📊 Full opportunity report: The Neocloud Cartel: How the AI Industry Started Renting Compute From Itself on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
In 2026, AI industry compute resources are largely rented from a small cartel of firms, led by Nvidia. This shift decouples ownership from use and creates a fragile, tightly controlled market.
In 2026, the AI industry largely relies on a network of companies renting compute resources from each other rather than owning hardware outright, with Nvidia acting as the central hub. This shift has transformed the compute layer into a tightly interconnected cartel, raising questions about market power and fragility.
Almost none of the leading AI firms own the hardware they run on; instead, they rent from a small group of GPU landlords, notably Nvidia, which has become the dominant player in this ecosystem. Companies like xAI, Anthropic, and Google lease massive supercomputers, often paying billions monthly, with the money flowing back to Nvidia through various financing arrangements.
This circular financing creates a closed loop where a handful of firms—Nvidia, Microsoft, AMD, and a few others—finance each other’s compute needs, inflating valuations and consolidating power. Nvidia’s control over GPU allocation, especially during shortages, effectively gives it veto power over who can compete in the AI race.
Meanwhile, the leasing agreements often include governance clauses, such as Musk’s clause allowing capacity reclamation if an AI harms humanity, turning supply contracts into strategic leverage points. Dependency is high; for example, CoreWeave derives most of its revenue from just two clients, illustrating the fragility of this system.
The Neocloud Cartel
Almost no one racing to build AI owns the machine it runs on. They rent — increasingly from each other — and the money loops back to one chip maker that’s also an investor in nearly everyone at the table.
The cartel isn’t a conspiracy — it’s the endpoint of extreme capital intensity, real scarcity, and one dominant supplier. But the same circularity that makes it powerful makes it a fuse: each cancelled order is someone else’s missing revenue. Don’t be a price-taker at the bottom of a loop you don’t control — own your inference, keep an open-weight fallback, diversify silicon.
Implications of the AI Compute Cartel for Industry Power
This development concentrates market power in the hands of a few firms, especially Nvidia, which controls access to the essential hardware. The circular financing and leasing model makes the market highly dependent on a small number of players, risking potential supply shocks or strategic manipulation that could impact the entire AI ecosystem.
Furthermore, the decoupling of ownership from use means that AI development is now governed by contractual and allocation controls, not hardware ownership, which could influence innovation, competition, and regulatory oversight in the future.
Local AI on Linux in Practice: Build Private LLM Servers, GPU Workstations, Ollama Apps, Dockerized AI Services, and Self-Hosted AI Infrastructure with CUDA, ROCm, vLLM, and Open WebUI
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Rise of the Neocloud and the Shift to Renting Compute
Over the past three years, the AI industry has shifted from owning hardware to renting compute resources, driven by GPU shortages and the need for rapid scaling. The emergence of ‘neocloud’ hyperscalers like CoreWeave and the strategic leasing deals among major firms have created a new market dynamic.
In May 2026, xAI’s leasing of its supercomputer to Anthropic and Google marked a turning point, illustrating how even labs that develop AI now act as landlords, leasing capacity to rivals. This has led to a small, interconnected cartel where Nvidia dominates supply and financing, controlling the flow of compute resources.
This evolution reflects broader industry trends toward circular financing, dependency, and control, with major firms committing hundreds of billions of dollars to lease and build AI infrastructure without direct ownership.
“A gigawatt of AI data center capacity costs roughly $50 billion, with about $35 billion flowing to Nvidia alone.”
— Jensen Huang, Nvidia CEO
AI training GPU rental
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Uncertainties About Market Stability and Future Risks
It is not yet clear how fragile this compute cartel might be if key players face financial stress, supply disruptions, or regulatory intervention. The extent to which Nvidia’s control could be challenged remains uncertain, as does the potential for new entrants to break the cycle.
Further, the long-term impact of decoupling hardware ownership from AI development on innovation and competition is still developing, with industry observers debating possible regulatory responses.
Practical RHEL AI: Designing, Deploying and Scaling AI Solutions with Red Hat Enterprise Linux
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Industry Consolidation and Regulation
Expect increased scrutiny from regulators concerned about market concentration and supply chain control. Industry players may seek to diversify supply sources or develop alternative hardware solutions. Nvidia’s role as the central gatekeeper will likely be a key focus for future policy and market shifts.
Additionally, the industry may see efforts to create more open, ownership-based compute models, challenging the current leasing cartel structure.
High Performance Computing: Modern Systems and Practices
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How does Nvidia control AI compute access?
Nvidia controls access through its dominant position in GPU manufacturing, allocation decisions during shortages, and financing arrangements that influence who can lease hardware.
Why are AI companies renting compute instead of owning hardware?
GPU shortages and the high costs of building data centers make renting the only feasible option for rapid scaling and access to cutting-edge hardware.
What risks does this compute cartel pose to the AI industry?
The concentration of power could lead to supply disruptions, manipulation of access, or regulatory crackdowns, potentially destabilizing the ecosystem.
Could this system change in the future?
Yes, industry shifts toward open hardware, regulatory interventions, or new technological developments could challenge the current leasing and control model.
Source: ThorstenMeyerAI.com
