Nvidia graphics processing units have become a key resource for AI infrastructure, pushing companies to change how they procure computing capacity. Firms are expanding beyond reliance on large cloud providers such as AWS, Microsoft Azure and Google Cloud, using neocloud specialists, online marketplaces, direct equipment purchases and customer-supplied hardware models. Demand continues to outstrip supply, and providers and customers are multiplying even as prices rise and prepayments and long-term contracts add to procurement costs.

South Korea will continue supporting two final teams developing a domestic AI foundation model through 2027, as originally planned, while launching a separate frontier AI development programme. The Science and ICT ministry plans to select two teams in February after evaluating the current round ending in late January. The frontier AI project would use a government stake to draw large private investment, with a tender possible in December after the 2027 budget is approved.