Open From Data Center to Desktop: NVIDIA Expands Nemotron Family With Highest Efficiency Super VL and Bolt Models
First-party signal from NVIDIA. Open the canonical release for the complete announcement.
MARKET PULSEH100 $4.18/GPU-HR 3PH200 $4.47/GPU-HR 1PMI300X $2.59/GPU-HR 1PL40S $1.62/GPU-HR 2PB300 $16.0/GPU-HR 1POBSERVED 2026-10-20 · 1H CACHE
INDEPENDENT GPU CLOUD COMPARISON
Compare GPU prices.
Compare GPU rental prices, H100 and A100 configurations, and the cost of your next AI workload. Every listing includes its source and observation date.
79public listings
across 4 providers
across 4 providers
Quick view
19 offers across 3 providers
| Provider | Configuration | Listed location | Per GPU / hr | Compute estimate | Shortlist |
|---|---|---|---|---|---|
| DigitalOcean On-demand | 1 × RTX 4000 Ada20 GB VRAM / GPU · 32 GB RAM | TorontoAvailability unconfirmed | $0.76USD | $5471 plan · 1 GPU · 720h | 20 Oct |
| DigitalOcean On-demand | 1 × RTX 6000 Ada48 GB VRAM / GPU · 64 GB RAM | TorontoAvailability unconfirmed | $1.57USD | $1,1301 plan · 1 GPU · 720h | 20 Oct |
| DigitalOcean On-demand | 1 × L40S48 GB VRAM / GPU · 64 GB RAM | TorontoAvailability unconfirmed | $1.57USD | $1,1301 plan · 1 GPU · 720h | 20 Oct |
| Atlantic.Net Public list | 1 × L40SL40S · 192 GB RAM | Orlando · Dallas · New YorkAvailability unconfirmed | $1.67USD | $1,2011 plan · 1 GPU · 720h | 20 Oct |
| DigitalOcean On-demand | 1 × MI300X192 GB VRAM / GPU · 240 GB RAM | AtlantaAvailability unconfirmed | $2.59USD | $1,8651 plan · 1 GPU · 720h | 20 Oct |
| Latitude Public list | 1 × H10080 GB VRAM / GPU · 192 GB RAM | United StatesAvailability unconfirmed | $3.37USD | $2,4261 plan · 1 GPU · 720h | 20 Oct |
Estimates cover whole-plan compute charges at the displayed rate. Storage, transfer, taxes, discounts and changing Spot rates can affect the final bill. Different GPU variants and network configurations are not performance-equivalent.
Public list observations · 20 OctGPU Cloud Compare (CC-BY 4.0) + Microsoft Azure
01 / MARKET BOARD
The price of intelligence.
Reference rates and advertised price history, with source coverage and observation dates. Compare individual rental offers above.
CURRENT
H100 minimum advertised plan rate
$4.41LOWEST / PLAN-HR
$4.41$4.41$4.41
2026-09-212026-10-062026-10-20
PUBLIC LIST PRICES · NOT A TRANSACTION INDEXCC-BY SOURCE ↗
CURRENTCHECKED 20 OCT, 11:07 UTCSOURCE UPDATES DAILY · WE CHECK DAILY SCHEDULED REFRESH PLUS HOURLY CACHE, ON REQUEST
SILICONTICKER REFERENCE INDEX
CL-FCR Frontier Compute Rate
One current-snapshot reference rate across eligible H100-family, H200, MI300X and B300 public cloud offers.
REFERENCE INDEX · NOT TRADEABLE$4.41USD / GPU-HR
- RANGE
- $2.59–$16.00
- COVERAGE
- 3 providers · 9 offers
- CONSTITUENTS
- H100 family · H200 · MI300X · B300
- OBSERVED
- 2026-10-20
GPU FAMILYMEDIANRANGECOVERAGE
H100H100 family · public cloud$4.18/ GPU hr$3.37–$4.413 providersBROAD · 4 offers · 2026-10-20
H200H200 · public cloud$4.47/ GPU hr$4.47–$4.471 providerSINGLE SOURCE · 2 offers · 2026-10-20
MI300XAMD MI300X · public cloud$2.59/ GPU hr$2.59–$2.591 providerSINGLE SOURCE · 2 offers · 2026-10-20
L40SL40S · public cloud$1.62/ GPU hr$1.57–$1.672 providersLIMITED · 2 offers · 2026-10-20
B300HGX B300 · public cloud$16.0/ GPU hr$16.00–$16.001 providerSINGLE SOURCE · 1 offers · 2026-10-20
RTX6KRTX PRO 6000 · public cloud$6.25/ GPU hr$6.00–$8.251 providerSINGLE SOURCE · 4 offers · 2026-10-20
CAPACITY CALCULATOR
What does your GPU month cost?
Data: gpucloudcompare.com (CC-BY-4.0)METHODOLOGY + LICENSE ↗
02 / INTELLIGENCE
Signals, not noise.
NVIDIA Expands Advanced Technology Access for Europe’s Scientific Community Through Open Learning Resources, Collaborations and Compute
First-party signal from NVIDIA. Open the canonical release for the complete announcement.
Introducing AMD Helion, Agentic AI Compute at Rack Scale
First-party signal from AMD. Open the canonical release for the complete announcement.
GPU RENTAL FIELD GUIDES
Make the price useful.
Understand the bill, compare hardware by useful output and price the risk of interruption.
GPU Rental Pricing: Hourly Rates and Total Cost
Understand GPU rental prices, per-GPU versus whole-plan billing, regional differences and the full cost of an AI workload. Compare dated public listings.
H100 vs A100: Compare GPUs by Workload Cost
Compare NVIDIA H100 and A100 memory, hardware variants and rental economics. Use measured workload throughput to choose the lower cost per result.
Spot vs On-Demand GPUs: Price, Capacity and Risk
Understand interruptible Spot GPU pricing, on-demand capacity and reservations. Compare total job cost, checkpointing and availability before renting.
03 / BUILT FOR MACHINES, TOO
Clean enough to read.
Structured enough to reason with.
Every signal has a date, unit, provenance and canonical URL. Humans get context; agents get deterministic endpoints.
# Editorial map and usage notes
GET /llms.txt
# Hourly-checked market + news data
GET /api/markets
GET /api/markets/azure-spot
GET /api/news
# Source rights, cadence and health
GET /api/sources200 PROVENANCE · FRESHNESS · VERSIONED SCHEMA
THE COMPUTE BRIEF