● Per-second billing UPI / Card / Wallet No commitment required

Simple, honest pricing.
Billed to the second.

No hidden fees. No reserved instance lock-in. Pay exactly for what you use — down to 100ms. Top up your wallet and start in under 60 seconds.

Get Started Free →
⚡ GPU Instances
Simple on-demand pricing — pay for what you use. All nodes are physically in India — Bangalore, Hyderabad, Mumbai, Delhi, Pune, Chennai.
NVIDIA CUDA
AMD ROCm Beta
Tier GPUs included VRAM Price (on-demand)
StarterRTX 3060 · GTX 1080 Ti8–12 GB₹12/hr
Starter+RTX 3060 12GB12 GB₹15/hr
PlusRTX 4060 · RTX 20708 GB₹17/hr
Plus+RTX 2080 · RTX 3060 Ti8 GB₹18/hr
StandardRTX 4060 Ti · RTX 2080 Super8–16 GB₹20/hr
StandardRTX 3070 · RTX 50608 GB₹22/hr
Standard+RTX 3070 Ti · RTX 2080 Ti8–11 GB₹24/hr
Standard+RTX 4070 · RTX 3080 (10/12 GB) · RTX 5060 Ti10–16 GB₹26–28/hr
Pro−RTX 3090 · RTX 3080 Ti · RTX 507012–24 GB₹29–30/hr
ProRTX 4070 Ti · RTX 4070 Super12 GB₹33/hr
Pro MaxRTX 4070 Ti Super · RTX 5070 Ti16 GB₹37/hr
Pro Max+RTX 4080 · RTX 4080 SUPER · RTX 3090 Ti16–24 GB₹42/hr
UltraRTX 4090 24GB24 GB₹48/hr
Ultra−RTX 508016 GB₹44/hr
ExtremeRTX 509032 GB₹51/hr
Billing is per-second after the first minute. Simple on-demand pricing — pay only for what you use. All taxes are collected once, when you top up your wallet (18% GST on the top-up — ₹500 credit costs ₹590); hourly rates then simply debit your wallet with no further GST.
🔮 Serverless Inference API
Scale to zero when idle. Pay only when your model is actually computing. No minimum charge between requests.
⚡
Active compute
from ₹0.005/sec
You pay for the card that serves you, per compute-second: ₹0.005/s on a GTX 1660 SUPER, ₹0.0083/s on an RTX 4060 Ti, ₹0.0213/s on an RTX 5090. Billed only while a worker is handling your request.
💤
Idle cost (scale-to-zero)
₹0
Workers scale to zero after 10 minutes with no requests, and then you pay nothing. While a worker is held warm and ready it costs 20% of its active rate (₹3.60/hr on a 1660 SUPER, ₹15.30/hr on a 5090). Cold starts are never billed.
🧮
Example: 1000 req/day
₹25
1000 requests × 3 seconds on an RTX 4060 Ti = 3000 compute-seconds × ₹0.0083 = ₹25/day. Renting that same card around the clock would cost ₹480/day.
📊
Compared to renting the card 24/7
95% off
At 10% utilization, serverless costs ₹108/day vs ₹1,080/day for always-on. 96% savings for low-traffic endpoints.
Serverless endpoints run on Indian GPU nodes — lower latency for Indian users vs US-based serverless. Pre-built endpoints for LLM (vLLM + Llama 3), Whisper STT, Stable Diffusion XL, and custom Docker. Browse endpoints →
💾 Storage — your work follows you, not your machine
Everything under /workspace is saved to our Singapore object store every ~30 seconds and restored automatically — when you resume, when your instance moves to another GPU, and if the machine you were on goes offline. It is a sync, not a mounted drive: your files come back with you wherever you land.
📦
Workspace save & restore
Included
Saved up to 4 GB per instance — first 1 GB free, then ₹6/GB per month. Survives stop, resume, migration and node failure — restored before your container starts, so nothing races it. Not included: model weights, the models/ folder and files over 256 MB — re-download those after Resume.
🆓
Scratch disk
Free
Local disk for the container lifetime. Deleted on stop. Size varies by GPU node.
🧠
Model weights
Not counted
Keep weights and datasets in /workspace/models/ — they are not synced and do not eat your 4 GB. Pull them at start, or bake them into your image.
We do not offer mounted network volumes. Our GPUs are individually-owned machines on Indian consumer broadband, not a datacenter — a remote mounted drive would be slower than simply downloading the file. What we do instead is the thing a datacenter does not: if the machine running your job dies, your work is already saved and comes back on another GPU.
📦 Plans
Pay-as-you-go is the default — no subscription needed. Teams & institutes can add an org plan for shared billing, per-member spend limits, priority dispatch (skip the +25% charge on Business & up), and bulk provisioning. Set one up from the dashboard.
Free
₹0
Pay-as-you-go. 5 concurrent instances · standard queue · community support. This is everyone's default — just sign up.
Team
₹2,499/mo
For small teams — up to 10 members. Shared org wallet · per-member spend limits · usage dashboards · email support.
Set up a team →
Business
₹9,999/mo
For institutes & companies — up to 50 members. 🚀 Priority dispatch (no +25% charge) · bulk provisioning · cohorts & time-boxed grants · dedicated support.
Set up a team →
Enterprise
Custom
Custom seats & monthly fee · priority capacity · postpaid (pay within 30 days) option · GST B2B invoicing · dedicated account manager · priority support.
Contact sales →
🆓 Included at no extra cost
These features are free for every account — no paid tier required.
✓
API Keys (up to 10)Create named API keys for CI/CD, scripts, SDKs
✓
SSH Public Keys (up to 10)Upload once — auto-injected into every instance
✓
Webhooks (up to 10)HTTP callbacks on instance start, stop, billing events
✓
Email AlertsInstance ready / stopped / low balance notifications
✓
GST InvoicePDF invoice every billing cycle — claim ITC
✓
Secrets Store (up to 20)Encrypted tokens, decrypted into your container at launch — not as env vars, which the machine’s owner can read
✓
Team / Org AccountShared wallet, up to 5 members, sub-user permissions
✓
Budget AlertsEmail when spend hits your threshold
✓
JupyterLab in-browserOne-click Jupyter on any running instance
✓
Web TerminalSSH in your browser — no local SSH client needed
✓
Pre-built Docker templatesPyTorch, TF, JupyterLab, vLLM, SD, ComfyUI + more
✓
Startup Credits ₹150Free credits on GSTIN verification for registered companies
🎁 Discount programs
We actively support students and early-stage startups in India.
🎓
Student Bonus — 15%
Verified .ac.in or .edu.in email gets +15% bonus credits on every wallet top-up. ₹1,000 topped up = ₹1,150 to spend. Up to ₹3,000 lifetime bonus. Automatically applied at checkout.
🚀
Startup Credits — ₹150 Free
Register your GSTIN in account settings. GridShare verifies instantly via GST API and adds ₹150 free credits to your wallet — unlocked with your first top-up of ₹1,000+. One-time per company.
👥
Referral — ₹100 each
Share your referral link. When a friend makes a first top-up of ₹500+, you get ₹100 and they get ₹100. You can earn up to ₹2,500 lifetime by referring friends (25 × ₹100).
vs. AWS / Azure / GCP
The same PyTorch workload on GridShare vs Western hyperscalers. INR equivalent at ₹83/$.
GPUGridShareAWS / GCP equivalentYou save
RTX 4090 24GB vs datacenter A100 — different GPU class from ₹48/hr ₹305/hr (AWS A100 80GB, ~$3.67/hr) Fraction of the cost
24 GB VRAM GPU ₹38–45/hr ₹1,660/hr (GCP a2-highgpu-1g) 97% cheaper
RTX 3080 10GB (NVIDIA) ₹25/hr ₹830/hr (AWS g5.xlarge A10G) 97% cheaper
RTX 3060 12GB (entry) ₹12/hr ₹415/hr (Azure NCv3, T4) 97% cheaper
Jarvislabs RTX 4090 ₹48/hr ₹62/hr 27% cheaper
RunPod RTX 4090 (USD) ₹48/hr ₹61/hr (USD conversion + fee) 26% cheaper + India latency
All GridShare GPUs are physically located in India — 5–20ms latency from Indian cities vs 150–250ms from US-based RunPod/Lambda. Your data stays in India.
🌱 Greener by design. Every GridShare GPU is existing hardware put back to work — no new datacenter built, no new silicon manufactured. Reusing idle consumer GPUs carries a fraction of the embodied-carbon footprint of spinning up fresh cloud capacity.
📊 Platform fee structure
Transparent split between GridShare and the GPU providers who power the network.
Provider earns
70%
Of every rupee you pay. Providers earn ₹8.40–₹40.60/hr depending on GPU. Paid out daily/weekly/monthly to their verified bank account (UPI coming soon).
GridShare platform
30%
Covers: relay infrastructure, billing, monitoring, support, node agent updates, matching engine, and insurance for provider downtime.
Payout fee (provider)
FREE
Monthly payouts to your bank are free — we absorb the transfer cost (pays out from ₹500). Want it faster? A flat fee covers the transfer: weekly ₹5 (from ₹300), daily ₹10 (from ₹150). Charged only when a payout is actually sent; balances below the minimum roll over, fee-free.
Billing FAQ
When does billing start?
Billing is per second, and the meter starts when your app is actually reachable — not when the container starts. You are not charged while the image pulls and the app boots. There is no minimum billing period. Serverless bills per 100ms of active compute.
What happens if my wallet runs out?
When your wallet balance drops below ₹10 (or your custom budget alert threshold), we send an email alert. If balance hits ₹0, your running instances are gracefully stopped. Data on the local scratch disk is lost. Anything under /workspace is already saved to our storage every ~30 seconds and restored automatically when you start again — that is where to keep work you want to survive a stop.
Can I get a refund on unused wallet balance?
Yes. Your unused loaded balance is refundable to your original payment method, less the ~2% payment-gateway fee — or you can take it as wallet credit with no fee. Delivered compute and bonus credits (student, referral) are non-refundable. See our Refund Policy.
Are prices exclusive of GST?
Yes. All prices on this page are exclusive of 18% GST. GST is added at checkout. Indian businesses with a valid GSTIN can claim Input Tax Credit (ITC) on compute spend — the GST invoice is automatically generated every billing cycle.
Is there a free trial?
Every new account gets ₹25 free trial credit on email verification — spendable immediately, so you can run a real GPU before paying anything. Beyond that: student accounts get 15% bonus on every top-up (up to ₹3,000 lifetime), startups with GSTIN get ₹150 free (unlocked on a ₹1,000+ top-up), and referral credits give you ₹100 per friend whose first top-up is ₹500+. Minimum top-up is ₹100.
Can I set a spending limit?
Yes. Set a budget alert threshold in account settings — you'll get an email notification when your spend hits that amount. Useful for training runs where you want cost guardrails. Auto-stop on zero balance also acts as a hard cap.
Start for ₹100. Stop anytime.
No monthly fee. No contract. Top up your wallet and launch your first GPU instance in under 60 seconds. Student? Get 15% extra credits on every top-up.
Create Free Account → Browse GPUs