📊 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 companies are increasingly renting compute from each other, forming a cartel centered around Nvidia. This shift decouples ownership from use, creating a fragile but influential network on AI infrastructure.
In 2026, the dominant trend in the AI industry is that companies no longer own the hardware they run on; instead, they rent compute from each other, with Nvidia acting as the central hub. This shift has transformed the compute layer into a network with concentrated control, affecting the distribution and availability of resources for AI development and scaling.
Recent reports reveal that major AI firms, including OpenAI, xAI, and Anthropic, are leasing hundreds of millions of dollars’ worth of GPU capacity from each other and from Nvidia, which supplies the majority of the hardware. Notably, xAI leased its supercomputer to Anthropic for about $1.25 billion a month and to Google for approximately $920 million a month, totaling roughly $26 billion annually.
This pattern indicates that ownership of hardware has become decoupled from AI deployment, with companies acting as landlords and tenants simultaneously. Nvidia, which supplies around $35 billion of the estimated $50 billion gigawatt cost of AI data centers, effectively controls the supply chain and GPU allocation. Nvidia’s investments extend beyond hardware, including equity stakes in firms like CoreWeave, Nebius, and xAI, and financing arrangements that reinforce its central role.
The circular financial relationships mean that a small number of firms—Nvidia, Microsoft, Amazon, and a handful of others—are both financiers and customers, creating a network with concentrated control over AI infrastructure. This arrangement allows these firms to influence access, pricing, or availability of compute resources based on contractual and strategic considerations.
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 Structure
This development indicates a shift in how AI infrastructure is managed, with a small group of firms exerting significant influence over AI progress. Nvidia’s dominant position in GPU supply and financing enables it to determine access to compute resources for AI development, which could affect innovation, competition, and market stability. The decoupling of ownership from use also introduces potential vulnerabilities, as the ecosystem relies on a tightly interconnected supply chain that could be susceptible to disruptions or strategic changes.
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 Its Market Dynamics
The concept of neocloud emerged around 2024–25 as a response to GPU shortages, with companies like CoreWeave and Meta leading the provision of GPU-as-a-service. By 2026, this sector has evolved into a network with concentrated control, with most AI firms leasing hardware from a limited number of providers, all heavily financed by Nvidia and other large tech firms. The industry trend favors leasing over ownership due to factors such as cost, speed, and supply constraints.
Previously, hardware ownership was more dispersed, but the GPU shortages of 2024–25 accelerated the move toward leasing, making the compute layer a strategic point of control. This transition has implications for competition and innovation in AI, as access to hardware becomes more strategically managed.
“A gigawatt of AI data center capacity costs roughly $50 billion, most of which flows to Nvidia.”
— Jensen Huang, Nvidia CEO
AI data center GPU rentals
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unclear Aspects of the Compute Cartel’s Stability
While the structure of the network and Nvidia’s role are well-documented, the long-term stability of this interconnected system remains uncertain. Potential disruptions, regulatory actions, or changes in supply chain dynamics could impact the current model, but specific vulnerabilities have yet to be fully assessed.
AI Systems Performance Engineering: Optimizing Model Training and Inference Workloads with GPUs, CUDA, and PyTorch
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Potential Risks and Future Industry Shifts
Industry analysts anticipate increased regulatory scrutiny and the emergence of new entrants seeking to challenge Nvidia’s influence. Additionally, the reliance on circular financing and leasing arrangements introduces vulnerabilities that could lead to disruptions if key participants experience strategic or financial shifts. Monitoring these developments will be important for understanding future industry dynamics.
Deep Learning at Scale: At the Intersection of Hardware, Software, and Data
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
Why are AI companies renting compute instead of owning it?
Due to recent GPU shortages and high costs, leasing has become a practical approach for scaling AI capabilities without making long-term hardware investments.
How does Nvidia control the AI compute supply chain?
Nvidia dominates GPU manufacturing, holds significant equity stakes in rental firms, and manages allocation through contractual and financial arrangements.
What are the risks of this compute cartel?
The interconnected nature and concentrated control pose risks such as supply disruptions and strategic manipulation, which could affect AI development and market stability.
Could regulatory action break up this cartel?
Potentially, but current regulatory measures have not targeted this structure, and the complex financial relationships present challenges for intervention.
What might change the current dynamics?
Advances in hardware manufacturing, entry of new providers, or regulatory measures could alter the current control dynamics within the AI compute ecosystem.
Source: ThorstenMeyerAI.com
