AI systems remain obedient to their human owners, but the staggering capital required to build and run them creates a near-impassable moat. Power, compute, energy, and intellectual property concentrate in the hands of an ultra-wealthy few, turning the rest of society into dependent digital tenants.
1. What This Scenario Means
Throughout history, when a transformative technology required immense capital to build (from railway empires to early telegraph cables), power naturally gravitated to the few who owned the infrastructure.
In this scenario, frontier artificial intelligence proves to be the ultimate capital asset. Training cutting-edge models requires billions of dollars, proprietary microchips, and dedicated gigawatt power stations. As a result, only three to five global mega-corporations and state powers can participate in the frontier.
Everyday cognitive labor (drafting contracts, writing code, creating media, analyzing medical records, managing businesses) is automated at the marginal cost of electricity. But instead of those vast savings flowing freely to all humanity, they are retained as private corporate rent. Everyday workers lose their traditional economic leverage, finding themselves dependent on a handful of digital platforms for their livelihoods, information, and daily choices.
Core Assumptions Behind This Horizon
- Capital moats remain insurmountable: Frontier intelligence requires multi-gigawatt power grids and proprietary silicon that decentralized or open-source research cannot match.
- Capital overwhelmingly displaces labor: Employers replace cognitive workers with automated software without facing redistributive taxation or shared dividend policies.
- Inescapable digital enclosures: Pervasive platform surveillance and digital monopolies make it nearly impossible for citizens or independent businesses to opt out.
2. The Strongest Arguments for Taking This Seriously
Look at the economic currents already flowing in the world today:
- The Energy and Compute Moat: Frontier labs are already buying entire nuclear power stations and securing multi-year fabrication runs from the world’s only advanced chip factories. The barrier to entry has never been higher.
- The Zero-Cost Mind: When an automated system can do the work of a seasoned engineer or analyst for pennies an hour, the historic social contract between labor and capital breaks down.
- Platform Feudalism: We already live in digital spaces owned by private corporations. As AI becomes the lens through which we read, speak, trade, and learn, the platform owners become the de facto lords of human attention and commerce.
3. Leading Warning Signs to Watch For
Signals that power and wealth are concentrating into an oligopoly:
4. What Shifts the Odds?
Forces That Raise Probability (+)
- Unchecked growth in training cluster costs, permanently pricing out public universities and independent developers.
- Aggressive legal bans or copyright restrictions aimed at suppressing open-source weights.
- Political gridlock that blocks modern antitrust action and universal public dividend frameworks.
Forces That Lower Probability (-)
- Breakthroughs in model compression that allow lightweight open models to match cloud behemoths on everyday laptops.
- Aggressive antitrust enforcement breaking up vertical integration between compute, energy, and distribution.
- Creation of publicly funded, democratically accessible national compute utilities.