Projects

AI Factory sites and equipment roadmap

Regional AI Factory capacity and modular container-room deployment plans across Japan, Korea, and the United States.

AI Factory Sites

Regional AI Factory portfolio.
Built for B300 and Pro 6000 deployment.

GoodVision is developing AI Factory sites across Japan, Korea, and the United States, with staged power expansion plans and GPU cluster deployment paths designed for high-performance AI inference.

Fukushima, Japan city view
01
Fukushima, Japan
2MW initial
Target expansion to 12 MW. Planned deployment includes NVIDIA B300 server clusters and Pro 6000 inference card clusters.
Gimpo, Korea city view
02
Gimpo, Korea
0.75MW initial
Target expansion to 5.5 MW. Planned deployment focuses on NVIDIA Pro 6000 inference card clusters.
Texas industrial aerial view
03
Texas, United States
2.4MW initial
Target expansion to 20 MW. Planned deployment includes NVIDIA B300 server clusters and Pro 6000 inference card clusters.
AI Factory Equipment
Prefabricated modular container rooms.
High-density compute, faster to operate.
NVIDIA B300 Servers

512 GPUs in one container

GoodVision works with global partners to co-develop prefabricated modular container-room solutions for high-density NVIDIA B300 server deployments.

Container format1 modular compute container
Power capacity1.5 MW
GPU density512 GPUs
Target workloadHigh-density AI training and inference workloads
  • Designed around NVIDIA B300 server clusters.
  • High-density containerized compute increases token output per unit area.
  • Supports liquid cooling and advanced thermal design for efficient operation.
NVIDIA Pro 6000 Inference

960 GPUs in two containers

For NVIDIA Pro 6000 inference-card deployments, GoodVision uses a two-container modular design to create dense inference capacity with manageable power and thermal requirements.

Container format2 modular compute containers
Power capacity1 MW
GPU density960 GPUs
Target workloadHigh-throughput AI inference workloads
  • Designed around NVIDIA Pro 6000 inference card clusters.
  • Two-container configuration supports high GPU count and modular expansion.
  • Optimized for inference density and token output per unit area.
Why It Works
From design to operation.
Deployable in under six months.
A
Fast deployment with N+1 power

N+1 power container-room AI Factory deployments can be designed, developed, manufactured, installed, and brought into compute operation within six months.

B
Advanced cooling efficiency

Leading liquid-cooling and immersion-cooling solutions help keep facility-level PUE below 1.2, improving the operating efficiency of the AI Factory.

C
Higher token output per square meter

High-density modular containers increase available compute in a smaller footprint, improving token output per unit area across AI inference workloads.