gcube

Easy to Share and Economical to Rent

Global GPU NETWORK

Use GPU economically by charging based on resource usage,
not fixed costs.

Reasonable Cost - Pay Only for What You Use

Up to 70% More Economical
GPU Cloud Platform

Use GPU economically by charging based on resource usage,
not fixed costs.

CES Innovation Awards®2026 Honoree in Artificial Intelligence

The CES Innovation Awards are based upon descriptive materials submitted to the judges.CTA did not verify the accuracy of any submission or of any claims made and did not test the item to which the award was given.

CES Innovation Awards®2026 Honoree inArtificial Intelligence

Efficient GPU Usage at Minimal Cost

Good performance but high cost, Low cost but insufficient performance.
We have overcome the limitations of GPU sharing services.

Global GPU Grid
gcube is a GPU sharing economy service that maintains strong cloud computing performance through Global GPU networking while providing it at an economical price.
Applicable to any environment
Customized Supply
Using cloud-native technology, we combine CSP GPUs and PC GPUs to provide various customized services for high-traffic services, irregular demand, and more.
Multiple domestic GPU resources
secured for High Stability
Starting with domestic-based services, we secure multiple suppliers to provide stable services through various GPU resources and fast networks.
Affordable for Everyone
GPU Available to Use
Direct contracts with suppliers and technology that anyone can easily supply reduce supply costs, making it more economical to use.

Up to 70% more economical GPU to use.
Use only what you need without worrying about costs, enjoy it.

Hourly GPU Usage Cost

39,746KRW
4,830KRW
Cost Advantage
H100 X 1
137,646KRW
34,720KRW
Cost Advantage
H100 X 8
57,344KRW
20,790KRW
Cost Advantage
A100
34,272KRW
1,610KRW
Cost Advantage
V100 X 2

gcube: Our Company
Other Companies: Other Companies

Based on 1 month: ??0,000 KRW from available.

gcube does not charge for unused time when using GPU. Costs are charged based on actual resource usage ratio through real-time monitoring.

Economical and Diverse GPUs

Choose from a variety of GPUs
with the purpose and specs you want at gcube.

Choose from a variety of GPUs
with the purpose and specs
you want at gcube.

TIER 1
For government-supported projects Large-scale data processing

Large capacity
High performance
24/7 stable

GPUs provided by Cloud service providers

TIER 2
For universities and research institutes Research purposes

Stable network
High performance
Customized

Dedicated server GPUs provided through direct contracts

TIER 2
For startups AI model development

Stable network
High performance
Customized

Dedicated server GPUs provided through direct contracts

TIER 3
For individual developers Resources for mining

Affordable
Diverse
Flexible

Personal / PC bang GPUs at economical prices

GPU Model: 0 types +

Hourly Price: 0 KRW~

Monthly Usage Cost: 0 KRW~

RTX 5000 Series
RTX 4000 Series
RTX 3000 Series
RTX 2000 Series
RTX A Series
RTX ADA Series
etc

Based on real-time GPU usage analysis data,
we establish a reasonable cost structure
and transparently disclose analysis data.

2025/06/08 15:30:27 routes.go : 1259 : INFO server config env=”map[CUDA_VISIBLE_DEVICES:
GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION:HTTPS_PROXYL
HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false
OLLAMA_GPU_OVERHEAD:0 OlLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false
OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY:
OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512
OLLAMA_MODELS:/root/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false
OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost
http://localhost http://localhost:* https://localhost:* http:127.0.0.1 https://127.0.0.1
http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:*
http://0.0.0.0* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false
ROCR_VISIBLE_DEVICES: http_proxy:https_proxy: no_proxy:]”

time=2025-06-18T15:30:27.553Z level=INFO source=images.go:757 msg=”total blobs: 5”
time=2025-06-18T15:30:27.553Z level=INFO source=images.go:757 msg=”total unused blobs removed: 0”
[GIN-debug] [WARNING] Creating an Engine instance with the Logger and Recovery middleware alreay attached.

[GIN-debug] [WARNING] Running in “debug” mode. Switch to “release” mode in production.
- using env: export GIN_MODE=release

2025/06/08 15:30:27 routes.go : 1259 : INFO server config env=”map[CUDA_VISIBLE_DEVICES:
GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION:HTTPS_PROXYL
HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false
OLLAMA_GPU_OVERHEAD:0 OlLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false
OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY:
OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512
OLLAMA_MODELS:/root/.ollama/models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false
OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost
http://localhost http://localhost:* https://localhost:* http:127.0.0.1 https://127.0.0.1
http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:*
http://0.0.0.0* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false
ROCR_VISIBLE_DEVICES: http_proxy:https_proxy: no_proxy:]”

time=2025-06-18T15:30:27.553Z level=INFO source=images.go:757 msg=”total blobs: 5”
time=2025-06-18T15:30:27.553Z level=INFO source=images.go:757 msg=”total unused blobs removed: 0”
[GIN-debug] [WARNING] Creating an Engine instance with the Logger and Recovery middleware alreay attached.

[GIN-debug] [WARNING] Running in “debug” mode. Switch to “release” mode in production.
- using env: export GIN_MODE=release
Container Log Real-time workload operation log inquiry is available after workload deployment.

WARNING
Startup probe failed: Get "http://10.249.7.253:15021/healthz/ready": dial tcp 10.249.7.253:15021: connection refused
Deployment Event Status Status change history is shared in real-time during workload operation.Also provided through console notifications, allowing you to immediately check disconnections or status changes and take necessary actions.

Monitoring We disclose hourly resource usage of connected GPUs during workload execution. We provide a reliable service by sharing GPU cost calculation data with users.