Reuters
SK Hynix says HBM4 on track despite spec changes
SK Hynix confirmed its HBM4 production timeline remains intact.
Reuters•
Back to home
Articles tagged with "ai-memory"
The Information
Inside NVIDIA's quiet pivot to Samsung HBM3E
After eighteen months of qualification delays, NVIDIA has quietly approved Samsung's HBM3E.
The Information•
Bloomberg
Microsoft, Micron strike long-term memory deal
Microsoft has secured multi-year HBM supply commitments from Micron.
Bloomberg•
The Register
Cerebras WSE-4 GA: wafer-scale inference is real
Cerebras has announced general availability of the WSE-4.
The Register•
BitDepth Editorial
Samsung's HBM3E qualification, finally. The full timeline and what it means for NVIDIA's 2026 allocation strategy.
For 18 months Samsung was the third HBM vendor on paper only.
BitDepth Editorial•
arXiv
DeepSeek V4 paper: 2x inference at iso-quality
DeepSeek's V4 technical report documents a new mixture-of-experts architecture.
arXiv•
BitDepth Editorial
Why every AI training run is now a packaging negotiation.
A bottleneck no longer measured in GPUs but in CoWoS slots.
BitDepth Editorial•
BitDepth Editorial
Cerebras WSE-4 is generally available. We ran the benchmarks. The numbers are real.
Wafer-scale inference has spent four years as a press release.
BitDepth Editorial•
No more articles to load