AI and Technofeudalism
Much to the public's dismay, artificial intelligence isn't developing the free market: it's tearing it down. States are increasingly using AI companies as geopolitical tools as they dominate more of the market. This dependence then strengthens the companies further, creating a feedback loop of a technofeudal equilibrium where power is consolidated among a highly exclusive group of firms to the detriment of the rest of the market, destroying market balance in the process. The irony lies in that although AI is a modern technology, it's reinforcing a centuries old playbook of feudalism, only this time in the digital space.
Economist Yanis Varoufakis defines technofeudalism as a pattern of political and social dominance among major technology companies, where markets reflect the roles of historical fiefdoms and profit comes in the form of cloud rent that must be paid for access. Companies are cloudalists collecting data while the general public are unwilling cloud serfs who give up personal data. Essentially, technofeudalism is when power is concentrated among a few powerful firms that control the technology infrastructure by exchanging services for data much like how historical feudalism allowed a small group of nobles to dominate the hierarchy by exchanging land for labor.
There are three main features that define technofeudalism. First, it assumes the weakening of open-market competition since a small number of firms dominate the market. Second, it establishes rent in the economy. Third, it instigates forms of economic and political inferiority because many individuals become dependent on other firms.
Technofeudalism can be expected to be even more prevalent in the age of AI. What originally was just major technology companies consolidating power and profiting from controlling cloud rent becomes AI companies taking on this role at an exponential rate.
AI, however, also acts as a catalyst pushing this phenomenon further, since AI optimizes data collection on a more considerable scale while also acting as the commoditized good itself. The companies behind generative AI models train their models on vast amounts of publicly generated data, yet the individuals whose data powers them receive neither compensation nor meaningful say in how their data is used. Meta's Llama 3 model was trained on 15 trillion tokens from "publicly available" sources, and AI lab Anthropic agreed to a $1.5 billion settlement with victimized authors, displaying data capture from unwilling sources.
Fundamentally, AI transforms the traditional roles within economics since it can act as the product, the producer, and the platform through which value is produced and distributed. For this reason, firms that utilize AI will find themselves controlling entire value chains all within one system. Just recently, SpaceX acquired coding software company Cursor for $60 billion to control the entire vertical stack.
This framework matters for AI because it helps explain why AI may strengthen this concentrated market instead of opening newer markets. If AI companies control the infrastructure and platforms then they are shaping the conditions of the market itself. Therefore, even when numerous startups seem to take part in AI innovation, they are actually dependent on larger companies. The surface looks competitive but remains hierarchical inside. Out of 50,000 American businesses, 34.4% use Anthropic's models, while 32.3% use OpenAI's models. This concentration of corporate AI usage illustrates the monopolistic system of AI incorporation into businesses.
The role of the state is important because this outcome is not inevitable. In the past, major technological industries like railroads and electricity became essential infrastructures that were controlled by powerful private groups. Soon, states intervened and regulated the industry by treating them as infrastructures with public significance. AI requires a similar political response.
Today, when governments invest in large AI corporations in the name of geopolitical competition and innovation, they reinforce the concentration of power that makes technofeudalism possible. Instead of diffusing the technological power across multiple entities, the state deepens the dependence on a few firms that already control the market. In that case, AI becomes part of a larger system where favored companies become powerful, establishing a long-term technofeudal equilibrium.
If this monopolistic feedback loop keeps up, the economic balance will tip, harming both competitors and users in the long-term. From a market player point of view everything from newer AI startups, other technology companies, and even businesses outside of the technological landscape will be displaced as the state depends heavily on AI firms that already dominate the market for competitive edge within the global market. While from a consumer point of view personal data will be used and tracked on much more rampant levels to fuel the system needed to sustain AI firms. The only true beneficiaries are the AI cloudalists and states that fund this cycle.
With technofeudalism running the market, the very foundation of capitalism becomes molded into a new economic hierarchy where the lords are digital, the rents are coded, and the rest of us are just trying to navigate an economy we do not control anymore.