AI training
Train and fine-tune models across eight GPUs. Validate framework support and inter-GPU requirements for your workload.
TRAINING · FINE-TUNINGDEDICATED GPU INFRASTRUCTURE
High-performance
GPU compute.
Solar-backed.
Eight RTX PRO 6000 Blackwell GPUs. One dedicated baremetal server. Direct contracting and a software environment aligned with your workload.
SC-BW6000-8BAREMETAL
8×
RTX PRO 6000
Blackwell
96 GB VRAM per GPUDedicated hardware
Capacity for demanding workloads.
Dedicated baremetal8 × Blackwell GPUs
Solar-backed infrastructureOwn photovoltaic capacity
Direct contractingConfiguration reviewed with you
01 / THE SERVER
One server. Serious capacity.
Dedicated hardware.
Select the commitment that fits your project.
Baremetal Server
RTX PRO 6000 Blackwell
SolCore Blackwell 8 / SC-BW6000-8
8 × 96 GBVRAM · per GPU, not unified
1,024 GBECC system memory
- GPU
- 8 × RTX PRO 6000 Blackwell 96 GB
- CPU
- 2 × 32 cores · AMD EPYC Turin 9355
- RAM
- 1,024 GB ECC
- Storage
- 15 TB NVMe SSD · TLC
- Operating system
- Ubuntu 24.04 LTS
- Internet
- 1 Gbit/s · IPv4 address
- Location
- Spain
Advertised configuration supplied for this offer. Availability and deployment details require confirmation.
SELECT YOUR COMMITMENT
€13,251.55/ month
1-month commitment
Equivalent to €17.81 per hour
Dedicated baremetal server Ubuntu 24.04 LTS 1 Gbit/s connectivity & IPv4
Hourly equivalents use a 744-hour comparison month and are rounded. This is a commitment plan, not hourly rental. Taxes, payment schedule and final terms are confirmed in your proposal.
02 / AN INFRASTRUCTURE ADVANTAGE
Compute first.
Solar at its foundation.
Our own photovoltaic generation supports the compute infrastructure in Spain, reducing reliance on grid electricity. Energy is part of the operating model, alongside dedicated hardware and direct coordination.
530.82 kWp
Installed photovoltaic capacity
Confirmed installed capacityInstalled power is not energy produced. It does not guarantee continuous renewable coverage. Production and coverage claims require verified measurements.
03 / BUILT FOR YOUR WORKLOAD
From models to production.
96 GB of VRAM per GPU.
A dedicated environment for demanding jobs.
AI inference
Run memory-intensive models and parallel inference workloads with 96 GB of VRAM per GPU.
LLM · INFERENCE3D rendering
Process GPU render batches and complex scenes with an environment matched to your renderer.
RENDER · PRODUCTIONSimulation
Execute CUDA-enabled numerical workloads and parallel simulations on dedicated hardware.
CUDA · SIMULATIONComputer vision
Train vision models and process image or video datasets with reproducible software dependencies.
VISION · DATASETSResearch workloads
Run experiments on a defined baremetal configuration, with software and access agreed before deployment.
RESEARCH · EXPERIMENTS04 / DIRECT COORDINATION
From plan to deployment.
Own infrastructure, operated from Spain. We review the workload and agree the environment and commercial terms directly with you.
- 01
Select your plan
Choose a monthly, half-yearly or yearly commitment.
- 02
Request availability
Describe your workload and preferred deployment timeline.
- 03
Review and confirm the setup
We validate availability, software requirements and terms.
- 04
Receive access details
After confirmation, we provide credentials and deployment details.
05 / BEFORE YOU COMMIT
Technical details.
Clear expectations.
What is included with the server?+
The advertised configuration includes eight RTX PRO 6000 Blackwell 96 GB GPUs, dual 32-core EPYC Turin 9355 CPUs, 1,024 GB ECC RAM, 15 TB TLC NVMe storage and 1 Gbit/s connectivity with IPv4. Availability and final terms are confirmed before contracting.
Is this a baremetal server?+
Yes. This offer is for a dedicated physical server. No virtual server option is offered on this page.
What operating system is provided?+
Ubuntu 24.04 LTS is proposed. Tell us your CUDA, framework and dependency requirements so we can review the environment.
Can I request a custom setup?+
Specify software and workload requirements in your request. We review compatibility with the advertised configuration before confirming the setup.
Is IPv4 included?+
An IPv4 address and 1 Gbit/s connectivity are part of the advertised configuration. Usage policies and any limits will be confirmed in the proposal.
How does the solar-backed infrastructure work?+
Our 530.82 kWp of installed photovoltaic capacity supports the infrastructure and reduces reliance on grid electricity. Installed power is not energy production and does not guarantee continuous renewable coverage.
Can I choose monthly, half-yearly or yearly billing?+
Request a 1-, 6- or 12-month commitment. Prices are expressed per month for comparison. Payment schedule, taxes and renewal conditions are confirmed in the proposal. Hourly figures are comparison equivalents, not hourly rental plans.
How is deployment handled?+
Deployment is coordinated directly after availability, setup and terms are confirmed. We provide credentials and deployment details; there is no instant automated provisioning.
Is support included?+
Support scope, response times and any service-level commitments are not yet published. They will be defined in your proposal before you commit.
06 / REQUEST YOUR SERVER
Tell us about
your workload.
RTX PRO 6000 Blackwell. Eight GPUs, dedicated to your project. Select your plan and describe the software and deployment requirements.
Monthly€13,251.55 / month1-month commitment · Subject to availability
Request preview
This form prepares a local draft you can copy or download. Submission is not connected yet. No request is sent or stored.
CORPORATE CONTACT
The right contact.
Contact us for general and sales enquiries, privacy or support.
General / Sales / Privacy
General enquiries, sales projects and privacy rights requests.
info@solcorecompute.comSupport
Customer enquiries. Scope and hours pending.
support@solcorecompute.com