South Korea's Free AI Chatbot Plan Signals Modest Chip Demand

South Korea has selected a consortium led by SK Telecom, Kakao, and KT to deliver free, unlimited-access AI chatbots to approximately 51 million citizens, with beta testing scheduled for September and full launch by year-end. The government is seeding the program with 512 NVIDIA B200 GPUs this year and plans ongoing operating subsidies starting in 2027.

The initiative represents a tangible sovereign AI deployment, but the compute allocation remains a fraction of what frontier AI labs consume. NVIDIA disclosed in its most recent 10-Q an expanded Korea AI factory ecosystem through partnerships with SK Telecom, NAVER, and Brookfield, plus a multiyear technology partnership with SK hynix for next-generation memory, according to 247 Wall St.

Why it matters

The Korean program illustrates the emerging neocloud model NVIDIA CEO Jensen Huang described on recent earnings calls: countries allocating land and power directly to regional cloud partners in ways they wouldn't for foreign hyperscalers. NVIDIA now expects non-hyperscaler business—including sovereigns and neoclouds—to represent roughly half of its data center revenue. Yet the 512-GPU allocation underscores the gap between policy ambition and the physics of inference at population scale. Rate limits, queues, and domestic-model quotas will constrain the "no limits" promise long before hardware does.

NVIDIA's sovereign AI momentum

NVIDIA reported Q2 FY2027 revenue of $96.22 billion, up 105.8% year over year, with Data Center revenue reaching $89.02 billion, up 117%. Management said sovereign AI business grew 35% sequentially and more than tripled year over year in Q2. The company specifically noted that South Korea's LG and Hyundai Motor Group are partnering with NVIDIA to build and scale AI.

Shares have risen 23.67% year to date to $230.36, with a price-to-earnings ratio of 46. However, the Korean disclosure remains qualitative—no GPU volumes, dollar values, or delivery schedules were provided for the LG or Hyundai partnerships. NVIDIA carries $279 billion in supply commitments tied largely to memory procurement and assumes no China Data Center compute revenue in Q3 guidance.

SK hynix supplies the memory layer

SK hynix provides the high-bandwidth memory stacks for every B200 shipped into Korea. The company reported preliminary Q2 FY2026 revenue of KRW 79.32 trillion, up 50.9% quarter over quarter and 256.8% year over year, with operating profit up 557.2% and net profit attributable to controlling interests up 1,240.8% year over year. The ADS has gained 17.2% over the past month to $177.

SK hynix held a groundbreaking ceremony for an HBM production base in Indiana on August 27, 2026, and is accelerating a 40 trillion won share repurchase program targeting over 50% of free cash flow returned to shareholders. NVIDIA management noted that memory scarcity driven by AI buildout has pushed price increases beyond prior expectations.

TSMC fabricates the silicon

Taiwan Semiconductor Manufacturing fabricates the B200 chips Korea is deploying and the custom accelerators Korean developers will design. TSMC reported Q2 2026 revenue of $40.20 billion, up 36% year over year, with high-performance computing at 66% of quarterly revenue and advanced nodes at 7nm and below representing 77% of wafer revenue. The stock has climbed 41.85% year to date to $428.91.

TSMC raised its 2026 capital budget to $60 billion to $64 billion and said CAPEX over the next three years will be "even more significantly higher" than the past three years. An additional $100 billion investment in Arizona brings total planned Arizona spending to $265 billion. Management identified advanced packaging as the current constraint, noting capacity is "so tight that now it's a little bit of my customers' growth."

The details were first reported by 247 Wall St.