📊 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
AI firms increasingly rent compute from each other, creating a tightly linked cartel led by Nvidia. This shift decouples ownership from use, raising questions about market power and fragility.
Major AI companies are now renting compute from each other, with Nvidia acting as the central gatekeeper. This development marks a significant shift in how AI infrastructure is financed and controlled, impacting industry power dynamics and supply chains.
According to sources from Thorsten Meyer AI, the AI compute layer in 2026 resembles a small cartel, with a handful of firms financing each other’s hardware purchases. Companies like Anthropic, xAI, Meta, and OpenAI are leasing massive GPU clusters, often paying hundreds of millions of dollars monthly, with the key supplier being Nvidia.
For example, xAI leased its Colossus 1 supercomputer to Anthropic for about $1.25 billion per month and to Google for approximately $920 million per month. This indicates a shift where companies no longer own the hardware but rent it, often from each other or from shared suppliers, blurring the lines between competitors and collaborators.
Furthermore, the financing loop is heavily centered on Nvidia, which has invested up to $100 billion in OpenAI and holds equity in multiple firms within the ecosystem. Nvidia’s control over hardware supply and allocation grants it significant power over the entire AI infrastructure market, effectively making it the choke point.
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 emerging cartel-like structure concentrates control over AI compute resources in the hands of a few firms, primarily Nvidia. It means that access, pricing, and capacity are dictated by a small circle of companies, potentially affecting innovation, competition, and pricing in the AI industry. The decoupling of ownership from use also raises concerns about market fragility, as the entire system relies on continuous financing and supply agreements among interconnected firms.
Nvidia GPU cloud computing hardware
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background on the Shift to Compute Renting in AI
Until recently, AI companies often owned their hardware or leased directly from general-purpose cloud providers. The GPU shortage of 2024–25 accelerated the move toward renting specialized AI hardware, leading to the rise of the ‘neocloud’—a market of hyperscale GPU-as-a-service providers like CoreWeave, Meta, and others.
By 2026, this ecosystem has evolved into a tightly interconnected network where companies lease from each other, creating a circular flow of capital and resources. Nvidia’s dominant role in manufacturing and financing has become the central pillar of this new structure, transforming the compute layer into a chokepoint.
“A gigawatt of AI data center capacity costs roughly $50 billion, with the majority flowing to Nvidia.”
— Jensen Huang, Nvidia CEO
AI GPU clusters for enterprise
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unclear Aspects of the AI Compute Cartel’s Stability
It remains unclear how sustainable this tightly linked system is, given its reliance on continuous financing and supply agreements. The fragility of the circular funding loop could lead to disruptions if any key player pulls back or if supply constraints intensify further.
Additionally, the full extent of Nvidia’s influence and whether regulatory actions might challenge this concentration of power are still evolving issues.
high performance AI training servers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Monitoring the AI Compute Market Dynamics
Industry observers will watch for signs of strain or fragmentation within this cartel structure, especially as new competitors or regulatory interventions emerge. Further developments may include shifts in financing arrangements, capacity expansion, or calls for increased transparency and regulation.
Key events to follow include Nvidia’s future hardware allocations, potential antitrust investigations, and the evolution of leasing agreements among AI firms.
AI infrastructure leasing services
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How does Nvidia control the AI compute market?
Nvidia dominates through manufacturing, financing, and allocation decisions, effectively acting as the central gatekeeper for GPU supply and pricing in the AI industry.
Why are companies leasing compute instead of owning hardware?
The GPU shortage of 2024–25 made owning hardware impractical for many firms, leading to a shift toward renting as a faster, more flexible way to scale AI infrastructure.
What risks does this cartel-like structure pose?
The reliance on a small number of firms and financing loops creates fragility; disruptions could occur if any key player withdraws or supply chains break down.
Could regulatory actions break up this cartel?
Potentially, as regulators may scrutinize Nvidia’s market dominance and the interconnected financing arrangements, but no major moves have been announced yet.
What does this mean for AI innovation?
Concentration of control could limit competition and innovation, as access to compute becomes more restricted and expensive.
Source: ThorstenMeyerAI.com
