By Gad Weiss.
In late 2024, OpenAI announced plans to overhaul its unusual “capped-profit” structure. The changes would have affected the governance features that had distinguished the market-leading AI startup from Silicon Valley’s common practice: notably, being completely controlled by its nonprofit parent entity, and placing caps on investor and employee returns. Together with its main rival Anthropic, OpenAI had been a leading example of what the paper terms Aligned Structuring: business designs that distribute control and cash flows among a startup’s entrepreneurs, investors, and stakeholders in ways meant to ensure that public interest is given priority over profits. The paper examines how policymakers should evaluate these structures – notably, when, if ever, they should block efforts to roll them back, as they were urged to do in OpenAI’s case? It argues that the significance of Aligned Structuring is largely overstated. Outsider decision-makers tasked with advancing the public interest often lack the tools or influence to meaningfully constrain the startup’s profit-seeking insiders. Profit caps are easy to game or bypass, particularly when equity holders have strategic (i.e., non-financial) investment motives or can benefit financially by indirect means. Many of the structural benefits ascribed to alternative business entities are already largely available within traditional corporation-based structures. Meanwhile, misaligned business design may effectively repel the capital and talent needed for safe and ethical innovation, to society’s detriment. Policymakers must look past formal structures and assess how power is actually distributed before endorsing a business structuring model that is likely to deliver less than it promises.