The goal of the vision fund is to make companies a monopoly by selling below cost (using the vc money), and waiting for the rival companies to flee the market. It does not really care about the specific area.

This strategy will backfire in a fragmented market. I.e. if the competitor hold on.

Yeah, I get that and my post was kind of an implicit difference between what they "are" versus what I think they "ought to be" (admittedly easier to be in my armchair investor seat than Softbank's).

But their returns from Uber, WeWork, and Wag kind of support my point, whereas ARM and Nvidia support yours/theirs.

My broader point is that most informed technologists would've said three years ago that the ARM and Nvidia investments made more long-term sense than Uber and WeWork, yet Uber and WeWork are a disproportionate amount of the portfolio.

So if the Vision Fund isn't even going to succeed at the "scale big companies globally" approach, I'm not sure they would've done much worse going with the approach I suggested (or doing more deals like ARM and Nvidia, maybe they could've taken a huge activist stake in IBM or other large existing tech cos).

I’m unsure the point of the vision fund was to make safe long term bets. At the end of the day, it’s unclear to me how nvidia would use an extra 10B dollars - other than pour that into R&D (in which case does that really cost $10B - and if your thesis is R&D, universities tend to do a better job anyways).

The thesis for Uber (and wework) was simpler - assuming they had sound unit economics, give them $XB to repeat all over the world. In fact if Uber didn’t mistakenly give Lyft a second life, Uber might be in a much better spot.

I think the problem is, companies like Uber, Wework and Wag work really well in wealthy cities like SF and NYC, but not as well in other cities. I live in SF and WeWork “works” there are glut of companies coming in and out of the VC machine that need temporary space that end up needing to also scale up quickly. However that’s unique to SF and SF is such a large market, maybe it looks sound until it doesn’t.

University R&D IP tends to be owned by universities that it's developed at - Google's PageRank patent was owned by Stanford. Investing directly in universities won't have the desired payout.