AI agents could create supercooperative firms with no internal limits

The Coaseian Boundary

To understand how AI changes business, we must first look at the Coaseian boundary. In 1937, economist Ronald Coase argued that a firm's size is limited by the cost of doing a task internally versus buying it from the market. If it is cheaper to hire an outside contractor, the firm stays small; if it is cheaper to do it in-house, the firm grows.

Historically, major inventions have shifted these boundaries in different directions. Technologies like the ledger, the telegraph, copy paper, and the internet have all functioned by reducing either internal or external costs. These tools changed how much work a single company could manage before it became too expensive to keep everything under one roof.

As exponentialview.co reported, the introduction of AI agents introduces a new variable into this equation. Unlike human employees, who require management and face communication friction, AI agents can operate with a different set of economic rules. This could fundamentally alter the point at which a company decides to stop growing or start outsourcing.

The Rise of Supercooperative Firms

AI agents differ from humans because they can share the same reasoning patterns, which are known as model weights. When one agent hands off a task to another, there is no loss of information or friction in the transition. This creates a environment where nothing leaks during the process, and no agent is motivated by personal career advancement.

This technical capability could lead to what some call "supercooperative firms." In these entities, becoming larger actually makes the company better at every scale. Because the agents share a working state, the company can expand its operations without the typical headaches of human coordination or organizational silos.

This efficiency might allow for the creation of giant AI-run firms that dominate deep, large markets. While human-led firms eventually hit a wall of complexity, AI-driven firms could potentially scale infinitely. The lack of friction in agent-to-agent communication removes the traditional penalties for expansion.

Scaling Across Industries

The ability to scale horizontally is a major shift for the modern economy. Traditionally, firms specialize in one area because managing diverse operations is costly. However, AI-capable firms might be able to enter many different markets simultaneously without significant overhead.

Anthropic serves as a current example of this trend, as the company is active in various fields including coding, biology, legal services, and finance. Because AI can be applied to these disparate domains, a single firm can expand its reach across multiple industries at once. This allows a company to move past the traditional Coasian boundary that usually keeps industries separate.

This horizontal expansion means that a single entity could become a dominant player in many sectors. Instead of a company being just a software firm or just a biotech firm, it could become a multi-industry powerhouse. The underlying AI technology provides the infrastructure to manage these diverse interests under one corporate umbrella.

Economic Concentration

The trend toward larger firms is not a new phenomenon, but AI may accelerate it significantly. Data shows that the largest 1% of American firms already held 97% of corporate assets in 2018. This indicates a high level of concentration in the current economy, where a small number of entities hold the vast majority of wealth.

AI is a technology that is more likely than not to accentuate this existing trend. By providing a way to scale horizontally and internally without human friction, AI gives large firms a massive advantage over smaller competitors. The firms that own the best models can expand into every market, potentially capturing a larger share of the total economy.

This concentration could lead to the rise of enormous entities that operate on a scale previously thought impossible. As these firms grow, they may become so large that they require new types of oversight. The economic landscape may shift from a collection of specialized companies to a few massive, AI-driven conglomerates.

New Regulatory Questions

The emergence of these supercooperative firms will likely prompt new questions for regulators. If a single firm can operate across coding, finance, and biology with infinite scalability, the traditional rules for antitrust and competition may no longer apply. We may need to develop new frameworks to deal with entities of this magnitude.

Regulators will have to consider how to manage firms that do not have traditional human boundaries. Because AI agents can share weights and states, the internal workings of these companies will be vastly different from traditional corporations. This transparency—or lack thereof—presents a unique challenge for policy makers.

The goal for the future will be to determine how to maintain a competitive market when AI allows for such rapid expansion. As firms become bigger and faster, the speed of their growth may outpace the ability of current laws to keep up. Watching how these enormous entities are governed will be a key area of focus for the coming years.