aiqbench takeChips & Infra
aiqbench analysis
The MemoryAI Photonic Appliance addresses the AI memory wall by using silicon photonics to extend HBM-class bandwidth across multi-rack distances, effectively decoupling GPU capacity from physical package constraints. This approach allows CPUs and GPUs to transparently share terabytes of memory at high data rates, significantly improving efficiency for agentic AI workloads without replacing existing on-die HBM.
3
Sources coveringHPCwire, Dealroom.co, 36Kr
Penguin Solutions Unveils Photonic Appliance for Shared CPU and GPU Memory at SC26
HPCwire1 min read

Other coverage of this story
Related coverage
AI Products
Lightweight Time-Frequency Vibration Sensing for Consumer Device Condition Monitoring
arXiv · Oct 12, 2026

Chips & Infra
Marvell and Penguin Solutions Collaborate to Address Data Center AI Memory and Accelerate Inference at Scale
Marvell Technology · Nov 16, 2026
AI Products
China's Light Industry Digital R&D Design Tool Adoption Rate Surpasses 86%
36Kr · Nov 5, 2026

Chips & Infra
Marvell and Penguin Solutions aim to end the RAMpocalypse with photonic memory fabric
The Register · Nov 16, 2026