Glossary
neocloud
A specialized cloud provider focused exclusively on GPU and AI workloads, operating outside the traditional AWS/Azure/GCP hyperscaler perimeter, with CoreWeave, Lambda, and Voltage Park as the canonical examples.
A cloud provider that sells GPUsiliconA massively parallel processor originally designed for graphics, repurposed since the 2010s as the dominant compute substrate for both training and inference of large neural networks. Open full entry compute as its primary or only product. Distinct from hyperscalers (AWS, Azure, GCP) in scope and customer base: neoclouds typically have no general-purpose VM tier, no managed databases, no enterprise SaaS, just GPUs by the hour or reserved instance. CoreWeave, Lambda, Voltage Park, Crusoe, Hyperbolic, Together AI, and Fireworks are commonly grouped under the label.
The economics depend on chip access. NVIDIA allocates Hopper and Blackwell GPUs to neoclouds on terms that vary; the most-favored operators (CoreWeave is the public example, with NVIDIA investments since its April 2023 Series B) get earlier and larger allocations than smaller peers. This creates a tier system within the neocloud category.
For open-source AI, neoclouds matter as the practical compute layer for teams that cannot buy capacity from a hyperscaler at acceptable terms but cannot self-host either. The sovereignty profile varies: some are hyperscaler-adjacent (CoreWeave hosts large Microsoft workloads); others target the independent-developer market (Hyperbolic, Voltage Park).