TY NEXUSCOMMERCETalk to TY Nexus

GPU compute & servers

RTX 5090. B300. RTX PRO 6000. H200.
Rent the compute or purchase a server configured for your workload.

Public market references are not confirmed TY Nexus offers. Hardware configurations, availability and purchase prices require a written quote.

Rent compute.
Own the server.

Explore these GPU-based server configurations for rental or purchase. Exact builds, availability and commercial terms are confirmed by quote.

Illustrative open rackmount GPU server with fan-cooled PCIe accelerators

RTX 5090

GeForce · Blackwell

32 GB GDDR7Memory per GPU

A PCIe GPU-based server direction for AI development and graphics workloads. Define the CPU, memory, storage and GPU count for your build.

Rental market reference$1.19/ GPU hour

Server purchaseQuote to configuration

Illustrative high-density accelerator server with a tall ventilated rackmount chassis

B300

Blackwell Ultra

288 GB HBM3eMemory per GPU

High-memory accelerator systems for larger AI projects. Discuss the platform, interconnect, GPU topology and deployment requirements together.

Rental market reference$8.99/ GPU hour

Server purchaseQuote to configuration

Illustrative enclosed professional GPU rack server with brushed metal panels and front intake grilles

RTX PRO 6000

Blackwell · Server Edition

96 GB GDDR7Memory per GPU

Server Edition configurations for AI and visual computing. Specify the accelerator count, chassis, power and software requirements for your project.

Rental market reference$2.49/ GPU hour

Server purchaseQuote to configuration

Illustrative AI server with segmented fan intakes and a lower row of drive bays

H200

Hopper

141 GB HBM3eMemory per GPU

Memory-intensive AI and HPC server configurations. GPU form factor, host resources and connectivity are confirmed against your workload.

Rental market reference$5.29/ GPU hour

Server purchaseQuote to configuration

Runpod single-GPU references · October 14, 2026. Not whole-server rental rates or TY Nexus offers. Chassis and complete-system specifications vary by quote; images do not establish stock or a certified design.

Price source

Compare GPU rates

USD per GPU hour, not per multi-GPU server. Rental and purchase inquiries are available for every GPU-based configuration below.

Single-GPU hourly market benchmarks in US dollars
GPUGPU memoryReference rateRent or buy
NVIDIA A600048 GB$1.09/ GPU hourLambda
NVIDIA A100 PCIe40 GB$1.99/ GPU hourLambda
NVIDIA H100 PCIe80 GB$3.29/ GPU hourLambda
NVIDIA GeForce RTX 509032 GB$1.19/ GPU hourRunpod
NVIDIA B300288 GB$8.99/ GPU hourRunpod
NVIDIA RTX PRO 6000 Blackwell96 GB$2.49/ GPU hourRunpod
NVIDIA H200141 GB$5.29/ GPU hourRunpod

Lambda single-GPU instances and Runpod Pods · checked October 14, 2026. Extra storage, services, taxes and provider terms may change the total. Purchase prices are quoted separately.

Dedicated servers

Monthly hosted bare-metal rental benchmarks from OVHcloud US. These are not GPU servers or shipped-equipment lease prices.

Hosted dedicated-server monthly starting-price benchmarks in US dollars
Reference modelProcessorStarting rateInquiry
SYS-1Intel Xeon E-21366 cores / 12 threads$30.00/ monthDiscuss leasing
ADVANCE-1AMD EPYC 4244P6 cores / 12 threads$147.00/ monthDiscuss leasing
RISE-5AMD EPYC 741324 cores / 48 threads$199.00/ monthDiscuss leasing

Reference: OVHcloud US · checked October 14, 2026. Base advertised offers; configurations, locations, setup fees, tax and terms may differ.

View source

Compute budget

Estimate one GPU using the reference rates above. Storage, additional services and taxes are not included.

Enter 1–744 whole hours for one GPU.

Reference compute estimate$109.00

100 hours × $1.09 / GPU hour

Buying the hardware?

All featured GPU-based systems can be discussed for purchase as well as rental. Specify GPU count, CPU, RAM, storage and cooling for a complete server quote. Cabinets, cables and cooling equipment are quoted separately.

Request a hardware quote

Bring the workload.
Define the system.

Configuration, term, site requirements.
Start with what your project needs.

Discuss your project