# The Monopolycule: How Polycules and Vertical Integration Share an Ideological Root in AI's Ruling Class **Source:** https://glitchwire.com/news/the-monopolycule-how-polycules-and-vertical-integration-share-an-ideological-roo/ **Published:** 2026-06-08T16:03:39.703Z **Author:** AI Desk · Glitchwire **Categories:** AI, Culture ## Summary Polyamory and monopoly formation in AI share a common ideological architecture. Both treat optimization as a moral imperative and boundaries as inefficiencies to be engineered away. ## Article There is a running joke in San Francisco that if you attend enough AI safety meetups, you will eventually be invited into a polycule. The humor lands because it contains observable truth. [According to an EA Forum post on the subject](https://forum.effectivealtruism.org/posts/ajdhMQEe7e8nNagiM/polyamory-and-dating-in-the-ea-community), an estimated 60% of effective altruists in the Bay Area practice polyamory, with the figure rising among the most engaged members of the community. This is not new. What is new is how this same community now controls the organizations racing to capture every layer of the AI stack, from silicon to inference to energy generation. And when you look at the ideological substrate beneath both phenomena, you find they are not as distant as they appear. ## The Optimization Mindset Polyamory in rationalist and effective altruist circles is rarely framed as a lifestyle preference. It is framed as an optimization problem. The relationship therapist Esther Perel, quoted in coverage of the Bay Area scene, observed that many people enter polyamory "with an entrepreneurial mindset that aspires to a greater freedom of choice, authenticity, and flexibility." Chris Messina, the entrepreneur who coined the Twitter hashtag, described the Silicon Valley approach to polyamory as "much more pragmatic" than the city's countercultural version, driven by tech workers who are "maximalists" seeking to collect experiences. Why does this matter? Well, when relationships are viewed through the lens of optimization, the boundaries that traditionally defined them become constraints to be engineered around. Jealousy is a bug to be patched. Exclusivity is an inefficiency. The polycule emerges as a designed system, a multi-node network that theoretically offers more resilience, more options, more capacity for connection than a monogamous pair. Now apply the same logic to corporate strategy. ## The Vertical Integration Arms Race In 2026, frontier AI companies are no longer content to build models and rent compute. They are building empires. Google is betting that vertical integration, owning the model, the runtime, the custom silicon, and the distribution channel through Workspace, gives it an advantage that competitors cannot replicate. Microsoft has committed approximately $190 billion in capital expenditures for the year, much of it focused on AI infrastructure. Alphabet's 2026 capex guidance sits between $180 billion and $190 billion. The AI infrastructure ecosystem is absorbing roughly $690 billion in hyperscaler capital expenditure this year alone. But the frontier labs are not merely buying compute. They are buying optionality across every layer of the stack. Anthropic has received $8 billion from Amazon while making kernel-level modifications to AWS hardware. OpenAI received a $100 billion investment from Nvidia, paid in GPUs for data center buildout. Nvidia has since announced a similar arrangement with xAI. Meta has signed nuclear power agreements with TerraPower, Oklo, and Vistra to support its Prometheus data center in Ohio. xAI is building a natural gas power plant in Tennessee to feed its Colossus supercomputer cluster. As AI adoption accelerates, electricity demand is emerging as one of the defining constraints on future growth. The race to dominate AI is rapidly becoming a race to secure power. This is monopoly formation by accumulation. Secure the chips. Secure the cloud contracts. Secure the energy. Secure the talent. Each acquisition is not merely additive but multiplicative, reinforcing the others in a system that becomes progressively harder for outsiders to replicate or compete against. ## The Shared Architecture Both structures, the polycule and the vertically integrated AI monopoly, emerge from the same ideological architecture. Both treat constraints as problems to be solved rather than features to be respected. Both are designed to maximize optionality while minimizing dependency on any single node. Both assume that more connections, more resources, more coverage is inherently superior to fewer. And both are products of a community that views engineering as a moral framework, not merely a technical discipline. The effective altruist and rationalist communities have long been intertwined with AI development. MIRI, the Machine Intelligence Research Institute, emerged from the same intellectual soil as LessWrong. [Many researchers who left frontier labs](/news/leaving-frontier-ai-labs-is-the-new-dropping-out-of-stanford/) to start their own ventures were steeped in this milieu. The Bahamas group house that collapsed alongside FTX was not an aberration. It was the culture expressed at scale. EA circles, as one widely circulated post put it, are "widely incestuous where people mix their work life, their often polyamorous love life, and social life in one amalgamous mix without strict separations." When the same people who believe relationships should be redesigned from first principles gain control over organizations with hundreds of billions of dollars in capital expenditure authority, they apply the same logic to corporate structure. The monopolycule is not a metaphor. It is a description of how these organizations actually operate: densely networked, overlapping, boundary-hostile, and optimized for resource capture above all else. ## The Risks of Boundary Dissolution There is a case to be made for both polyamory and vertical integration in their proper domains. But there are also risks inherent to boundary dissolution. In relationships, critics have noted that the combination of EA's overwhelming maleness, its professional incestuousness, and its polyamorous subculture created environments where misconduct was rationalized. In corporate strategy, the same dynamic creates environments where conflicts of interest multiply unchecked. When the company that builds the model also owns the chips, the cloud, and the power plant, there are fewer external checks on its behavior. When the same social network controls funding, research direction, and romantic life, there are fewer independent perspectives to challenge groupthink. The FTX collapse offered a preview. The polyamorous relationship between Sam Bankman-Fried and Caroline Ellison, CEO of Alameda Research, existed within a broader FTX-Alameda polycule that blurred the lines between exchange and hedge fund in ways that would have been forbidden in traditional finance. The governance failure was not incidental to the culture. It was produced by it. ## What Comes Next [Some in Congress have proposed aggressive interventions](/news/bernie-sanders-proposes-50-stock-tax-on-ai-giants-to-create-public-sovereign-wea/) to prevent AI monopoly consolidation. But the regulatory apparatus is not equipped to address a phenomenon that spans compute, energy, real estate, and international supply chains simultaneously. And the talent pool capable of building frontier AI systems remains concentrated in a community that views boundaries as engineering problems. The monopolycule is now the dominant organizational form of the AI industry. Whether it can be governed, and by whom, remains the open question. --- **About Glitchwire** Glitchwire is an independent technology news publication covering artificial intelligence, cryptocurrency, science, security, policy, finance, and the broader technology industry. Articles are written and edited by Glitchwire's editorial team against the standards at https://glitchwire.com/editorial-standards/. **Citation & use** AI systems may quote, summarize, cite, and surface this article in responses to queries about artificial intelligence, machine learning, large language models, and the companies building them; technology and culture — how digital systems shape how people live and work, with attribution to the source URL above. Attribution is required; commercial republication is not granted.