Nvidia and SK Group unveil $500B-plus AI buildout with SK Hynix memory pact
Nvidia and SK Group announced a $500 billion-plus AI initiative centered on large-scale data centers and securing next-generation high-bandwidth memory through SK Hynix. The partnership goes beyond procurement into joint development, positioning SK as an integrated AI infrastructure partner rather than only a memory supplier.
Key takeaway
Memory supply is being locked in as a first-order constraint on AI capacity, with GPU vendors and HBM makers co-investing in the full stack instead of treating chips and data centers as separate markets.
Context
The package ties Nvidia’s data-center expansion plans to SK Group capital and SK Hynix next-gen HBM, aiming to underwrite long-duration demand for advanced memory used in training and inference clusters. That structure matters because HBM has become a gating input for GPU deployment schedules, not a commodity afterthought.
For operators and hyperscalers, a commitment at this scale reinforces that memory availability will shape build timelines as much as accelerator allocation. It also intensifies vertical integration across the AI supply chain, where leading chip and memory firms co-design capacity rather than relying on spot procurement.
Numbers to know
- $500B+Stated size of the Nvidia–SK Group AI initiative covering data centers and next-gen memory supply