Someone we know got access to a huge late-stage AI round. Massive allocation and a founder whose name you probably know.

Our friend offered us an allocation. We took a look because a lot of angels we know want late-stage exposure alongside the early-stage checks they're already writing.

Our first move was an email with five questions about the structure.

1. What data do investors actually get?

We’re being asked to make an investment decision. On what, exactly?

Sometimes you get revenue, growth, and a real data room. Sometimes you get a one-paragraph blurb and a famous name. Those are wildly different data sets, and you should know what you’re getting before you commit.

2. What happens if the vehicle misses its minimum?

The vehicle needed to raise a minimum of $50M(!) in order to go through. So what if the group our friend recruits only pulls together $25M?

Does the deal still go through at half size? Does 100% of everyone's money come back?

3. If the group above them misses their allocation, does our access evaporate?

Our friend had allocation through a syndicate, which got its allocation from the bank leading the round. If that syndicate doesn't fill its number, does everything below it collapse?

Your funds may sit in limbo for six months and then come back. That lag is not ideal.

4. What's the timeline?

When does the money get called, and when does this actually close?

Knowing you're wiring in three weeks versus three months changes what else you can do with that cash in the meantime.

5. Is the fee negotiable?

Oftentimes syndicate leads will reduce carry for checks over a certain size. Worst case, they say “no.” So we asked. If we decided to go for it, we felt like we could bring enough to the table to get access to lower fees.

Getting into the deal feels like the win. It isn't.

Plenty of people have gotten into great companies through terrible structures and watched the returns get eaten by three layers of carry and insane management fees.

The company is one variable. The vehicle is another one, and it's critical to evaluate both before you wire.