On the one hand it is ideal for an organization to be auto-catalyzing, and hopefully a private ARPA would create massive value. On the other hand It is hard for a private ARPA to capture value without hamstringing impact and pressure to capture value or create measurable value too quickly could constrain the organization out of working on things whose impact depends on the results being public or have long time-horizons. There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators
At the same time, I suspect that The market in for-profit organizations is fairly efficient (at least as of 2020.)
There are two key constraints on for-profits: first, they need either generate cash or be a good investment and second, they need to be able to capture the value they create. The whole goal of a private ARPA is to work on things that aren’t product-ready so self-catalyzing through cash generation from day one is not an option. The tricky thing about investment quality is that it is judged relative to other possible investments. So even if an investment could possibly generate a positive 7% per year over a long time-period, if a lower-risk investment could generate that same (or better!) return or return it sooner it just makes sense as an investor to go with the latter.
For-profits need to capture enough of the value they create to generate those investable returns. Many companies are able to create massive amounts of value and capture a small but still massive amount of that value. This is great! People should do this whenever possible. However, it is incorrect to assume that there is anywhere near a direct relationship between creating massive value and capturing massive value.^1Current value capture mechanisms are crude] - you either need to be able to package what you’ve created into a product and sell it or command enough attention that you can sell ads, seminars, etc. Perhaps even more significant than value-capture issues, there is a whole class of innovations whose impact (and value) depends on their results being public. There is a significant class of innovations that would create drastically less value for the world if their value had been captured by their creators. An organization that isn’t working on a well-defined product or a working on something whose value depends on public access is just a bad investment. ^2 I suspect that unlike the marginal good investment that isn’t supported because it looks like a bad investment, valuable programs that are just bad investments that aren’t happening are profligate.
So while ideally an organization with long-term aspirations would be for-profit, the nature of the programs a private ARPA should support impose constraints that make a for-profit structure unlikely.
^1:The most extreme version of assuming a direct relationship between creating and capturing value is that anything that creates value will be able to capture it and therefore anything that can’t generate enough money to sustain itself is obviously not that valuable.
^2: By “bad investment” I specifically mean “bad monetary investment.” However, I think that attempts to measure non-monetary ROI are extremely sketchy and at the end of the day professional investors are measured in dollars.