Market Actors: The "Magnificent Seven"
The Weight of Hyperscalers in the US Economy
The current paper examines the dynamics of concentration and competition because of the drastic imbalance that can be observed within the field of AI and its place in the economy as a whole. As it currently stands, all but one1 of the seven top ranked companies globally (and within the S&P 500) by market capitalization (Nvidia, Apple, Alphabet, Amazon, Microsoft, Meta and Tesla, the so-called "magnificent seven") -- with a total valuation of 22 trillion dollars out of a total global market capitalization of 115 trillion(1) -- have their activity intimately tied with the scaling use of compute in AI and advanced compute features. Notably, out of these seven companies, only 3 (Apple, Google, and Microsoft) were in the top 10 most valuable companies in 2015, and the sum of the top 10 was of "only" 3.6 trillion out of 62 total global market capitalization.
This begs the question of what constitutes the main drivers of this extreme concentration of capital on a global level, and whether AI is set to have a strong impact in either accelerating this concentration (which seems to currently be the case), or mitigating it, depending on how it impacts different competition dynamics. In particular, AI has taken a growing place both in the public discourse and in the advertisement strategies of all of these companies. At the same time, access to AI compute, which all of them have in common as a resource and which they control most of the market for, is increasingly touted as the main component of "ever more advanced" AI systems. In this work, we interrogate whether the relationship between AI and compute makes further concentration unavoidable given the current resources of companies outside of the "magnificent 7", by looking at the compute costs of "artificial intelligence" under different development models and comparing those to the needs and expenses of the broader set of commercial actors.
High Priority TODO: Outline how AI as a ramp to global digital monopoly can justify the gap between AI investment and expenses, i.e. it's another iteration of Amazonification and Uberification more broadly applied.
Footnotes
-
While Tesla's line of business is not directly integrated with AI, they do have multiple direct links to AI market structure: in terms of compute, they produce Megapack batteries for data centers and are developing a 500 MW data center in Saudi Arabia; in terms of AI development and usage, they develop their own models for autonomous driving applications and can be directly linked, via their common CEO, to xAI, which is an increasingly prolific provider of AI as a service and whose value as of January 2026 is estimated at up to $230 billion. ↩