FAQ
How we run bare-metal GPU clusters, where they live, how they scale, and what working with us actually looks like.
Amaya Cloud was founded by system integrators and venture capitalists to provide reliable GPU infrastructure for our portfolio companies. Strong demand has led us to expand the platform for the broader market. We now rapidly grow our dedicated compute for startups, AI labs, research teams, and tech companies.
We come from system integration, so we focus on the layer we know best: dedicated compute and networking. Rather than competing with our customers at the orchestration or application layer, we act as their infrastructure partner. Bare metal provides better economics, stronger security and isolation, predictable performance and availability, and greater control.
Our primary infrastructure is in a Tier III, SOC 2 Type II audited data center in the San Francisco Bay Area, near our office and engineering team. We have a local operational presence, continually expand capacity, and partner with data centers to support reliable long-term growth.
We use dedicated, single-tenant infrastructure with isolated customer environments in enterprise-grade data centers. Our infrastructure is designed for customers that care about workload isolation, physical security, operational controls, and long-term infrastructure reliability.
Yes. We support multi-node GPU deployments with high-speed networking, including InfiniBand configurations for distributed training, large-model workloads, and other communication-intensive AI applications.
Yes. Many customers begin with a smaller dedicated deployment and expand as their models, workloads, or teams grow. We continuously expand our GPU fleet and can plan additional capacity around customer growth.
Yes. We can work with customers on custom GPU clusters, networking, storage, connectivity, security, and other infrastructure requirements that do not fit a standard configuration.
Yes. Qualified customers can run a proof of concept or benchmark their workload on Amaya Cloud infrastructure before scaling into a larger production deployment.
Yes. Our infrastructure team can work with customers transitioning from hyperscale cloud environments to dedicated GPU infrastructure, including compute, networking, storage, and deployment requirements.
No. We do not charge egress fees, allowing customers to move datasets, checkpoints, and model outputs without unexpected data-transfer charges.
Customers work directly with our infrastructure team for provisioning, hardware, networking, deployment, and infrastructure-related support. Our focus is on long-term infrastructure relationships rather than simply providing access to GPUs.
AMAYA CLOUD