Andrew Jones

Large firms exist because of market friction. If you want to be well known and easily trusted, size has traditionally been the answer. The downside of scale is greater coordination friction; larger firms become more inefficient. This is the basic economic principle laid out by Ronald Coase in his 1937 essay, The Nature of the Firm. For a long time, it remained unchanged. And then, in the 1990s, the IT revolution kicked it up a gear. As computers greatly reduced market frictions in some areas, the optimal firm size fell considerably. However, in other categories they reduced internal coordination frictions, enabling much larger firms. Essentially, it was an economy-wide barbell effect, as firm scale became more concentrated at either end of the distribution, with fewer medium-sized firms. It is likely that AI takes this even further, pushing firms much further out on the distribution. This is especially true for venture capital, where the market is already divisible into: - A scaled service-oriented industry with an elastic product (private market beta / solving puzzles) - A performance-driven cottage industry with an inelastic product (venture alpha / solving mysteries) In the former, the scaled firms will get larger in AUM and headcount as AI reduces coordination friction by handling more of the investment process. In the latter, the ideal size of a firm will fall to one, as AI eliminates external market friction by quantifying reality and making that more accessible. A simple way to frame this: The best venture investors are individuals, not firms. Peter Thiel, Cyan Banister, Oren Zeev, Charlie Songhurst, Elad Gil, Chris Sacca... They might also have firms, but their success is their own. On the other hand, a sovereign wealth fund has no real interest allocating to individuals, because their priority is scale and (lower) risk, not performance. This is the role that firms like Andreessen Horowitz play. One category of LP wants banging DPI. The other wants name brand firms, stable IRR and a five-star investor relations team. This is where the market splits in two, and why AI will push participants further apart on the distribution as market frictions are reduced.