How does controlling chips, models, and data translate into market power?
Control over the core building blocks of AI—specialized chips, large-scale models, and massive datasets—creates a self-reinforcing cycle that concentrates market power. A 2025 study of the AI chip market found that NVIDIA's dominance in GPUs (graphics processing units) gives it a structural advantage that it can sustain through continued R&D investment and a deep software ecosystem [2]. Similarly, a 2026 panel analysis of the U.S. telecommunications industry from 2000 to 2025 showed that firms with larger investments in AI, bigger data assets, and greater computing capacity consistently achieved stronger market positions and higher concentration, as measured by the Herfindahl-Hirschman Index (a standard metric of market concentration) [5]. In agriculture, a 2026 study found that firms with higher AI adoption demonstrated stronger productivity and greater pricing influence, with digital data infrastructure reinforcing competitive advantages for larger agribusinesses [8]. The pattern is clear: whoever controls the inputs controls the market.
What happens to competition and innovation when a few firms dominate?
Competition tends to weaken, and innovation can become lopsided or stalled. A 2024 study of Ethereum's block-building auctions—a market where builders compete to create blocks for the blockchain—found that a small set of dominant builders used advantages in latency and access to maximal extractable value (MEV) to consolidate power, reduce auction efficiency, and increase centralization [4]. This mirrors findings in AI markets: a 2026 paper on the socio-legal dimensions of AI competition warns that control over training data and algorithmic infrastructure leads to the formation of "digital oligopolies," where a few firms dominate and users become objects of manipulative algorithmic influence rather than active market participants [3]. However, the picture is not entirely bleak. A 2025 study on AI ecosystems in Africa notes that backward and forward integration—where firms move into adjacent markets—can further enhance ecosystem power, but it also suggests that proactive competition policy can prevent the worst outcomes [1]. The key is that without intervention, concentration tends to increase, but regulation can change the trajectory.
Can anything stop the concentration of power?
Yes, but it requires deliberate policy action. The strongest evidence comes from the 2026 telecom study, which found that regulatory oversight significantly moderated the effect of AI-driven advantages on market power, limiting the translation of technological advantages into persistent dominance [5]. This suggests that antitrust enforcement, data-sharing mandates, and transparency rules can work. A 2025 paper on AI in oncology calls for global collaborations to ensure equitable access to computing resources and protect crucial AI infrastructure, arguing that trade barriers and export controls (like those limiting Chinese firms' access to advanced chips) create an unsustainable environment for AI development worldwide [6]. The 2026 socio-legal paper proposes a framework of "adaptive regulation" that combines preventive antitrust control, energy transparency requirements, and guarantees of digital autonomy for users [3]. Meanwhile, a 2025 study using game theory and reinforcement learning shows that in an oligopoly, a leader firm can anticipate and influence follower strategies, but the framework also demonstrates that followers can adapt and compete effectively under dynamic conditions—suggesting that market structure is not destiny [7]. The bottom line: without regulation, concentration worsens; with smart, proactive policy, the worst effects can be contained.
About These Sources
This answer is built on 8 peer-reviewed studies — published from 2024 to 2026, 8 from 2024 or later, 1 in Q1 journals — selected as the most relevant from 8 studies that passed quality screening, drawn from 87 papers retrieved from a database of over 500 million.
Sources used in this answer
Competition, Market Power, and Artificial Intelligence Ecosystems in Africa
Using network analysis, this 2025 study shows that AI ecosystems in Africa are highly integrated, meaning market power in one input market (e.g., chips) can amplify power across the whole ecosystem, and that African firms are largely absent from chip design and fabrication.
Nvidia in the AI Chip Market Competition: Based on SWOT Model and Financial Analysis
A 2025 SWOT and financial analysis of NVIDIA finds that its dominance in the AI chip market is sustained by continued R&D investment, a deep software ecosystem, and supply chain management, positioning it to maintain high growth in an increasingly competitive market.
Socio-Legal Dimensions of Competition Among Artificial Intelligence Systems for Users and Energy Resources
This 2026 study identifies legal risks of asymmetric AI competition, including the formation of digital oligopolies through control over training data and algorithmic infrastructure, and proposes adaptive regulation combining antitrust control, energy transparency, and user digital autonomy.
From Competition to Centralization: The Oligopoly in Ethereum Block Building Auctions
Using empirical game-theoretic analysis of Ethereum's block-building auctions, this 2024 study finds that a small set of dominant builders leverage latency and MEV access advantages to consolidate power, reduce auction efficiency, and increase centralization.
Market Power in the Age of Artificial Intelligence: An Empirical Panel Analysis of Data Scale and Government Regulation in the United States Telecommunications Industry
A 2026 panel analysis of the U.S. telecom industry (2000–2025) shows that greater AI investment, larger data assets, and enhanced computing capacity are associated with stronger market positions and higher concentration, but regulatory oversight significantly moderates these effects.
The need for expansion of global collaborations on AI in oncology
This 2025 commentary on AI in oncology highlights that export controls on advanced chips create an unsustainable environment for AI development and calls for global collaborations to ensure equitable access to computing resources and protect AI infrastructure.
Strategic Decision-Making in Oligopoly Markets Based on Stackelberg Game and Reinforcement Learning Algorithms
A 2025 study combining Stackelberg game theory with reinforcement learning shows that in an oligopoly, a leader firm can anticipate and influence follower strategies, but followers can adapt and compete effectively under dynamic market conditions.
Artificial Intelligence, Digital Agriculture, and Market Power in the United States Agricultural Sector
A 2026 econometric study of the U.S. agricultural sector finds that firms with higher AI adoption demonstrate stronger productivity and greater pricing influence, and that digital data infrastructure reinforces competitive advantages for larger agribusinesses, contributing to increased market concentration.