Getting Started: Requesting your first cluster
Capacity is currently open for reservation. Provisioning timelines are confirmed at contract.
AlphaSolutions IT will GPU clusters on request. There is no self-service console. All cluster requests are handled by our engineering team, who will confirm hardware availability, configure your network environment, and hand over credentials once your cluster is commissioned, to the timeline confirmed in your order form.
To request your first cluster, contact us at [email protected] with the following information: the GPU model you require, the number of nodes, your preferred billing model (on-demand hourly or reserved), and your target start date. If you have specific networking requirements, such as a dedicated VLAN, a private interconnect to an existing environment, or a particular IP range, include those details in your initial message.
Once your request is received, our team will send a formal quote. The quote will specify the hardware configuration, per-hour or monthly rate, and the estimated provisioning timeline. Reserved cluster contracts include a minimum term of six months and are subject to a signed order form before provisioning begins.
After the order is confirmed, we will provision your cluster and send you SSH access credentials, the management IP range, and any relevant network configuration details. For multi-node clusters, we will also provide the InfiniBand or RoCEv2 fabric topology so you can configure your distributed training framework accordingly.
All clusters run on bare-metal hardware. There is no hypervisor layer between your workload and the GPU. You receive root access to each node and full control over the software stack, including the choice of CUDA version, container runtime, and job scheduler. Our engineering team supports customers through provisioning and initial configuration.
