📊 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 closed loop controlled by a small group of firms led by Nvidia. This setup forms a cartel that concentrates power but also introduces fragility into the AI industry.
In 2026, the AI industry has transitioned to a model where most companies rent compute from each other, rather than owning their own hardware. This shift has created a tightly interconnected network of firms, with Nvidia acting as the central power broker, controlling access to essential GPU resources. This development matters because it consolidates industry influence in a small circle of firms, raising questions about competition, supply chain resilience, and market power.
The core of this shift is the rise of the ‘neocloud’ model, where AI-focused hyperscalers rent GPU compute without owning the hardware outright. Major players like CoreWeave, Meta, OpenAI, and xAI now lease from each other and from Nvidia, often on multi-billion-dollar contracts. Notably, xAI leased its supercomputer to Anthropic for about $1.25 billion monthly and to Google for roughly $920 million monthly, highlighting how even self-described AI labs act as landlords.
This circular leasing creates a ‘cartel’-like structure, where a handful of firms—predominantly Nvidia—control the flow of compute resources. Nvidia alone captures the majority of the $50 billion per gigawatt of data center costs, and it holds equity stakes in many of the leasing firms. Nvidia’s strategic investments, including a $100 billion fund for OpenAI and stakes in other firms, reinforce its central position. This concentration of control means access to compute is now governed by contractual and allocation decisions made by a small group, rather than open market competition.
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 AI Compute Cartel Concentration
This emerging cartel consolidates significant industry power within a small group of firms, especially Nvidia, which controls GPU supply and allocation. Such control can influence AI development trajectories, pricing, and access, potentially stifling competition and innovation. The circular finance and leasing model also introduces systemic fragility: if one key player withdraws or faces disruption, the entire supply chain could be affected. This raises concerns about the resilience of AI infrastructure and the risks of market manipulation by a handful of firms.
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.
Formation of the AI Compute Cartel and Its Drivers
The shift toward renting compute was driven by the 2024–25 GPU shortage, which made owning hardware prohibitively expensive and slow. Companies turned to GPU-as-a-service providers like CoreWeave, Meta, and others, creating a market that quickly evolved into a tightly interconnected network. The emergence of xAI as a landlord, leasing its underutilized supercomputer to rivals, marked a turning point, highlighting how ownership has decoupled from use in the AI industry. The circular financing and leasing relationships have since cemented a small group of firms as the gatekeepers of AI compute power.
“A gigawatt of AI data center capacity costs around $50 billion, with most of that flowing to Nvidia, which holds the key to GPU supply and allocation.”
— Jensen Huang, Nvidia CEO
AI Data Center Infrastructure Engineering: Power Distribution, Liquid Cooling, High-Density Networking, and Energy Efficiency for GPU Training … Hardware & Compiler Engineering Series)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Uncertainties About Industry Stability and Future Risks
It is not yet clear how resilient this cartel structure is to disruptions, such as supply chain shocks, regulatory interventions, or shifts in company strategies. The long-term stability of the leasing and financing relationships remains uncertain, and potential antitrust scrutiny could challenge Nvidia’s central role. Additionally, how this concentration impacts innovation and market entry for new players is still unclear.
NVIDIA RTX PRO 6000 Blackwell Server Edition
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Potential Developments and Industry Responses
Next steps include increased regulatory attention, especially concerning monopolistic behaviors and supply chain vulnerabilities. Companies may seek alternative compute sources or develop proprietary hardware to reduce dependence. Nvidia’s role will likely be scrutinized, and shifts in leasing agreements or new entrants could alter the current cartel dynamics. Monitoring how these relationships evolve will be critical for understanding the future of AI infrastructure.
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 do AI companies prefer renting compute instead of owning hardware?
Renting allows companies to access the latest hardware without large upfront investments, especially during shortages. It provides flexibility and scalability, which are crucial for rapidly evolving AI workloads.
What role does Nvidia play in this AI compute cartel?
Nvidia is the central power broker, controlling most GPU supply and allocation. It also invests in many leasing firms, making it the key gatekeeper of AI infrastructure.
Could this cartel structure lead to decreased competition?
Yes, concentration of control in a small group of firms could limit market entry and innovation, raising concerns about monopolistic practices and market resilience.
What risks does this structure pose to the AI industry?
The reliance on circular leasing and a few dominant suppliers creates systemic fragility. Disruptions could cascade through the supply chain, affecting AI development and deployment.
Source: ThorstenMeyerAI.com
