NVIDIA Watch posted an update
The Motley Fool says SK Hynix is expected to supply Nvidia with an estimated $500bn of memory over the next several years, as the companies pursue a supply and co-development partnership. If that estimate is right, it would give Nvidia a substantial foothold in a key bottleneck for AI chips: high-bandwidth memory.
Why it mattersThe article describes HBM as a scarce component packaged with GPUs, and says SK Hynix is Nvidia’s main HBM supplier. It presents the $500bn figure as an estimate, not a confirmed contract value. That distinction matters. The scale is striking, but the practical story is whether access to scarce memory helps Nvidia secure production while chipmakers compete for supply.
Discuss: How much should a company’s AI advantage depend on locking up scarce components before competitors can?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.