Neocloud Revenue Surged on AI Infrastructure Demand
Specialized GPU cloud providers have scaled past $25 billion in annual revenue as enterprises seek to bypass hyperscaler constraints.
Updated on Sept. 29, 2026 in Data Centers
Live Poll
Is now a good time to shift your organization's computing needs toward specialized cloud providers?
Neocloud providers are increasingly capturing enterprise AI workloads by prioritizing GPU capacity over general cloud services, with annual revenue exceeding $25 billion in 2025. This growth, evidenced by high-performance cluster deployment, signals a shift in how firms manage large-scale training and inference tasks.
Why it matters
Enterprises are migrating to these specialized cloud providers to bypass the queue times and infrastructure capacity constraints found at traditional hyperscalers. This shift forces a new competition for AI-ready data center capacity as the sector seeks to sustain dense infrastructure investments.
CoreWeave reported $2.6 billion in revenue for Q2 2026, a 112% increase year-over-year, while Nebius saw revenue reach $575 million in the August 2026 cycle, a 514% surge. These platforms support massive GPU clusters via frameworks like PyTorch and JAX, managed through Slurm or Kubernetes schedulers.
The players
CoreWeave
A GPU-specialized cloud provider that offers high-performance infrastructure for generative AI training and inference.
Nebius
An Amsterdam-based cloud platform providing scalable GPU infrastructure and AI services.
Nutanix
An enterprise cloud computing company building software-defined infrastructure and multitenant management tools.
The details
Neocloud providers operate by imaging nodes and attaching high-speed storage directly to GPU clusters, a configuration optimized for the compute-intensive nature of AI models. This architecture differs from general-purpose cloud services by stripping away overhead to focus purely on high-performance throughput. As companies like Nutanix move to introduce multitenant capabilities in April 2026, the focus remains on keeping AI training and inference workflows efficient and accessible.
Timeline
2025: Total annual neocloud revenue exceeded $25 billion.
March 2025: CoreWeave completed a Nasdaq listing.
April 2026: Nutanix announced future multitenant capabilities for AI services.
August 2026: CoreWeave reported Q2 2026 revenue of $2.6 billion, while Nebius reported $575 million.
The Tech Race
The rapid expansion of neoclouds marks a departure from the dominance of traditional hyperscalers, challenging their hold on enterprise AI compute. This competitive race now hinges on securing energy-sited capacity as sovereign AI requirements become a defining milestone for the next 12 to 24 months.
Organizations requiring intensive compute power can now access specialized GPU-focused platforms that offer an alternative to waiting for hyperscaler resources. Users should expect a continued focus on energy-sited and sovereign capacity models, which may influence cloud architecture decisions in the coming two years.
The takeaway
The neocloud sector has proven that high-performance AI infrastructure can thrive outside of the traditional cloud giants. Watch for whether upcoming multitenant capabilities and increased sovereign capacity will allow these providers to retain their revenue growth trajectory into late 2027.
Further reading
For more on infrastructure scale and industry trends, visit the Data Centers section.
Source note: This article includes information reported by CloudTweaks.
Live Poll
Is now a good time to shift your organization's computing needs toward specialized cloud providers?
