Skip to content

NVIDIA DGX Spark · GB10 Blackwell

128 GB of unified memory, from $0.59 an hour.

Rent a DGX Spark to prototype and serve open-weights models — full root SSH and Docker, CUDA 12.8, billed per second. Cluster up to four units for 512 GB unified, then step up to the RTX PRO 6000 fleet when you scale.

$0.59 / hr, from128 GB unified memoryRoot SSH + DockerTerminate anytime
root@spark — 128 GB unified
# full root · Docker · CUDA 12.8
$ nvidia-smi --query-gpu=memory.total --format=csv,noheader
131072 MiB
$ docker run --gpus all -p 8000:8000 vllm/vllm-openai
INFO serving on :8000 (OpenAI-compatible)
$ curl https://calm-node-2214.skyforgecompute.com/v1/models \
    -H "Authorization: Bearer $SKYFORGE_API_KEY"

01Hardware

One box, sheet of record.

Unified memory big enough for serious models — run up to roughly 200B parameters without sharding, with full root.

SpecValueNotes
ComputeGB10 BlackwellSM 10.0 · NVFP4 · Transformer Engine 2.0
Unified memory128 GB @ 273 GB/scoherent CPU+GPU memory on one box
Storage4 TB NVMelocal scratch for datasets and checkpoints
Accessroot SSH + DockerCUDA 12.8 — your stack runs as-is
Model headroom≤ ~200B paramson a single box, without sharding
Clusteringup to 4 units512 GB unified across the cluster

From $0.59 / hr per instance · billed per second

Shapes & pricing →

02Shapes & pricing

One Spark, or a 512 GB cluster.

Priced per instance, per hour, billed per second. No commitment, no tier gates — launch from the console.

ShapeUnified memoryPriceNotes
1× Spark128 GB unified$0.59/ hrone GB10 box — the cheapest way onto the fleetLaunch →
2× Spark256 GB unified$1.10/ hrtwo linked units — a step before the full clusterLaunch →
4× Spark cluster512 GB unified$2.10/ hrfour linked units — bigger models, parallel jobsLaunch →

Priced per instance, per hour — billed per second from boot to terminate. Root SSH + Docker, CUDA 12.8 on every shape.

Own a DG1 workstation →

base_url = https://<your-slug>.skyforgecompute.com/v1 · api_key = one SkyForge key, every instance

03How it compares

The cheapest way onto Blackwell.

A Spark is a dev box, not a training cluster — for bigger serving or training, step up the fleet.

OptionMemoryPriceBest for
DGX Spark128 GB unified$0.59 / hrdev, prototyping, ≤ ~200B models
RTX PRO 600096 GB / GPU$2.25 / GPU-hropen-weights inference
B200 / B300192 GB-class HBMbuilt to orderlarge-scale training

A Spark is where you prototype — step up when throughput matters.

RTX PRO 6000 →

04What runs on it

Ready-to-run, from the first minute.

01

Open frontier models

GLM, DeepSeek, Qwen 3, MiniMax, and Llama — served behind your private, OpenAI-compatible endpoint.

02

Inference server

vLLM, SGLang, TGI, or Ollama behind an OpenAI-compatible endpoint.

03

Fine-tuning

Axolotl, Unsloth, or LLaMA-Factory with datasets staged on local NVMe.

04

Agent harness

Tooling and orchestration for agentic workloads and high request fan-out.

05

Dev environment

JupyterLab or VS Code Server with the CUDA toolkit, ready to code.

06

Embeddings & RAG

Generate embeddings and back retrieval pipelines from the same endpoint.

05FAQ

Questions, answered.

Billing, model headroom, and when to step up — the short version.

06 — When you outgrow it

Start on a Spark. Scale on the fleet.

Prototype at $0.59 / hr, then move the same endpoint workflow to RTX PRO 6000 at $2.25 / GPU-hr. Want to own one? A DG1 workstation on a one-year term is an enterprise order.