NAVER–NVIDIA AI Factory Power Expansion Roadmap
NAVER × NVIDIA · Infrastructure Partnership
AI Factory Power Expansion Roadmap
Phased deployment from 55 MW to gigawatt-scale global AI compute infrastructure
2027H1
1
Phase 1 · GAK Sejong Launch 55 MW
NVIDIA DSX First Deployment at GAK Sejong
Initial NVIDIA DSX platform integration at NAVER's hyperscale data center GAK Sejong. NAMU III hybrid cooling commissioned. DSX MaxLPS activates dynamic power allocation across initial Blackwell GPU clusters.
5.5% of 1 GW target
2027H2
2
Phase 2 · Global Expansion 100 MW
Sovereign AI Markets: Middle East & Europe Go-Live
Capacity doubles as overseas sovereign AI deployments come online. DSX Flex grid-responsive orchestration activates across distributed sites, synchronizing compute queues with local power grids in real time.
10% of 1 GW target
2028
3
Phase 3 · Asia-Pacific Cluster 200 MW
Ultra-Dense AI Factory Cluster: APAC + EMEA Grid Linkage
Interconnected AI factory clusters span Asia-Pacific, Europe, and the Middle East. Immersion cooling and direct liquid cooling fully deployed. DSX Exchange bridges IT/OT signals across all sites for unified fleet management.
20% of 1 GW target
2029+
GW
1 GW
Global AI Grid — Full Scale
Hundreds of thousands of next-gen NVIDIA GPUs running uninterrupted across a planetary compute grid. ≈ 4× the maximum design capacity of GAK Sejong (270 MW).
Source: NAVER–NVIDIA Strategic Partnership Announcement, June 2026 · NAVER Cloud Technical Whitepaper