The Power Elite
The power elite is Mills's name for the concentrated decision-making apparatus at the top of modern institutional life — not a conspiracy, not a ruling class in the classical sense, but an interlocking directorate whose members occupy structurally equivalent positions across the major command institutions. They attend the same schools, sit on the same boards, move between the same institutions, and recognize one another as members of the same class. Their decisions reinforce one another with the regularity of a machine whose operators have never needed to discuss its purpose. In the AI age, the tripartite structure maps with precision onto the concentrated apparatus controlling frontier model development, capital allocation, and deployment terms — a few hundred people whose decisions determine the productive lives of billions.
In the AI Story
Mills articulated the power elite thesis against two dominant alternatives in 1950s American social thought: the pluralist account, which held that power was distributed across competing interest groups, and the Marxist account, which located power in ownership of the means of production. Mills argued both were inadequate. The pluralist account missed the structural convergence of the command positions. The Marxist account missed the military and political dimensions that had become as consequential as the economic. His synthesis identified three command posts — the corporate rich, the warlords, and the political directorate — whose interlocking produced a coherent governing apparatus without requiring formal coordination.
The AI power elite reproduces this architecture with organizational precision. Fewer than a dozen organizations — OpenAI, Anthropic, Google DeepMind, Meta, xAI, and a small number of state-adjacent peers — control frontier model development. The capital requirements ensure the barrier to entry is insurmountable for any entity outside the concentrated wealth that the largest corporations, sovereign wealth funds, and wealthiest individual investors can mobilize. Within these organizations, consequential decisions are made by executive teams that could fit in a conference room.
The decisions are specific and political. Pricing determines who can afford to participate in the new economy of intelligence. Terms of service draw the boundaries of the permissible without legislative deliberation. Training data selection determines whose knowledge is encoded and whose is erased. Capability release decisions determine the boundaries of the possible for every person who depends on the tool. These are not technical decisions dressed as engineering. They are acts of governance performed by institutions accountable to shareholders but not to the populations whose lives the decisions reshape.
The response to the power elite is not moral exhortation but institutional construction. Mills was explicit: telling the elite to be more responsible is futile because the irresponsibility is structural, not personal. The response must be governance arrangements that connect the exercise of power to the experience of its consequences — a demand the AI transition has made globally urgent and institutionally unmet.
Origin
Mills published The Power Elite in 1956 after two years of research synthesizing biographical data on corporate executives, military leaders, and political officeholders. The book received the American establishment as an act of aggression. Reviewers dismissed it as conspiracy-minded; the establishment press treated it as fringe. The analysis has aged better than the reviews.
The thesis gained renewed urgency in the era of platform monopoly and has reached a new pitch with the concentration of AI capability. Contemporary readers encountering Mills for the first time often experience a specific vertigo: the sensation that a framework built for mid-century military-industrial America describes the current moment with more precision than frameworks constructed specifically for it.
Key Ideas
Structural position, not personal character. The elite is defined by the command posts its members occupy, not by the virtues or vices of the individuals occupying them. Replacing the individuals leaves the structure intact.
Interlocking, not conspiring. The coordination among command posts operates through shared background, shared interests, and institutional circulation — not through secret meetings or formal alliances.
Insulation from consequences. The elite's decisions affect populations from which the decision-makers are geographically, temporally, and epistemically separated, producing what Mills called the higher immorality.
Democratic appearance, oligarchic substance. Formal democratic institutions persist while consequential decisions migrate to command posts that democratic mechanisms cannot reach.
Debates & Critiques
Critics from the pluralist tradition — Robert Dahl most prominently — argued Mills overstated the coherence of the elite and understated the genuine conflict among its factions. Later scholarship has largely sided with Mills on the structural analysis while accepting that the elite is less monolithic than his sharpest formulations suggested. The AI case has, if anything, simplified the analytical problem: the command structure is more concentrated, the interlocking more visible, and the accountability gap more measurable than anything Mills examined.
Appears in the Orange Pill Cycle
Further reading
- C. Wright Mills, The Power Elite (Oxford University Press, 1956)
- G. William Domhoff, Who Rules America? (McGraw-Hill, 1967, revised editions through 2022)
- Kate Crawford, Atlas of AI (Yale University Press, 2021)
- Shoshana Zuboff, The Age of Surveillance Capitalism (PublicAffairs, 2019)
- Meredith Whittaker, 'The Steep Cost of Capture' (Interactions, 2021)