The AI ecosystem is uniquely prone to monopolisation due to massive barriers to entry. To build and maintain cutting-edge AI (like Large Language Models), three critical resources are required:

• Compute power. Access to massive, expensive data centers and specialised hardware such as advanced GPUs (graphics processing units).

• Data hegemony. Vast, proprietary datasets generated by billions of users, which are used to train and continuously refine AI models.

• Talent drain. The financial capacity to hoard elite AI researchers and engineers, drawing them away from academia and public research.

Because of these requirements, smaller startups and public institutions are increasingly forced to rely on the infrastructure of dominant firms, effectively turning "Big Tech" companies into the "gatekeepers" of the AI era.

Beyond market structure, monopolisation also manifests in algorithmic behaviour. Regulators and researchers have increasingly focused on algorithmic collusion - when pricing algorithms independently learn to maintain artificially high prices without any direct communication between sellers. 

AI systems can also be used to entrench dominance by leveraging data advantages, locking in platform users and partners, and making it structurally difficult for rivals to compete. 

Why it matters

Healthy competition usually lowers costs, improves quality, and expands consumer choice. If AI markets become too concentrated, society may get fewer independent innovators, less resilience, and more dependence on a few powerful companies. In public interest terms, that can also affect democracy, access to information, and who gets to decide how AI is used. 

Impacts & consequences

• Economic inequality. Wealth and economic surplus are concentrated heavily within a few corporations, while smaller businesses are squeezed out or aggressively acquired before they can compete. 

• Price inflation. Algorithmic pricing tools can artificially inflate prices across entire markets, from housing to airline tickets.

• Innovation stagnation. Dominant firms can acquire or neutralise potential disruptors, reducing the diversity of approaches and slowing progress outside of incumbent priorities.

• Censorship and control. Gatekeepers can quietly alter algorithmic policies, manipulate search/information visibility, and restrict access to specific tools, effectively shaping public discourse and behaviour.

• Regulatory capture and governance gaps. Unlike banking, insurance, telecommunications, and media, where markets are already regulated, the AI and ML industry has no comparable framework, allowing large tech companies to pursue vertical integration to control the full value chain.

• Geopolitical risk. AI monopolisation has a national security and sovereignty dimension, as dependence on a handful of foreign-owned AI platforms creates strategic vulnerabilities for governments and economies.

• Vendor lock-in. Developers, enterprises, and public institutions become structurally dependent on proprietary AI ecosystems, losing flexibility, negotiating power, and independence.

Causes & contributing factors

• High capital requirements. The astronomical cost of training frontier models (often running into hundreds of millions of dollars) naturally locks out smaller competitors.

• Network effects. Platforms grow more valuable as more users join, making displacement increasingly difficult.

• Aggressive mergers & acquisitions. Dominant firms routinely engage in "killer acquisitions" - buying promising AI startups early to absorb their tech and talent or eliminate them as competitors.

• Regulatory lag. In many jurisdictions, competition law has failed to keep pace with the speed and opacity of AI-driven market dynamics.

• Regulatory capture. Large tech firms possess the lobbying power to shape AI safety regulations in ways that compliance becomes too expensive for open-source developers and startups to survive.