guarantee than a signed document and a stronger one in practice, since a fund lives on being shown things first.
There is also the quieter reason: ideas are not the scarce asset. Execution, speed, and the specific insight you develop by building are. A fund looking at three or four similar deals has heard the idea before. What it has not seen is you executing it.
Staging what you share
The real discipline is matching the depth of what you share to the depth of the conversation:
- The deck. Your narrative, market view, and headline traction. Nothing in it should sink you if it appeared on the open internet, because decks get forwarded. Design for that reality.
- The data room. The sensitive layer: detailed financials, customer contracts, pipeline names, cohort data. It opens after real interest, in stages, during diligence.
- The crown jewels. Proprietary algorithms, unpublished IP details and trade secrets stay out of fundraising materials entirely. If deep technical diligence eventually requires them, an NDA at that stage is normal and reasonable, and serious investors sign them there without friction.
Mechanically, trackable links beat attachments: you see who opened what, you can revoke access when a process ends, and forwarding becomes visible instead of invisible.
Assume the deck circulates inside the firm and beyond it. Partners forward decks to colleagues, to founders in the portfolio, and to people they use as informal experts on your market. That is how diligence happens, and a confidentiality footer on slide one does nothing to stop it. It is also the reason the deck is the wrong place for a customer's name you have not asked permission to use.
Where an NDA is normal
The norm is narrower than "investors never sign anything". Corporate and strategic investors, whose parent company may compete with you, are a different conversation, and founders routinely ask for protections there that a financial VC would refuse. Technical diligence that involves showing source code, unpublished research or a regulated dataset comes with an agreement almost by default. So does anything covered by an existing customer contract, where the obligation is yours and the investor inherits it.
The pattern is that an NDA attaches to a specific disclosure with a defined scope, late in a process, and never to the first meeting. Asked for at that point it reads as competence. Asked for before a first call it reads as inexperience, and it costs you the call.
The signal you send instead
Walking into a first meeting without an NDA request signals that you understand the game: that you know execution is the moat, that you have staged your information sensibly, and that you are optimizing for speed of trust.
Where legal edge cases exist, patent timing questions or regulated data, route them through your counsel instead of a template NDA aimed at a VC. Where an invention is patentable, a filing made before you start pitching protects it in a way an NDA cannot, and counsel decides whether that applies to yours.
Enjoyed this article? Share it.
