Intel has launched the Xeon 6+ Clearwater Forest processors, featuring up to 288 E-cores to address increasing data center power and space constraints. These chips utilize the Intel 18A manufacturing process to double the core count and provide a five-fold increase in last-level cache compared to previous generations. The platform is specifically designed to help organizations consolidate legacy server estates into high-density environments, particularly for 5G core and cloud-native workloads.

Source

Microsoft is updating the Windows 11 kernel to support the Nvidia RTX Spark, a heterogeneous Arm-based system-on-a-chip that combines a 20-core Grace CPU with a Blackwell GPU. A new Workload Profile Scheduling system has been implemented to intelligently distribute tasks across these diverse processing cores while balancing power and thermal limits. These architectural changes also include updates to the Prism emulation layer, which translates x86 instructions for Arm hardware to ensure legacy software compatibility.

Source

AMD has officially extended the support window for its AM5 desktop socket until at least 2029. This announcement updates previous commitments that only guaranteed platform longevity through 2027. The extension suggests that the socket will accommodate at least two more generations of processors beyond the current Zen 5 architecture.

Source

AMD announced two new octa-core processors featuring 3D V-cache technology during its Computex 2026 keynote. The Ryzen 7 7700X3D is a new entry for the AM5 platform, utilizing the Zen 4 microarchitecture with a 4.5 GHz boost clock. Positioned just below the existing 7800X3D, this chip offers 96 MB of L3 cache and a 120W TDP to provide a high-performance gaming option at a lower price point.

Source

Intel has officially launched its Xeon 6+ processor series, codenamed Clearwater Forest, which is the first data center CPU built on the Intel 18A process node. The flagship SKU, the Xeon 6990E+, features 288 Darkmont E-cores and is designed specifically for high-density workloads rather than raw single-core performance. This architecture utilizes advanced packaging technologies including RibbonFET, PowerVia, and Foveros Direct3D to stack twelve 24-core tiles atop base dies containing memory controllers.

Source