SK hynix fuels AI race with HBM4E sample delivery
Seol, June 18
South Korean semiconductor company SK hynix is fueling the artificial intelligence race as its has officially shiped its samples of 12-layer HBM4E to major customers, as reported by Korea Herald on Thursday.
This comes approximately three weeks after Samsung Electronics claimed the industry's first such shipment. Now the comapany has pushed itself into the customer-qualification stage, as reported by Korea Herald.
SK hynix said its 12-layer part improves energy efficiency by over 20 per cent as compared with previous HBM4 generation. It delivers 48 gigabytes of capacity and pin speeds of up to 16 gigabits per second, as reported by the Korea Herald.
The South Korean semiconductor major has further applied its Advanced MR-MUF packaging process, which reduces gaps between stacked chips with a protective material to reinforce structural stability and dissipate heat. As per the company, the process reduces thermal resistance by nearly 17 per cent compared with HBM4. This is a significant achievement as faster memory generates more heat in dense computing environments.
In high-bandwidth memory, which is packaged alongside GPUs to ease the data bottlenecks in AI training and inference, sample delivery is the entry point to customer validation rather than proof of mass production. The stage carries added weight for HBM4E because the generation is moving toward more customized designs: logic base dies are tailored to individual customers, raising the value of early validation slots.
"To be clear, both companies have only shipped samples so far. Neither company has said it has cleared customer qualification or begun volume production, the stages that will decide which supplier locks in orders for next year's AI platforms," an industry official noted, as reported by Korea Herald.
With a revenue share of about 58 per cent as of the first quarter of 2026, SK hynix leads the HBM market, noted Korea Herald citing Counterpoint Research. On the other hand, Samsung and Micron each are trailing at roughly 21 per cent.
— ANI
Reader Comments
Impressive numbers from SK hynix – 20% better energy efficiency and 17% less thermal resistance. But let's be real, sample delivery is just the start. Samsung also shipped samples weeks ago. The real battle will be in customer qualification. I hope Indian cloud companies like Jio and Tata are working with both to get early access for our AI infrastructure.
The customization angle is crucial – HBM4E logic dies tailored to individual customers. This means closer partnerships between memory makers and AI chip designers. NVIDIA probably already has its pick. Samsung might have the manufacturing edge, but SK hynix has the early mover advantage in market share. Fascinating race to watch.
As someone working in Indian tech, this makes me question why we're still lagging in memory chip production. Our IT minister talks big about semiconductor fab, but where's the strategic focus on high-bandwidth memory? These Korean companies are moving at lightning speed while we're still in planning stages. Need a reality check, honestly. 😕
48GB capacity per stack at 16 Gbps speeds – that's going to be a game changer for AI training clusters. The thermal management improvement is particularly smart given how hot these dense computing environments get. SK hynix's MR-MUF process seems like a real innovation. Samsung better watch out, they're losing ground fast.
The article mentions 58% market share for SK hynix – that's dominance! But I'm skeptical about the customer qualification timeline. Both companies have only shipped samples, not production units. AI companies
We welcome thoughtful discussions from our readers. Please keep comments respectful and on-topic.