Nvidia GPUs have become the most sought-after processors in the AI industry, driving the company's stock to a record high and pushing its market cap close to $6 trillion. The demand for these chips is so intense that companies are exploring various avenues to access them, from cloud providers like Amazon, Microsoft, and Google to emerging 'neoclouds' such as CoreWeave. Nvidia's revenue is projected to hit $108 billion in the October quarter, an 89% year-over-year increase, highlighting the unrelenting growth in the sector.
While hyperscalers like Amazon, Google, and Microsoft have long been the go-to providers for AI workloads, the landscape is diversifying. Industry research firm SemiAnalysis identified 323 Nvidia GPU providers as of September, up from 209 just 11 months earlier. Nvidia CEO Jensen Huang anticipates the rise of new neoclouds with significant backlogs, offering companies more options to meet their computing needs.
Hyperscalers offer a trusted reputation, which is crucial for enterprises. However, they often struggle to meet the high demand for GPUs. Companies like Anthropic and OpenAI have committed over $500 billion to Amazon and Microsoft, but even these giants admit they cannot fulfill all the demand. This gap is being filled by neoclouds, which provide more flexible and immediate solutions. Startups like Modal and Reactor are turning to these smaller providers for better service and quicker access to the necessary hardware.
For companies needing GPUs immediately, smaller or specialized neoclouds offer an alternative. These providers often require upfront payments and long-term commitments, but they can deliver the specific GPUs and customer service that larger clouds cannot. Additionally, Oracle is allowing clients to bring their own GPUs, a tactic that could appeal to companies with the capital but lacking the infrastructure or skilled labor to manage their own data centers.