SK and Nvidia outline $500B AI infrastructure plan
SK Group and Nvidia have outlined a large multi-year AI infrastructure and memory partnership spanning cloud capacity and next-generation chips.
The largest AI infrastructure projects increasingly depend on alliances between chip designers, memory manufacturers, telecom operators and long-term capital providers.
What happened
SK Group and Nvidia signed letters of intent covering a headline $500B-plus AI infrastructure and memory initiative. The plans include a Korean AI cloud of up to 2GW operated through SK Telecom and a long-term partnership between Nvidia and SK hynix covering high-bandwidth memory and other next-generation memory technologies.
The first AI factory is planned for 2027 and would use Nvidia’s Vera Rubin platform alongside SK hynix memory.
The $500B-plus figure is not a completed financing or committed investment round. It describes the projected scale of a multi-year initiative governed initially by letters of intent, so actual spending and deployment could differ significantly.
Why it matters
SK can provide memory, telecom infrastructure and data-centre operations, while Nvidia supplies the broader computing architecture. That combination could create a more integrated Korean AI platform and secure long-term demand for both companies.
The bigger picture
AI infrastructure is moving toward vertically coordinated ecosystems rather than isolated hardware purchases. The strategic signal is strong, but execution will depend on power availability, construction timelines, customer demand and whether preliminary agreements become binding projects.