South Korea's SK Group and Nvidia are widening an existing partnership that spans so-called AI factories and memory, according to a report from the electronics trade publication Evertiq surfaced via Google News.
The available source item is a single summary line, so the specifics — the size of the commitment, the timeline, the facilities involved, and any executive comments — are not detailed in it. What it does establish is the shape of the deal: two areas, AI factories and memory, and an expansion rather than a first-time tie-up.
That framing is worth unpacking for non-specialists. "AI factory" is the industry's term for a data center built specifically to train and run AI models, packed with accelerators rather than the general-purpose servers that handled traditional computing. Memory is the other half of the equation: modern AI chips are frequently limited less by raw processing power than by how fast data can be fed to them, which makes high-performance memory a bottleneck technology rather than a commodity add-on.
SK Group is one of South Korea's largest conglomerates, and Nvidia is the dominant supplier of the processors used to train large AI models. An expanded arrangement between a major Korean industrial group and Nvidia touches both ends of the AI supply chain — the components that go into systems, and the buildings full of those systems.
Why it matters: the AI buildout increasingly hinges on securing memory and data center capacity, so deals like this signal where the physical constraints on AI progress are being addressed — and by whom.