DEDICATED COMPUTE. ZERO COMPROMISE.

Small footprint.
Big intelligence.
Your Spark.

Your own NVIDIA DGX Spark. Full root access.
One machine or a private cluster — the power to build
what comes next, without owning the hardware.

Explore the hardware ↓
From ₹45/ Spark / hour·Billed in INR
✓ No shared GPUs✓ SSH as root✓ Prepaid, no lock-in
BARE-METAL / GB10 SUPERCHIP YOUR PRIVATE FABRIC
NVIDIADGX SPARK
01
NVIDIADGX SPARK
02
NVIDIADGX SPARK
03 — 06
cluster-01 2 × SparkPrivate VLAN · RoCE
cluster-02 4 × SparkSeparate tenant fabric
root@spark-01 BARE METAL
~ $ nvidia-smi
NVIDIA GB10 · 128 GB unified memory
~ $ ▌
01 / 06 NODESBUILT FOR YOUR NEXT BREAKTHROUGH ↗
THE SUPERCHIPNVIDIA GB10Grace Blackwell architecture
UNIFIED MEMORY128 GBPer dedicated Spark
AI PERFORMANCE1 PFLOP*FP4 sparse theoretical peak
YOUR ENVIRONMENT100% yoursFull system. Full root access.

NOT A SLICE. THE WHOLE MACHINE.

Less infrastructure.
More possibility.

From your first inference to multi-node experiments.
Real hardware, a private fabric, and nothing
between you and your next idea.

⌘
01

Root means root.

Install your stack. Change your kernel. Run containers, fine-tune models, or start from scratch. No GPU sharing. No arbitrary software restrictions.

ssh root@your-spark↗
▧
03

Your cluster. Your boundary.

Separate Ethernet and RDMA VLANs, enforced on MikroTik switches. Hosts are sanitized between rentals, not just given a new password.

Dedicated tenant network↗

SERIOUS COMPUTE. SENSIBLE PRICING.

Your ambition.
Your cluster.
Your budget.

No dollar surprises. Configure your cluster
and see exactly what you reserve in rupees.

01

First Spark at the base rate.
Each additional Spark costs 20% more than the previous one.

02

Prepay for 1–24 hours.
Stop early. Unused time is refunded.

A little more about billing ↗
CONFIGURE YOUR CLUSTERLoading capacity…

A little Spark goes a long way.

Choose how much power you need.

128 GB unified memory20 CPU coresDedicated system
First Spark₹45.00 / hr
Additional Sparks—
YOUR HOURLY RATE₹45.00 / hour
INR

Loading pricing…

FROM IDEA TO TERMINAL

Your workflow. Upgraded.

01 / CONFIGURE

Pick your power.

Choose your Sparks, reserve your time, and add an SSH public key. Your quote is locked when you book.

02 / CONNECT

Make yourself at home.

Once isolation and provisioning checks pass, connect as root using your private key. The entire machine is yours.

03 / CREATE

Build without limits.

Train, infer, experiment. Stop when you’re done. Access is revoked and hardware is sanitized before its next rental.

GOOD QUESTIONS.

A few things
worth knowing.

Is this a virtual machine?

No. Each rental reserves dedicated physical DGX Spark systems, including their CPU, GPU, memory, and local disk. Your root access is not a container shell.

Can two customers share a six-Spark installation?

Yes. A two-node and a four-node cluster consume all six machines. They use distinct access and RDMA VLANs enforced by managed switches. Tenant gateways must also deny cross-tenant routing, including via public IPs.

Do Sparks connect with NVLink?

No inter-machine NVLink is exposed on DGX Spark. Clustering uses ConnectX-7 Ethernet with RoCE. NVIDIA specifies up to 200 Gb/s per QSFP port; actual bandwidth depends on the switches, cables, and validated configuration. Multi-node memory is not one transparent shared pool.

How do pricing and early stops work?

Rates are in INR using an administrator-configured USD base price and reference exchange rate. Each additional Spark costs the configured percentage more than the previous Spark. With a 20% premium and a base rate of ₹100, individual machines cost ₹100, ₹120, ₹144 and ₹172.80 per hour; a four-Spark cluster totals ₹536.80. Cluster totals are rounded once to the paise. Reservations are prepaid; unused time is refunded after verified access revocation. Applicable taxes are extra. No automatic card charging is enabled.

What happens to my files?

Local storage is ephemeral. Back up your work before stopping or before your reservation expires. Hardware stays quarantined until the external reset process has verified sanitization. Root access does not permit abuse, attacks, or access to other tenants or provider infrastructure.

THE NEXT BIG THING STARTS SMALL.

Give your idea a Spark.

YOUR COMPUTE, YOUR RULES

My clusters.

A dedicated home for your next big idea.

Fund your wallet

Recent payment attempts

If checkout closes or your connection fails, check the existing payment before paying again. Captured live payments fund your wallet once; test payments never do.

DateAmountStatusAction

Your infrastructure

Updates every 4 seconds
WELCOME TO GB10.IN

Your next idea starts here.

Create an account to configure your first cluster.

YOUR NEXT BREAKTHROUGH

Make it your cluster.

Never share your private key.

Metering begins only after provisioning checks pass. Automatic stop at expiry. Unused time refunded after verified access revocation.

ACCOUNT SECURITY

Confirm your email address.

Confirm that this email belongs to you. This does not sign you in or change your password.

MACHINE ONBOARDING

Spark recovery profile

Use discovery on the actual Spark for disk serial, exact capacity, wired MAC and both netdevs per QSFP port. Editing decommissions this node.

OPERATOR CONTROLS

Spark operation

Reimage permanently destroys all local data. Stop rented clusters first. Failures quarantine the node and require inspection or USB recovery. Simulation never touches hardware.

One-use enrollment

Expires in 15 minutes. This replaces any old node credential. Save the token into a root-readable file on the Spark, enroll, then delete it. Do not put it in shell history or images.