Choose RunPod if…
- You want to launch from a console with a card, no quote.
- You want per-second billing and will run your own operations.
- You run serverless endpoints that scale to zero between requests.
- You need GPUs in several countries at once.
RunPod bills by the second in dollars. We rent the card by the month in rupees, managed if you want.
By Arun Bansal · RunPod facts checked 13 Oct 2026
RunPod figures from its public pages (Secure Cloud), checked 13 Oct 2026, excluding tax. Rechecked quarterly.
Businesses we’ve hosted and supported since 2009
RunPod is a self-serve GPU cloud. You pay by the second in US dollars, launch from a console with no commitment, and pick from 31 regions, one announced in India.
ZenoCloud rents you a dedicated NVIDIA GPU with its host, by the month, hosted in India, on an INR invoice with GST. Managed deployment and 24/7 care are priced add-ons.
| Compare on | RunPod Secure Cloud, per second, 13 Oct 2026 | ZenoCloud Dedicated GPU by the month, hosted in India |
|---|---|---|
| What you rent | A Pod: a GPU container billed by the second. Secure Cloud, the tier compared here, “operates in T3/T4 data centers” (RunPod’s words) | A whole NVIDIA GPU card with its host, yours for the month |
| Billing | Per second. Standard accounts prepay credits from $10; post-paid invoicing for enterprise and larger accounts via sales | Monthly, one-month minimum, on a written quote. No hourly plan |
| Currency and tax | US dollars. Digital sales tax “in certain jurisdictions”. GST not found on the pages we checked | INR. GST applies; input tax credit depends on your tax status and applicable law |
| H100 80GB PCIe | $2.89 an hour, Secure Cloud | From ₹1,80,000 a card a month. On request |
| H200 141GB | $5.29 an hour, Secure Cloud | From ₹2,20,000 a card a month. Ready now |
| L40S 48GB | $1.09 an hour, Secure Cloud | From ₹55,000 a card a month. Ready now |
| RTX PRO 6000 96GB | $2.49 an hour, Secure Cloud | From ₹1,10,000 a card a month. Ready now |
| A100 80GB PCIe | $1.79 an hour, Secure Cloud | From ₹99,000 a card a month. On request |
| Storage | Metered per GB a month by disk type and state, $0.05 to $0.20 | Capacity and any charge are listed in the quote |
| Data transfer | “no fees for data ingress or egress” | Network listed in the quote |
| Where the GPU is | 31 regions, chosen at deploy. AP-IN-1 in India announced April 2026, H100 focus; city not stated | India. The quote names the location |
| Who runs it | You, from their console, templates and SSH | You. Or we deploy (from ₹50,000) and run the host (Care from ₹25,000 a node a month, ₹50,000 account minimum) |
| How you reach them | Help centre and email for customers; phone and WhatsApp for sales. Hours not stated | Our 24/7 operations desk |
| Trying it | Start with $10 of credits | A short paid trial node. Keep it and it becomes production |
RunPod: Secure Cloud rates from its pricing page and docs, read 13 Oct 2026, US dollars, before tax. Ours: list prices in INR, “from”, taxes extra; the quote confirms price and availability.
Not us: hourly or burst GPUs.
For those we run AWS GPU instances in your AWS account. Managed AWS.
Illustrative only: the quote confirms GPU variant, host and price, and deployment or care cost extra. Our H100 80GB is from ₹1,80,000 a card a month.
RunPod’s H100 PCIe Pod on Secure Cloud was $2.89 an hour on 13 Oct 2026 (188 GB RAM, 16 vCPUs). Here is where the lines cross on an assumed 730-hour month.
crossover hours = our monthly price ÷ (RunPod hourly rate × INR per USD)
= 1,80,000 ÷ (2.89 × 95.78)
= 1,80,000 ÷ 276.80
= 650 hours a month (650.3 unrounded)
Inputs: ₹1,80,000 (ZenoCloud price book, 14 Oct 2026, “from”); $2.89 an hour (RunPod pricing page, H100 PCIe, Secure Cloud, 13 Oct 2026); USD 1 = ₹95.78 (ECB reference rate published 12 Oct 2026).
Run the card more than about 650 hours a month, roughly 21 hours a day, and our monthly price is lower. A full 730-hour month on RunPod costs ₹2,02,067.
Below that, RunPod’s metered price is lower.
The other cards: RunPod Secure Cloud rates as on 13 Oct 2026; crossover hours on a 730-hour month
| Card | Our price (from) | RunPod Secure rate → crossover hours (% of 730) |
|---|---|---|
| H200 141GB | ₹2,20,000 | $5.29 → 434 h (59%) |
| RTX PRO 6000 96GB | ₹1,10,000 | $2.49 → 461 h (63%) |
| L40S 48GB | ₹55,000 | $1.09 → 527 h (72%) |
| A100 80GB PCIe | ₹99,000 | $1.79 → 577 h (79%) |
| H100 80GB PCIe | ₹1,80,000 | $2.89 → 650 h (89%) |
Crossover = our price ÷ (rate × 95.78), computed from unrounded inputs. Above the hours shown, our monthly price is lower; below, RunPod’s metered price is lower.
RunPod also sells a cheaper Community Cloud tier, run on individual providers’ machines; we compare data-centre tier to data-centre tier.
RunPod lets you choose among its 31 regions and announced AP-IN-1 in India in April 2026, focused on H100s. Its compliance page says coverage varies by region; ask them.
RunPod Inc. is a US company. Charges are in dollars, with sales tax in some jurisdictions. India is on its tax-ID list; GST was not found on the pages checked.
With us the card and host are yours for the month, in India, location named in the quote. We describe where data flows so your compliance team can document it.
You get an INR invoice with GST. India hosting alone does not make you DPDP compliant; your policies, consent handling, access controls and audits still matter.
Pods run containers. We test your container and its dependencies on a paid trial node before agreeing the migration steps. Keep RunPod running until acceptance passes.
GPU, container image, volume size, hours a month, timing. Quote within one business day.
A short paid trial. You run your workload, then decide. Keep it; it becomes production.
Runtime, drivers, container, serving framework, observability, closed by an agreed test. From ₹50,000.
Agree storage and transfer steps first, then move your volumes to the quoted host storage.
Run it yourself, or add GPU Infrastructure Care, monitored 24/7.
Care is from ₹25,000 a node a month with a ₹50,000 account minimum. Deployment is a one-time project.
For steady, month-long workloads, yes. We rent dedicated NVIDIA GPUs with a host, hosted in India, by the month, on an INR invoice with GST, and can deploy and run the host for you. We do not sell per-second GPUs or serverless endpoints.
One GPU card for one month with a baseline host; the written quote lists CPU, RAM, storage, network, card count and location. Prices are “from”, confirmed with availability in the quote. Deployment (from ₹50,000) and GPU Infrastructure Care (from ₹25,000 a node, ₹50,000 minimum) are separate.
RunPod announced AP-IN-1 in India on 20 April 2026, on Secure Cloud, focused on H100 80GB cards; the city is not stated on its public pages. Check its console for today’s GPUs there. Our GPUs are hosted in India, with the location named in your quote.
It depends on configuration and billable hours. Example: an H100 PCIe on Secure Cloud at $2.89 an hour costs ₹2,02,067 for 730 hours at ₹95.78 a dollar; ours is from ₹1,80,000. Below about 650 hours a month RunPod costs less, before tax and storage.
We do not sell hourly GPUs. If you need burst or hourly capacity with an engineer behind it, we run AWS GPU instances in your own AWS account: 24/7 monitoring, approved changes, incident response and cost reviews, from ₹50,000 a month. Request a Managed AWS quote.
Yes. We quote within one business day, hand over a paid trial node so you can test your container and dependencies, then deploy and accept the runtime, framework and observability from ₹50,000. Storage and transfer steps are agreed first. Keep RunPod live until acceptance passes.
Dedicated NVIDIA GPUs hosted in India, L40S from ₹55,000 a card a month. INR invoice with GST. Managed care optional.