📊 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 companies increasingly rent compute from each other, forming a small cartel led by Nvidia. This shift decouples ownership from use, creating a fragile but powerful chokehold on AI development.
In 2026, the AI industry has transitioned to a model where companies no longer own the machines they run on but instead rent compute from each other, with Nvidia emerging as the central figure in this new cartel. This shift significantly impacts how AI development is financed and controlled, making access to compute a tightly guarded commodity.
Recent reports reveal that the AI industry increasingly relies on a network of GPU landlords, including firms like CoreWeave, Meta, OpenAI, and xAI, that rent hardware from Nvidia and each other. Notably, xAI leased its supercomputer to Anthropic and Google for over $26 billion annually, despite its own underutilized capacity, signaling a move towards renting as the primary means of scaling AI operations.
This circular renting system is driven by a small group of dominant firms that finance, supply, and control access to compute resources. Nvidia, in particular, holds a commanding position, investing heavily in AI firms and controlling GPU allocation, effectively acting as the choke point for the entire industry. The financial arrangements often involve circular investments, with Nvidia and major tech companies like Microsoft and Amazon funneling billions into AI startups, which then spend on Nvidia hardware and services.
This network has created a tightly interconnected ‘cartel’ where access, pricing, and capacity are governed by a handful of firms, making the industry highly dependent on Nvidia’s supply chain and strategic decisions. The contracts often include clauses that give landlords governance rights, such as Musk’s lease to Anthropic, which allows capacity reclamation if certain conditions are met, adding a layer of control over AI development.
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 reshapes the industry’s power dynamics, concentrating control over AI development in the hands of a few firms, primarily Nvidia. It creates a situation where access to compute is no longer a matter of ownership but of contractual and financial leverage, which can be used to influence AI progress and market competition.
Furthermore, the circular financing and dependency create a fragile ecosystem: if Nvidia or key financiers withdraw support, the entire AI buildout could face disruption. This concentration of power also raises concerns about transparency, market fairness, and the potential for bottlenecks that could slow innovation or lead to monopolistic behaviors.
Linux Hardware Mastery: Drivers, Firmware, GPUs, Wi-Fi, and Making Any PC Work (The Modern Linux Mastery Series)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Origins and Evolution of the AI Compute Cartel
Historically, AI companies owned their own hardware, but the GPU shortage of 2024–25 forced a shift towards renting compute resources. CoreWeave, Meta, and other hyperscalers emerged to meet demand, but by 2026, a new pattern formed: companies began leasing from each other and from Nvidia, creating a tightly knit network of financial and hardware dependencies.
The involvement of xAI in leasing its supercomputer to rivals marked a turning point, illustrating that even AI labs are now acting as landlords. This evolution was driven by supply constraints, high costs, and the need for rapid scaling, which made renting the only viable option for many firms. The industry’s financial flows increasingly resemble a circular loop, with Nvidia at the center, financing and controlling the flow of hardware and capital.
“A gigawatt of AI data center capacity costs roughly $50 billion, with about $35 billion flowing to Nvidia — making us the gatekeeper of AI compute.”
— Jensen Huang, Nvidia CEO
AI training GPU rental services
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unconfirmed Aspects of the AI Compute Cartel’s Future
While the current structure appears stable, it remains unclear how long Nvidia’s dominance can be maintained without attracting regulatory scrutiny or market shifts. The fragility of the circular financing loop raises questions about potential disruptions if key players withdraw support or if new competitors emerge.
Additionally, the long-term impact of contractual governance clauses, such as capacity reclamation rights, on industry innovation and competition is still uncertain. It is also unclear how governments or regulators might respond to this concentration of control in the AI supply chain.
NVIDIA DGX Spark™ – Personal AI Desktop Supercomputer – Desktop GB10 Grace Blackwell Chip
Supercomputer performance directly to your desk in a compact, energy-efficient design, enabling enterprise-scale AI and high-performance computing right…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Potential Developments and Industry Responses
The industry may see increased regulatory scrutiny as the concentration of compute control becomes more apparent. Nvidia’s strategic investments and contractual leverage could face antitrust investigations or calls for more open hardware access.
Meanwhile, startups and new entrants might attempt to develop alternative hardware or leasing models to bypass the existing cartel. The next few years will likely determine whether this structure consolidates further or begins to fracture under external pressures.
enterprise GPU server rentals
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 over owning hardware?
Renting allows faster scaling, lower upfront costs, and flexibility to adapt to demand fluctuations, especially during hardware shortages like those in 2024–25.
How does Nvidia control the AI compute supply chain?
Nvidia dominates GPU manufacturing, invests heavily in AI firms, and controls hardware allocation through strategic contracts, making it the central choke point.
What risks does this cartel structure pose for the AI industry?
The reliance on a small group of firms creates vulnerability; if Nvidia or financiers withdraw support, the entire AI development ecosystem could face disruption.
Could regulatory action break up this compute cartel?
Potentially, yes. Increased scrutiny over monopolistic practices and hardware control could lead to regulations aimed at increasing transparency and competition.
What alternatives might emerge to challenge Nvidia’s dominance?
Development of alternative hardware, open-source compute platforms, or new leasing models could provide competitors with options outside the current cartel system.
Source: ThorstenMeyerAI.com
