Semiconductor News & Analysis Feed

AMD has officially unveiled its first Zen 6 CPU, the EPYC 9996 'Venice', a 256-core/512-thread chip that delivers significant performance advantages over Intel Xeon 6 and NVIDIA Vera. Built on TSMC's N2 process, the chip supports up to 1.6 TB/s memory bandwidth and shows marked improvements across v

Intel has launched its new Xeon 6+ processor, codenamed Clearwater Forest, built on the advanced 18A process node, supporting up to 288 cores and 576 MB of L3 cache. Designed for data center applications, the processor adopts a disaggregated architecture combining multiple nodes and packaging techni

Intel has officially confirmed the launch of its next-generation Xeon 7 processor, codenamed Diamond Rapids, scheduled for 2027. Built on the refined 18A-P process node, this chip is expected to deliver 50% more cores and double the memory bandwidth compared to the Xeon 6. It will also support PCIe

NVIDIA's upcoming ARM-based server CPU, Vera, has demonstrated impressive performance in early benchmarks, outpacing both Intel Xeon and AMD EPYC processors. The chip, featuring 88 custom Armv9.2 cores and supporting up to 1.5TB of LPDDR5X memory with 1.2TB/s bandwidth, delivers a 1.63x performance

NVIDIA has released initial Linux benchmark results for its new 88-core server CPU, Vera, showing strong performance in specific workloads that rivals AMD EPYC and Intel Xeon processors. Vera features a custom 'Olympus' CPU core based on the ARM instruction set, similar to Apple's chips, but without

NVIDIA's new Vera CPU processor demonstrates significant performance advantages in benchmark tests, particularly excelling over leading industry processors like Intel Xeon and AMD EPYC in specific workloads. This breakthrough represents NVIDIA's strategic move into the server CPU market, showcasing

NVIDIA's new Vera CPU has demonstrated superior performance in its first benchmarks, outperforming AMD EPYC and Intel Xeon processors. Built on ARM architecture with 88 custom Olympus cores, Vera is designed specifically for Agentic AI and inference tasks. It delivers a 63% performance boost over it