ASUS showcased the ExpertCenter Pro ET900N G3 at NVIDIA GTC 2026, a deskside AI supercomputer powered by Grace Blackwell Ultra with 748GB unified memory and up to 20 PFLOPS AI compute.
Plugable’s TBT5-AI Thunderbolt 5 enclosure lets developers add a powerful GPU to a Windows laptop and run local AI models without relying on the cloud.
Arduino VENTUNO Q is a new AI-focused development board powered by Qualcomm Dragonwing hardware. It targets robotics, edge AI, and autonomous machines running locally.
OpenAI and Foxconn are collaborating on US-made AI infrastructure hardware, aiming to strengthen supply chains and accelerate next generation data center development.
SAPPHIRE’s new EDGE AI mini PCs pack AMD Ryzen AI 300 chips with 50 TOPS NPUs, Radeon graphics, and Copilot+ readiness in a 117 mm chassis built for real-time edge AI.
Cadence has taped out the first LPDDR6 memory IP system, built to handle the bandwidth demands of AI inference, chiplet architectures, and next-gen SoCs.
Kioxia’s UFS 4.1 flash with CBA tech boosts performance and power efficiency in AI smartphones, Linux devices, and embedded hardware.
PCI-SIG has officially released the PCI Express 7.0 specification to members, delivering a massive performance jump to 128 GT/s per lane and up to 512GB/s of bi-directional bandwidth in a x16 configuration. This marks another milestone in PCIe’s evolution, continuing the group’s tradition of doubling bandwidth roughly every three years. While the specification is now … Read more