GPU Rent Index: cloud GPU prices per hour in ₹ and $

Hourly on-demand prices for one GPU on Indian clouds (E2E Networks, Jarvislabs India region) and global ones (RunPod, Lambda, Vast.ai), side by side in rupees and dollars. Cheapest first in each table.

FX rate: 1 USD = ₹94.93 (ECB reference rate for 2026-10-19, via Frankfurter). Prices were copied from each provider's public pricing page, except Vast.ai rows which come from its public offer API. All prices exclude GST/taxes, storage and egress, and USD-billed providers may add card FX charges. Spot, community and reserved prices are lower; confirm the live quote before renting. Data: gpu_prices.json. No links on this page are affiliate links.

RTX 4090

Provider₹ / GPU-hour$ / GPU-hourSource
RunPod (Community Cloud)
1x pod, community hosts
₹32.28
converted
$0.34Pricing page
checked 2026-10-20 · manual
Vast.ai (marketplace)
cheapest of 1 verified 1x on-demand offers; hosts vary
₹43.57
converted
$0.46Pricing page
checked 2026-10-20 · public API
RunPod (Secure Cloud)
1x pod
₹84.49
converted
$0.89Pricing page
checked 2026-10-20 · manual

L4

Provider₹ / GPU-hour$ / GPU-hourSource
Vast.ai (marketplace)
cheapest of 1 verified 1x on-demand offers; hosts vary
₹34.36
converted
$0.36Pricing page
checked 2026-10-20 · public API
Jarvislabs (India region)
1x, on-demand, per-minute billing
₹41.31$0.43
converted
Pricing page
checked 2026-10-20 · manual
E2E Networks
1x, on-demand
₹49.00$0.52
converted
Pricing page
checked 2026-10-20 · manual
RunPod (Secure Cloud)
1x pod
₹56.01
converted
$0.59Pricing page
checked 2026-10-20 · manual

L40S

Provider₹ / GPU-hour$ / GPU-hourSource
Vast.ai (marketplace)
cheapest of 1 verified 1x on-demand offers; hosts vary
₹63.51
converted
$0.67Pricing page
checked 2026-10-20 · public API
E2E Networks
1x, on-demand
₹102.00$1.07
converted
Pricing page
checked 2026-10-20 · manual
RunPod (Secure Cloud)
1x pod
₹103.47
converted
$1.09Pricing page
checked 2026-10-20 · manual

A100 80GB

Provider₹ / GPU-hour$ / GPU-hourSource
Vast.ai (marketplace)
cheapest of 5 verified 1x on-demand offers; hosts vary
₹52.12
converted
$0.55Pricing page
checked 2026-10-20 · public API
Jarvislabs (India region)
1x, on-demand
₹140.94$1.49
converted
Pricing page
checked 2026-10-20 · manual
RunPod (Secure Cloud)
1x SXM pod
₹169.92
converted
$1.79Pricing page
checked 2026-10-20 · manual
E2E Networks
1x, on-demand
₹189.00$1.99
converted
Pricing page
checked 2026-10-20 · manual
Lambda
8x SXM instance only (per GPU)
₹264.85
converted
$2.79Pricing page
checked 2026-10-20 · manual

H100 80GB

Provider₹ / GPU-hour$ / GPU-hourSource
RunPod (Community Cloud)
1x SXM pod, community hosts
₹255.36
converted
$2.69Pricing page
checked 2026-10-20 · manual
Vast.ai (marketplace)
cheapest of 1 verified 1x on-demand offers; hosts vary
₹295.04
converted
$3.11Pricing page
checked 2026-10-20 · public API
Jarvislabs (India region)
1x SXM, on-demand
₹330.48$3.48
converted
Pricing page
checked 2026-10-20 · manual
RunPod (Secure Cloud)
1x SXM pod
₹378.77
converted
$3.99Pricing page
checked 2026-10-20 · manual
E2E Networks
1x, on-demand
₹379.05$3.99
converted
Pricing page
checked 2026-10-20 · manual
Lambda
1x SXM instance
₹407.25
converted
$4.29Pricing page
checked 2026-10-20 · manual

Edge vs cloud: when does a board beat renting?

A cloud GPU wins for big models and short bursts: an L4 at about ₹41–49/hour runs 7B–13B models far faster than any board. A board wins when you need a small model running all day, offline or privately: a Raspberry Pi 5 or Jetson costs a one-time price, while 24×7 on even the cheapest GPU here is roughly ₹30,000/month. See what boards actually manage on the EdgeLLM leaderboard and board pages.

Last updated 2026-10-20. Spotted a wrong or stale price? Tell us.

📬 The EdgeLLM Weekly

Every week, free: new real-world benchmarks · India price drops on boards · one build idea you can try this weekend.

No spam. Unsubscribe anytime.