DISCLAIMER: GoldInvestors.news is not a registered investment, legal or tax advisor or broker/dealer. All investment/financial opinions expressed by GoldInvestors.news are from the personal research and experience of the owner of the site and are intended as educational material. Although best efforts are made to ensure that all information is accurate and up to date, occasionally unintended errors and misprints may occur.

Nvidia graphics processors have become the most sought after chips in artificial intelligence, pushing the company’s stock to another record and lifting its market value close to $6 trillion. Customers now have more ways than ever to secure access, but growing choice has created its own complications.

Companies can rent Nvidia GPUs from the giant cloud platforms operated by Amazon, Microsoft and Google, turn to neocloud providers such as CoreWeave, use online marketplaces or purchase the costly hardware directly. Nvidia expects $108 billion in revenue for the October quarter, an 89% increase from a year earlier.

The customer base is also broadening beyond the cloud infrastructure companies that have ranked among Nvidia’s biggest buyers for years. Five clients accounted for at least 10% of Nvidia’s accounts receivable in the July quarter, up from three in January, according to a filing.

SemiAnalysis counted 323 Nvidia GPU providers as of September, compared with 209 less than 11 months earlier. “You’re going to see a whole new crop of really, really exciting neoclouds with hundreds of billions of dollars backlog together,” Nvidia CEO Jensen Huang said at a Goldman Sachs technology conference in San Francisco last month.

Amazon, Google and Microsoft remain natural choices for large companies already spending tens of millions of dollars annually on cloud services. Since ChatGPT launched in 2022, businesses have increasingly used these hyperscalers to obtain GPUs for generative artificial intelligence workloads.

The largest providers also carry reputational weight with enterprise customers. Anthropic and OpenAI have committed to spending more than $500 billion between Amazon and Microsoft, which controlled 59% of the cloud infrastructure market in 2025, according to Gartner.

Gartner analyst Hardeep Singh said the hyperscalers benefit from more than 10 years of broad technology capabilities, though they do not always possess enough GPUs to satisfy enterprise demand. Amazon CEO Andy Jassy told analysts in July that the company would not be able to serve all the demand it expected during the year, and he anticipated the same dynamic in 2027.

That capacity shortage has opened the door for neocloud operators. Modal, which runs virtual sandboxes where artificial intelligence agents work separately from primary information technology environments, moved from hyperscalers to major neoclouds and now uses 25 providers, according to CEO Erik Bernhardsson.

Even the hyperscalers are purchasing capacity from neoclouds, with Google and Microsoft tapping CoreWeave despite competing with the company. Marc Boroditsky, chief revenue officer of Nebius, said some hyperscalers had approached his company about serving customers they could not accommodate when needed.

Reactor uses GPUs through both Nebius and hyperscalers, CEO Alberto Taiuti said. Data center location matters because the video generation startup wants content created by users to appear immediately, while Nebius supplies the specific chips, service, hardware and software Reactor requires at an attractive price.

Prominent neoclouds can demand upfront payments, while promised chips may take months to become available as providers raise financing and install equipment. CoreWeave would struggle to provide 10,000 GPUs to a new customer with one day of notice, and CEO Mike Intrator said in August that near term capacity was essentially sold out.

Smaller specialist providers can offer more flexibility, including bare metal GPUs that provide customers with greater control while requiring more technical management. Customers remain focused on two fundamental questions: when the processors will become available and how much they will cost.

Oracle is offering another route by allowing customers to bring their own GPUs. The software company has more debt than Amazon or Microsoft and a lower credit rating, limiting its flexibility to embark on a GPU spending spree, but it is willing to operate customer owned hardware.

OpenAI has committed to spending more than $300 billion with Oracle over five years, although it has not said anything about supplying its own GPUs. The arrangement could appeal to companies able to buy artificial intelligence chips but lacking sufficient power, data center space or skilled workers.

Large private arrangements are also emerging. SpaceX provided excess capacity through separate deals with Google and Reflection, agreed in April to supply Cursor with GPUs before buying the artificial intelligence coding startup for $60 billion, and reached a May agreement to rent GPUs to Anthropic for $1.25 billion per month through mid 2029.

Companies are still installing GPU servers inside their own data centers to balance capability and cost control. The hourly spot price for an Nvidia B200 GPU has more than doubled since March, while Dropbox and Everpure are among the companies using GPUs they control for internal computing needs.

DISCLAIMER: GoldInvestors.news is not a registered investment, legal or tax advisor or broker/dealer. All investment/financial opinions expressed by GoldInvestors.news are from the personal research and experience of the owner of the site and are intended as educational material. Although best efforts are made to ensure that all information is accurate and up to date, occasionally unintended errors and misprints may occur.