Friday, December 8, 2023
Benjamin posted an excellent discussion of OpenAI a few days ago and I am just here to piggy-back a little bit. Like me, he recognizes and explains why nonprofit joint ventures involving profit takers is primarily a problem of private benefit rather than private inurement. Private benefit is typically manifested by consistently paying too much and private inurement is paying too much to fiduciaries even just once. Both are concerned about the diversion of charitable money to personal consumption by noncharitable beneficiaries. Control, as Ben noted, is key for private benefit. Control by those without a personal financial stake in the joint venture, presumably those appointed by the charity. If Microsoft controls the board that controls the joint venture, then the nonprofit's participation in the joint venture is deemed to generate private benefit. If OpenAI (the 501(c)(3), I mean) controls board that controls the joint venture than private benefit is assumed away according to the guidance Ben discusses.
Private inurement is when the insiders pay themselves too much of the charity's assets. It might also occur when the insiders take a cut of the 501(c)(3)'s profit -- revenue sharing, a topic that still has not been regulated upon in the excess benefit regulations. Private inurement is always private benefit but private benefit isn't always private inurement. We learned that rather painfully after about 10 years of litigation in United Cancer Council. At first, I was sure OpenAI's joint venture with Microsoft (I am going to call it a two-entity partnership just for simplicity) was about private benefit of the non-inurement variety. But then OpenAI insiders all became employees of the joint venture, with profit sharing rights. Profits derived from OpenAI's participation in the "whole hospital" joint venture.
Two things happened when insiders too stock in the joint venture, it seems to me. The underlying assumption that OpenAI's control over the governing authority ensures charity takes precedence over profit is no longer rational. The insiders are no longer mere altruistic managers whose loyalty may be assumed. They might still be altruistic somehow, but they are also profit taking from the charitable activity via an intermediary. Because insiders have stock and there is an outstanding tender offer expected to bring in $86 billion if profit-takers, including the insiders of OpenAI, sell. Charity might be in control of the joint venture in name only. But the second problem is private inurement. By an indirect route through Microsoft (or some other entity to collect Microsoft's profit from the joint venture), insiders get a cut of profit from OpenAI (the nonprofit, whose assets are entirely at the joint venture's disposal). There is a good case that Microsoft's coup d'etat resulting in Altman's reinstatement proves private benefit. But even if not, the insider's profit taking via an intermediary of the sort described in Treas. Reg. 53.4958-4 seems a huge problem.
darryll k. jones