Frequently Asked Questions

Axe Compute (NASDAQ: AGPU) provides enterprise-grade, bare metal GPU infrastructure for AI workloads. We deliver dedicated NVIDIA GPU clusters through our access to the Aethir decentralized global compute network, serving teams that need high-performance hardware for training, inference, and research.

We offer a range of NVIDIA GPUs including H100, A100, L40S, and other models depending on availability. Our inventory is updated in real-time — browse the GPU Catalog for current offerings, specs, and pricing.

Bare metal. You get direct hardware access with zero virtualization overhead. This means full CUDA core, tensor core, and NVLink utilization for maximum training and inference throughput.

Pricing is per GPU per hour. We offer month-to-month and committed terms (up to 36 months) with volume discounts available for longer commitments. All pricing is transparent with no hidden fees. Committed terms may require a prepayment.

We offer flexible terms ranging from month-to-month to 3, 6, 12, 24, and 36-month commitments. Longer terms come with discounted rates. Contact our sales team to discuss custom arrangements for large deployments.

We have GPU infrastructure across multiple global regions. Available regions depend on current inventory — check the GPU Catalog to see what's available in your preferred location.

Many of our offerings include high-speed InfiniBand or RoCE interconnect for distributed training workloads. Interconnect type and speed vary by offering and are listed on each GPU card in the catalog.

Browse our GPU Catalog to see available inventory, select a configuration, and proceed to checkout. For custom requirements, large deployments, or questions, contact our sales team directly.