South Korea’s Ministry of Science and ICT says Nvidia has reconfirmed that it plans to proceed, at the company level and as previously announced, with the supply of 260,000 advanced GPUs to Korea and its cooperation with Korean companies, following the resignation of Nvidia’s chief executive. Deputy Prime Minister and Science Minister Bae Kyung-hoon wrote in an official post on X on 13 November that the ministry’s 2nd Vice Minister chaired an issues-review meeting with Nvidia that morning.
The post is in Korean; quotations here are our translation. Bae wrote that many people had asked whether GPU supply and cooperation projects were proceeding as scheduled after the resignation. Korea is the first government we have seen publicly confirm that its Nvidia deals stand since the chief executive change.
Business Pill · Sovereign AI
A short Business Engineer explainer on sovereign AI: why governments buy their own compute. It teaches the general idea only and does not discuss this story.
The key insight: South Korea’s science ministry says Nvidia reconfirmed, at the company level, the supply of 260,000 advanced GPUs and its cooperation with Korean companies after the resignation of Nvidia’s chief executive. The ministry says Seoul has secured about 21,000 of the 35,000 GPUs it buys directly, 60% by our arithmetic.
According to the post, the two sides agreed to keep pursuing “linkage between Korean AI chips and the Nvidia platform using NVLink”, building AI computing infrastructure, joint research on physical AI and talent development, and holding GTC, Nvidia’s developer event, in Korea.
Korea’s government plan of 50,000 GPUs (B200 compute-equivalent): about 21,000 of the 35,000 bought directly are secured, about 14,000 are due by early 2027, and 15,000 come through the National AI Computing Center. Source: the science minister’s official X post, 13 November 2026 (our translation). The 60% share is FourWeekMBA’s arithmetic.
Seoul’s Own 50,000 GPUs
Bae wrote that the government’s plan to secure 50,000 GPUs is proceeding as scheduled. The post says that of approximately 35,000 the government will secure directly, approximately 21,000 have been secured so far, and the remaining approximately 14,000 are scheduled to arrive by early 2027. Both figures are on a B200 compute-equivalent basis. A separate 15,000 are being secured through the National AI Computing Center.
By our arithmetic, 21,000 of 35,000 is 60% of the directly secured tranche, leaving 40% to arrive. The 35,000 and the 15,000 add up to the 50,000 plan. On 9 November Vice Minister Ryu Jemyung said 10,000 government-funded GPUs were already in use, as reported by Reuters and covered in Korea Says Nvidia GPU Deliveries on Track, 10,000 in Use; the post does not say how that count relates to the 21,000.
Where the 260,000 Sit
Nvidia’s 31 October 2025 APEC announcement, as reported in our 9 November piece, put the Korean build-out at over 260,000 GPUs: the government over 50,000, Samsung Electronics over 50,000, SK Group over 50,000, Hyundai Motor Group 50,000 and NAVER Cloud over 60,000. The government plan is about 19% of the total by our arithmetic, so the reconfirmation covers a programme 5.2 times the size of Seoul’s own share.
As we read the post, the ministry is speaking as a customer: it reports a supply check with Nvidia on the morning after the resignation news, not a view on the leadership change itself.
As we read the figures, the government’s own plan is a small slice of the total: 50,000 of the 260,000, about 19% by our arithmetic, with the larger orders sitting with four company programmes.
As we read the agenda, the items kept on the table (NVLink linkage with Korean AI chips, physical AI research and a GTC in Korea) are partnership terms rather than delivery volumes.
Bae Kyung-hoon, South Korea’s Deputy Prime Minister and Minister of Science and ICT, official X post, 13 November 2026 (our translation)
“The government’s plan to secure 50,000 GPUs is also proceeding as scheduled”
Three Implications
FOR KOREA’S SOVEREIGN AI PROGRAMME The ministry says about 14,000 of the government’s directly bought GPUs are due by early 2027, with 15,000 more through the National AI Computing Center.
FOR NVIDIA’S KOREAN PARTNERS The ministry says the NVLink linkage with Korean AI chips, physical AI research and a GTC in Korea stay on the agenda.
FOR THE FOUR COMPANY PROGRAMMES Samsung, SK Group, Hyundai Motor Group and NAVER Cloud hold about 81% of the announced GPUs by our arithmetic; the post reports no statements from them.
The analysis says: “The total cycle: $725B Big Four + $50B Oracle + $13B Apple + $60B Neoclouds + $80B China + $60B Sovereigns + $48B Other = ~$1.04 trillion in 2026 compute capex, the first trillion-dollar year of compute capex in human history.”
As we read it, Korea’s government programme is one of the sovereign buyers in that map, and the ministry’s post addresses supply continuity for that share. The analysis predates the post and does not discuss Korea’s contracts.
Nvidia’s 31 October 2025 Korea allocation by recipient, minimum figures. The 19% government share and 5.2 times ratio are FourWeekMBA’s arithmetic.
What Is Not Established
The account of the meeting is the ministry’s, and the post does not quote Nvidia directly. It does not say which Nvidia executives attended, does not give delivery dates for the company programmes, and does not report statements from Samsung, SK Group, Hyundai Motor Group or NAVER Cloud. The 60%, 19% and 5.2 times figures are our arithmetic.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Korea’s science ministry says Nvidia reconfirmed the 260,000-GPU supply and its Korean partnerships after the chief executive change, and that Seoul has secured about 21,000 of the 35,000 GPUs it is buying directly, 60% by our arithmetic, with about 14,000 due by early 2027.
A note on sourcing. The ministry’s account is from the official X post of Deputy Prime Minister and Science Minister Bae Kyung-hoon, 13 November 2026, written in Korean; quotations are our translation. The 31 October 2025 allocation is from Nvidia’s newsroom release as reported in our 9 November piece. We did not contact the ministry or Nvidia. The 60%, 19% and 5.2 times figures are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
