SK hynix has reportedly already started supplying "small quantities" of its next-generation HBM4 memory to NVIDIA, ready to power the company's next-gen Rubin AI GPUs.
In a new report by Korean media outlet Dealsite and picked up by insider @Jukanrosleve on X, we're hearing that SK hynix has been pre-supplying NVIDIA with its new HBM4 memory in small quantities. US-based memory giant Micron has already started shipping its own HBM4 memory to NVIDIA, but the industry believes that SK hynix will have a "significant portion" of the initial HBM4 volume.
The industry believes that HBM3E memory forced a price premium of around 20% but next-gen HBM4 is expected to drive that premium upwards of over 30% as the design of HBM4 is much more complex which has driven up manufacturing costs because of the number of I/O pins doubling from 1024 to 2048, and the die size increasing.
Micron is closely behind SK hynix with HBM4 supply to NVIDIA, but it also faces limitations in its volume allocation, with Dealsite's industry insider source explaining: "NVIDIA is actively building a 'multi-vendor' strategy. It is a natural step for Micron to enter the HBM4 supply chain. However, Micron's capacity is still limited, and its yield is known to be lower than SK hynix's, so there are limits to securing large-scale volume".
- Read more: SK hynix showcases world's first HBM4: 16-Hi stacks, 2TB/sec bandwidth
- Read more: SK hynix ships world's first 12-layer HBM4 samples for NVIDIA Rubin AI GPUs
- Read more: JEDEC releases HBM4 standard: ready for next-gen AI GPUs
- Read more: Micron begins shipping HBM4 memory to key partners for next-gen AI systems
- Read more: NVIDIA's next-gen Vera Rubin NVL576 AI server: 576 Rubin AI GPUs, 12672C/25344T CPU, HBM4
NVIDIA's expected market share has recently increased, with SK hynix's shipment volume analyzed to rise together with NVIDIA, with an industry official noting that "the global AI GPU volume is expected to double in a year. Along with this, NVIDIA's global market share will expand by around 20% from 45% to 65%".
Meritz Securities projected in a recent report that "within its main client, SK hynix is expected to first secure a share of over 60% for HBM3E 8-stack and over 75% for 12-stack in 2026. For HBM4, as the first supplier, it is estimated to have secured supply priority of over 75%".
SK hynix had originally planned to transition to mass production of HBM4 for October 2025, but with the HBM4 samples shipping to NVIDIA to its new Rubin AI GPUs and expected to be used in September, earlier than the industry expectede, it looks like SK hynix is pre-emptively supplying small quantities. NVIDIA will be launching its new Rubin AI GPUs in the second half of 2026, the entire progress is progressing quite quickly.
Another industry source said: "recently with SK hynix, there was talk about whether the October mass production transition target was set too aggressively. Because of this, there were also forecasts that proper mass production would begin in earnest from early next year".




