Neocloud Companies, Explained: The New Landlords of the AI Buildout

What is a neocloud company?

A neocloud is a cloud provider built for one job: renting out high-performance computing for artificial intelligence. Where the traditional hyperscalers, Amazon Web Services, Microsoft Azure and Google Cloud, offer a sprawling menu of hundreds of services, a neocloud sells something far more specific. It sells access to graphics processing units, most often the Nvidia chips that train and run modern AI models, delivered as capacity rentable by the hour, the month, or the multi-year contract. The industry shorthand for the core product is GPU-as-a-service.

That specialization is the whole point. A neocloud is not trying to host your website, your database and your email. It is trying to give an AI lab thousands of the latest accelerators, wired together with fast networking and tuned software, faster and often cheaper than a general-purpose cloud can. The customer is not buying square footage or generic servers.

It is buying scarce, cutting-edge compute at the exact moment that compute is the most contested resource in technology.

Why neoclouds exist: the capacity gap

Neoclouds exist because demand for AI compute arrived faster than the traditional cloud could build for it. Training a frontier model, or serving it to millions of users, requires enormous, tightly clustered GPU capacity, and the hyperscalers, for all their scale, could not stand up data centers quickly enough to meet the surge. Building a new hyperscale facility can take three to five years of permitting, construction and grid interconnection. A neocloud that locked up power and space early can rack GPUs and bring capacity online in a fraction of that time.

That speed advantage created an opening, and the economics are unusual. Many neoclouds sign multi-year, take-or-pay contracts, in which a customer commits to a fixed amount of GPU capacity at a fixed price for years, whether or not it uses every hour. That structure gives these companies something rare for a young, capital-intensive business: years of contracted revenue visibility. The trade-off is that building the capacity requires staggering amounts of upfront capital, most of it borrowed against the GPUs themselves and the future revenue those contracts promise.

The clearest proof of the opportunity is who is doing the renting. The buyers are not only startups. They are the largest and most sophisticated technology companies on earth. Microsoft has committed tens of billions of dollars to neocloud partnerships to cover its own capacity gap. The same hyperscalers that neoclouds appear to compete with have become some of their biggest customers, because even they cannot build fast enough.

Major Players

The public neocloud universe sits on a spectrum, from pure AI-cloud platforms to companies that started as something else and pivoted toward AI infrastructure.

CoreWeave (CRWV) is the flagship neocloud. It listed on the Nasdaq in March 2025 and has become the largest public pure-play with customers that include OpenAI, Microsoft, Anthropic and Meta. The company has surpassed one gigawatt of active power and targets more than eight gigawatts by 2030.

Nebius Group (NBIS) is the fast-growing challenger. Emerging from the former Yandex, it has positioned itself as a full-stack AI cloud with a heavier software layer, which management argues supports better pricing and margins.

IREN (IREN) represents a crossover/pivot story. It began as a Bitcoin miner and has been converting its power sites and data centers into AI and high-performance computing capacity. That heritage gives it something valuable, namely control of power and low-cost sites, but it also means the market is still watching for the moment AI cloud revenue structurally overtakes the legacy mining business.

Applied Digital (APLD) sits closer to the power-and-campus end of the spectrum. It is less a software platform than an execution story built on delivering large, powered data-center campuses to meet hyperscaler demand. It is a reminder that the neocloud opportunity is not one business model but several, layered across power, real estate, GPU access and software.

What Meta's cloud ambitions really signal

In July 2026, Bloomberg reported that Meta was developing a cloud business, known internally as Meta Compute, to sell access to its AI models and, potentially, raw computing capacity of the kind neoclouds rent today. The market's first reaction was harsh. Shares of CoreWeave and Nebius fell sharply on the news while Meta rose, and the reason was not simply a new competitor. Meta is one of the largest customers both companies have, with commitments reported at around 21 billion dollars to CoreWeave and up to 27 billion dollars to Nebius. A customer that builds enough capacity to sell its own excess is a customer that might one day rent less.

That risk is real but taking a step back, the same event tells a more constructive story about the size of the opportunity. Meta is weighing this move because it is spending well over $100 billion dollars a year on AI infrastructure and still finds itself compute-constrained. When one of the most demanding technology companies in the world decides that owning and renting compute is a business worth entering, it the economic potential that neoclouds operate in. Compute is the scarce, valuable layer of the AI economy, and demand for it runs deep enough to pull even the hyperscalers' own customers onto the supply side.

The nuance for investors is that Meta's entry does not erase neocloud demand so much as validate how large and durable it is, while raising the bar on customer concentration and pricing discipline. It is a signal to watch, not a verdict. And notably, Meta remains a buyer today, still contracting for outside capacity even as it explores selling its own.

Why investors should care

The neocloud story is a story about where value accrues in artificial intelligence. Enormous attention goes to the model builders and to the chip designer at the center of it all. Neoclouds occupy the layer in between, the infrastructure that turns chips and power into usable AI capacity, and that layer has become indispensable precisely because everyone, from startups to hyperscalers to Meta, needs more of it than they can build themselves.

The companies building the physical backbone of AI have moved from a niche most investors had never heard of to one of the most consequential layers of the entire trade.

Glossary

A graphics processing unit (GPU) is a specialized electronic circuit designed for digital image processing and to accelerate computer graphics, being present either as a component on a discrete graphics card or embedded on motherboards, mobile phones, personal computers, workstations, and game consoles.

This information is provided solely as general investment education. None of the information provided should be regarded as a suggestion to engage in or refrain from any investment related course of action. Investing involves risk, loss of principal is possible.

Not an offer: This document does not constitute advice or a recommendation or offer to sell or a solicitation to deal in any security or financial product. It is provided for information purposes only and on the understanding that the recipient has sufficient knowledge and experience to be able to understand and make their own evaluation of the proposals and services described herein, any risks associated therewith and any related legal, tax, accounting or other material considerations. To the extent that the reader has any questions regarding the applicability of any specific issue discussed above to their specific portfolio or situation, prospective investors are encouraged to contact 1-855-561-5728 or consult with the professional advisor of their choosing.

Forward-looking statements: Certain information contained herein constitutes “forward-looking statements,” which can be identified by the use of forward-looking terminology such as “may,” “will,” “should,” “expect,” “anticipate,” “project,” “estimate,” “intend,” “continue,” or “believe,” or the negatives thereof or other variations thereon or comparable terminology. Due to various risks and uncertainties, actual events, results or actual performance may differ materially from those reflected or contemplated in such forward-looking statements. Nothing contained herein may be relied upon as a guarantee, promise, assurance or a representation as to the future.

Use of Third-party Information: Certain information contained herein has been obtained from third party sources and such information has not been independently verified by Roundhill Financial Inc. No representation, warranty, or undertaking, expressed or implied, is given to the accuracy or completeness of such information by Roundhill Financial Inc. or any other person. While such sources are believed to be reliable, Roundhill Financial Inc. does not assume any responsibility for the accuracy or completeness of such information. Roundhill Financial Inc. does not undertake any obligation to update the information contained herein as of any future date.

Any indices and other financial benchmarks shown are provided for illustrative purposes only, are unmanaged, reflect reinvestment of income and dividends and do not reflect the impact of advisory fees. Investors cannot invest directly in an index.

Except where otherwise indicated, the information contained in this presentation is based on matters as they exist as of the date of preparation of such material and not as of the date of distribution or any future date. Recipients should not rely on this material in making any future investment decision. The performance data quoted represents past performance. Past performance does not guarantee future results. Current performance may be lower or higher than the performance data quoted.