With HBM in short supply, AMD's next-generation of adaptive SoCs will be switching from HBM to LPDDR5X memory. The Versal Premium Gen 2 Memory on Package chips target the same compact form factor, but with a 15+ year projected lifecycle
We take a look at the AMD EPYC 8005 "Sorano" and see one of the biggest generational upheavals in server CPUs we have seen since 2019
For Computex 2026, Asustor had their upcoming third-generation Flashstor NASes on display. With support for either 6 or 12 M.2 SSDs, these all-flash NASes continue to be an interesting offering
With NVIDIA DGX Station systems now shipping, Supermicro had their Super AI Station on display a Computex, showing off the GB300 system in all of its 208 billion transistor glory
We check out next-generation of 800V DC gear for powering huge next-gen AI accelerators. Even the busbars are liquid-cooled
Today IBM is unveiling their nanostack transistor architecture. Meant to drive chip construction in the sub-1nm era in the 2030s, nanostack aims for building better and smaller chips by building them taller via wafer stacking
We are at Qualcomm Investor Day 2026 where the company is talking High Bandwidth Compute, its HBM alternative, Its Dragonfly C1000 CPU, and AI accelerators
Today OpenAI and Broadcom showed off the new OpenAI Jalapeño platform and gave a few details on the hardware platform
The June 2026 Top500 supercomputer list has been released, and an Arm-based system has retaken the #1 spot. Delivering 2.2 EFLOPs, China's CPU-only LineShine has become the world's fastest supercomputer
We found a HPE Spaceborne Computer 4 compute module bringing AI inference to the moon later this year and the storage may surprise you