Nodes

Supply health, regional distribution, earnings, capacity, and quality.

Total nodes+9.1%
4,603
Unique node operator sessions
Active nodes+4.2%
3,164
Online + processing + idle
Total GPUs+11.4%
11.8K
Physical GPU units across all nodes
Active GPUs+5.1%
8,146
GPU units in active nodes
Processing now+6.4%
1,703
666 idle · 997 offline
Degraded-0.6%
181
RS < 60 or high temperature events
Median RS-1.8%
71
Reliability Score 0–100
Avg uptime+0.6%
83%
Completion rate+0.4%
97.8%
Tokens 30d+22.1%
16.3B
Avg 66 tps / GPU
Earnings 30d+14.4%
$4.5K
Pending payouts+6.4%
$1.1K
Owed to operators
  • Pause nodes with RS < 50 for a 24h cooldownNode Team

    The bottom 1.5% of the network (RS < 50) account for ~38% of failed jobs. A cooldown + automated retry historically improves median RS by 3 pts.

    Open
  • Expand Tier-A capacity for Llama 3.1 70BAI / Infra

    Demand for this model outpaced Tier-A supply at peak UTC hours. Queue depth reached 11%. Routing to idle Tier-A nodes in SE/NL is a short-term fix.

    Open

Active nodes over time (30d)

Node status breakdown

4,603
Total
  • online795
  • processing1,703
  • idle666
  • offline997
  • paused226
  • degraded181
  • banned35

Regional supply analysis

Active nodes vs active developers per region — higher dev/node ratio = capacity gap

EU
Active nodes1,086
Active devs1,093
Devs / node1.0
Asia
Active nodes796
Active devs894
Devs / node1.1
US
Active nodes454
Active devs480
Devs / node1.1
MENA
Active nodes386
Active devs384
Devs / node1.0
LATAM
Active nodes245
Active devs296
Devs / node1.2
Oceania
Active nodes100
Active devs119
Devs / node1.2
Africa
Active nodes97
Active devs102
Devs / node1.1

Global avg: 1.1 active devs per active node. Regions above 1.5 are flagged as under-served.

Active nodes by country (top 15)

15 countries
CountryRegionActive nodesTotal nodesActive devsDevs / nodeShare
United StatesUS3474983521.0
IndiaAsia1832782291.3
United KingdomEU1732471691.0
GermanyEU1632331711.0
FranceEU1151571151.0
CanadaUS1071631281.2
NetherlandsEU1041531071.0
United Arab EmiratesMENA102145810.8
PolandEU93146891.0
SingaporeAsia92134870.9
BrazilLATAM911371031.1
JapanAsia84121901.1
IsraelMENA82121851.0
South KoreaAsia79115801.0
AustraliaOceania77111971.3

Node earnings over time (30d)

Reliability Score median (30d)

All nodes
800 nodes · sortable, searchable, exportable
NodeStatusGPUVRAMOSRSUptimeJobs 30dTokens 30dLatencyEarnings 30dLast seen
Beatriz Schneider
Hanoi, VN
processing16×RTX A6000768 GBacross 16 GPUsWindows 118989.3%92.4K220.7M1.28s$57.12—
Julia Eriksen
Seattle, US
online8×RTX A6000384 GBacross 8 GPUsWindows 117290.7%53.9K120.4M1.02s$38.19—
Theo Wang
Melbourne, AU
online16×RTX 4080256 GBacross 16 GPUsWindows 117097.4%37.8K89.3M778ms$36.49—
Daniel Lim
Busan, KR
processing10×RTX 3090+1 other336 GBacross 16 GPUsWindows 118188.7%39.5K92.7M631ms$35.74—
Diego Sato
Taipei, TW
online8×RTX A6000384 GBacross 8 GPUsWindows 119693.6%55.4K128.7M1.02s$27.28—
Omar Brown
Manama, BH
online16×RTX A6000768 GBacross 16 GPUsWindows 117799.1%94.1K64.9M414ms$27.09—
Leila Khalil
Los Angeles, US
idle16×RTX 4090384 GBacross 16 GPUsWindows 117499.3%26.3K61.6M380ms$25.36—
Vikram Khalil
Santiago, CL
online10×RTX 4070+1 other408 GBacross 16 GPUsWindows 119398.7%32.3K58.9M743ms$24.73—
Julia Ortega
Dubai, AE
online16×RTX 3090384 GBacross 16 GPUsWindows 117593.9%39.2K68.4M276ms$24.43—
Ravi Petrosyan
Singapore, SG
processing8×RTX 3090192 GBacross 8 GPUsWindows 119395.8%24K52.5M690ms$22.03—
Simon Halim
Brno, CZ
processing6×RTX 3090+1 other192 GBacross 8 GPUsWindows 118195.2%24K54.4M757ms$21.49—
Marko Holm
Stockholm, SE
processing13×RTX A5000+1 other348 GBacross 16 GPUsWindows 119894.6%35.3K51.7M1.12s$21.16—
1–12 of 800