posted by Neuralrack AI 2 months ago

$0.48 / gpu / hour - 64 RTX 5090s available from 03/01/2026 to 08/31/2027 (FL, US)

Unverified
Interconnect: Ethernet 100GbE
Cores per node: 64
RAM per node: 384 GB
NVME storage per node: 9600 GB
Location: FL, US
Cloud provider: Neuralrack AI
Cluster interface: Bare metal SSH (deploy your own clustering software)
Minimium bookable GPUs: 8
Minimium bookable duration: 4 weeks
Additional details
Neuralrack AI: 3 years of AI infra experience, 1000+ GPUs deployed in total, redundant BGP 10Gbps internet (upgradable), and tier 3 DC partners (SOC2/ISO27001/HIPAA). All equipment fully insured. Blackwell available on neuralrack.ai in March/July 2025, 60 days after official launch.
*List price of $0.484/hr is for a 12-month contract, paid monthly, with standard specs. Month-to-month is $0.57/hr. Customize your specs/commitment here: https://neuralrack.ai/gpu-server-pricing*
We have up to 64 RTX 5090 32GB cards available for rent in 8x bare metal configurations or with virtualization. Pay $0.29-$0.57/GPU/hr based on rental period and specs. Save with long-term commitments, upfront payment or ownership. Bulk pricing available on request: https://neuralrack.ai/contact
All servers are in-stock, deployable in 48 hours. Locations: Tampa, FL. Accepted payment methods: ACH, wire, or crypto.
Standard specs for 8x5090 (upgradable): 384GB DDR5 4800 ECC memory, 2x AMD EPYC 9354 32-core CPUs, 1x 1.92TB Gen4 NVMe boot + 2x 3.84TB Gen4 NVMe data storage, 100GbE / 25GbE local network, all-way PCIe Gen5 x16 (100GB/s per GPU)
Custom clusters available (RTX Pro 6000 96G, RTX Pro 5000 48/72G, RTX 5090 32G, RTX 5080 16G, B200 180G, H200SXM 141G, H200NVL 141G, MI325X/MI355X 288G, and more) at https://neuralrack.ai
contact seller

seller email domain is neuralrack.ai