AI Demand Pushes Neoclouds into the Channel Conversation

Neocloud providers are gaining momentum as enterprises look for more GPU capacity to support AI training, fine-tuning, and inference.

For MSPs, resellers, and system integrators, that shift could expand the cloud infrastructure conversation beyond hypescalers. As AI workloads push customers to compare every available option, partners have an opportunity to advise on cost, capacity, workload placement, and deployment complexity.

Neoclouds compete on GPU capacity and flexibility

Even though neoclouds existed before ChatGPT’s launch, the sector’s growth has come hand in hand with AI. 

Quite a few began life as cryptocurrency miners, giving them a head start in the AI infrastructure market because they already understood power procurement, chip-heavy facilities, cooling, and the economics of running high-density compute at scale. 

The GPU-first design of most neoclouds also makes them efficient for AI developers’ tasks, such as model training, fine-tuning, and inference.

These neocloud companies, including CoreWeave, Crusoe, Lambda, Nebius, Nscale, and others, are positioning themselves as the specialist layer for enterprises that need high-performance AI infrastructure.

The size of neoclouds is minuscule compared to that of hyperscalers such as AWS, Microsoft, and Google Cloud, but that can work to their advantage. Their smaller scale can make them more flexible in handling special customer requests. 

The price per hour can also be less than half that of comparable hyperscaler capacity, according to research cited by STL Partners, making them attractive to AI developers who primarily need raw, high-performance compute.

Neoclouds and hyperscalers race to support demand

Even though these neoclouds are signing major contracts, the hyperscalers are still securing a lot of the AI customers. AWS, Microsoft, and Google have sharply increased their capex plans for 2026, with the three expected to spend around $570 billion this year. The majority of this will go toward cloud infrastructure.

Once, or if, they reach a point where supply meets demand, there may be a shift back toward the hyperscalers, as they can offer a wider range of tools, services, and applications. 

All three have their own AI operations, with Google CEO Sundar Pichai acknowledging that Google Cloud’s revenue is lower than it could be due to capacity constraints linked to the company’s own AI products.

Neocloud market expands as hyperscalers and AI developers seek capacity

These capacity constraints have led both Microsoft and Google to rent capacity from neoclouds. 

Microsoft signed a five-year deal with IREN in November last year, indicating that it expects capacity constraints to persist for some time. CoreWeave said in June 2025 that it had signed Google as a customer, reportedly set up as part of the OpenAI-Google deal.

Even some AI developers are getting into the neocloud business. Anthropic recently bought all of the compute capacity at xAI’s Colossus 1 data center, providing the Grok chatbot maker with billions of dollars in revenue. 

xAI reportedly burned $7.8 billion in the first nine months of 2025, according to Bloomberg, so reducing the net loss could be timely ahead of a potential SpaceX IPO in the summer.

In this high-demand market, neoclouds have carved out an important role, and it does not look like this will change any time soon. 

According to Synergy Research, neocloud market revenue increased from $9 billion in 2024 to $25 billion in 2025. The research firm estimates that revenues will be close to $400 billion by 2031.

What neocloud growth means for channel partners

This creates both a disruption and an opening for partners. Resellers, MSPs, and system integrators that have built cloud practices around the hyperscalers’ many offerings may find AI buyers asking different questions. 

Instead of simply asking which cloud to use, customers may focus more on the right infrastructure stack, how to control AI training costs, whether their data pipelines are ready for AI workloads, and how to manage workloads across hyperscale, neocloud, and on-premise environments.

This opens the door to more conversations and optimizations, as partners can help customers fine-tune infrastructure choices and buildout plans. 

Partners can help customers avoid AI infrastructure complexity

At the same time, customers are still likely to run some of their core applications on AWS or Azure while using a neocloud provider for experimentation or high-performance workloads.

For channel partners, the rise of neoclouds is less a threat to traditional cloud practices than an expansion of the AI infrastructure portfolio.

The partners best positioned to benefit will be those that can explain when neoclouds make economic sense, how they fit alongside hyperscaler and on-premises environments, and how customers can adopt them without creating unnecessary operational complexity.