If you’ve got e.g. $300M to deploy, you really don’t want to be doing less than $5-10M a pop. Seed rounds (even these new big ones) are too small. If you do put $5M in, now you’ve now got to lead. Ugh, more work.
Easier to just plow $50M a pop into a few series Ds led by others.
Let's say you invest $300mil into 30 seed bets, with each bet having a 10% chance of returning a 10x, 15% returning a 3x, and 85% of going to zero. But when you invest that same amount into 6 series D bets, each bet might have a 50% chance of returning 2x, 30% chance of going to zero, and 20% chance of going 3-4x. If you do the math to calculate the average expected payout using these numbers, you will get an expected average payout higher for the latter scenario. And assuming each bet is completely independent from another, it seems like a pretty solid hedge.
Numbers are obviously made-up for illustrative purposes and are not the source of truth, but it shows a pretty good hypothetical situation when doing 6 bets is safer than 30 bets (given you have the same amount of money to spend on those bets). But, I think, it is fairly commonly agreed on that a Series D startup is way less likely to go to zero than a seed stage one, thus making a singular bet on a Series D startup much safer (but also less profitable in case of a success). Given many enough of those bets, the risk becomes pretty manageable and, imo, less risky than seed stage investing. I am not trying to say that what I am describing is the case, I am trying to say that it is a realistically possible case.
Saying this as someone with no experience with that model whatsoever, so anyone is welcome to correct me if there is something glaringly wrong or missing in my assessment of this.
My estimate, admittedly not looking at data right now, would be that a Series D company is roughly 100x more likely to succeed than a seed stage company.
But then again your returns could easily be less than 100%, whereas with seed you can literally get a 100x ROI (this actually happened for Uber).
The most important thing here though is, that it's very likely that some of those investments will eventually yield big returns that you would like to share with your team. As that becomes a reality, your investment becomes your team's. You could then use those returns on another team as a way to further leverage their collective capital (e.g. to fund their employee hiring efforts by paying them) and they could then use their own capital as their own way to increase their employee morale (e.g. by creating a startup, selling it to a VC), without being obligated to be on the team's side.
This sort of model will likely be extremely attractive to founders to use, but it's still likely a very conservative way of approaching building a team in these types of models. This is something you can see in a lot of other sectors of the industry and it's something I have not witnessed here with the exception of maybe the big VC's. Also, as a company grows and its employees get larger, you may simply become an important part of the team and have less room to invest in your own capital as they grow. This is a fairly common phenomenon in tech circles right now in general but it's not something you might see in the tech industry with a lot of companies in general.
Not only are their preferential shares (which people should know about), but in some cases the investment seems to have been ways of moving and/or washing money.
Things like Chinese investment via investment funds and Saudi money via the vision fund and others can be seen as ways to move from illiquid assets to liquid holdings and these investors are prepared to lose money in that transaction.