Naver turns a ₩400B state loan into a Korea compute push
Original: Naver Secures ₩400 Billion for South Korea AI Data Center Expansion View original →
Naver’s latest AI move is not just a company capex story. It is industrial policy with racks and power bills attached. An ETDatacenters report published on April 24, 2026 says South Korea’s Financial Services Commission has backed Naver with a ₩400 billion (US$270 million) loan to expand its Gak Sejong AI data center. The reported rate is about 3 percent, which makes this more than financing. It is a clear signal that domestic compute capacity is now a national asset.
The money is aimed at very specific bottlenecks. ETDatacenters says the expansion will support GPU server deployment, strengthen Naver’s HyperCLOVA X large language model, and deepen AI integration inside Naver’s search services. That matters because South Korea, like many countries outside the U.S. hyperscaler core, is trying to avoid a future where every serious model improvement depends on foreign cloud availability and foreign policy risk. A sovereign model strategy without domestic infrastructure eventually runs into the wall of someone else’s queue.
Gak Sejong therefore becomes more than another data-center campus. It is the physical layer of a national AI stack: compute, model training, inference capacity and service integration living inside a policy framework that explicitly talks about reducing reliance on global big tech. The report says the FSC wants the financing to help the domestic AI industry secure its own sovereignty through large-scale data-center construction. That is unusually direct language. Governments used to subsidize chip fabs and broadband. In 2026, they are increasingly underwriting model-era electricity and GPUs.
The market backdrop explains the urgency. ETDatacenters cites forecasts showing South Korea’s AI market could rise from US$5.47 billion in 2024 to US$53.87 billion by 2032. Those projections can move around, but the directional point is hard to miss: if local demand is compounding this quickly, local infrastructure cannot remain an afterthought. Naver still has to execute, and the expansion will only matter if HyperCLOVA X and related services keep improving. But the message from Seoul is already clear. Korea does not want AI sovereignty to be a slogan imported from Europe. It wants it wired into a domestic compute base while the race is still open.
Related Articles
Safe Superintelligence will expand its compute by an order of magnitude through NVIDIA investment and access to Vera Rubin systems. The deal turns a productless frontier lab into a strategic customer for upcoming AI infrastructure.
NVIDIA detailed Vera, a CPU designed for agentic AI workloads where tool calls, code execution, retrieval, and verification sit between model calls. The company claims 50% higher IPC than Grace and 1.8x sustained per-core performance versus x86 on agentic execution workloads.
AP reports that AI data-center spending likely topping $700B this year is feeding into memory-chip costs, electricity prices, consumer electronics, and Fed inflation debates. The AI buildout is now a macro variable, not just a capex line.