Ballistic Electron Transport Observed in Single-Crystalline Copper Thin Films
Researchers from POSTECH, Pusan National University and Mississippi State University have experimentally demonstrated ballistic electron transport in single‑crystalline copper thin films as thin as 80 nm and 150 nm wide. The copper films, grown by Atomic Sputtering Epitaxy, exhibit a surface roughness of ~0.2 nm and show negative bend resistance, confirming scattering‑free electron flow at temperatures below 85 K. This marks the first observation of ballistic transport in an industry‑standard metal at dimensions relevant to semiconductor interconnects. The breakthrough suggests a pathway to reduce resistance, delay and power consumption in future nanoscale wiring.
HighPoint Unveils High-Performance Gen5 AI Compute Platform for Ultra-High Inference Throughput Empowered by Hailo
HighPoint Technologies and AI‑chip maker Hailo announced the Enterprise Edge AI Compute Platform, built around HighPoint's Rocket 1604L Gen5 retimer and up to four Hailo‑8 or Hailo‑10H M.2 acceleration modules. The solution packs 160 TOPS (INT4) of inference power into a single...
Supermicro and Arm Advance Compute for the Agentic AI Era
Supermicro unveiled a new line of servers powered by Arm’s AGI CPU, a 136‑core, 300 W processor designed for high‑density, power‑efficient AI inference. The portfolio includes liquid‑cooled Open Rack Wide systems that can house up to 336 CPUs per rack and...
NextSilicon to Productize Arbel RISC-V Core Into 64-Core Enterprise Processor for AI and HPC
NextSilicon announced it will productize its Arbel RISC‑V core into 64‑core and 128‑core enterprise processors slated for early 2028. The chips will run at 3.4 GHz on a newer TSMC node, retaining a 10‑wide issue pipeline, 480‑entry reorder buffer and four...
D-Matrix Corsair AI Inference Platform Enters Full Production to Meet Customer Demand
d-Matrix announced its Corsair AI inference accelerator platform has entered full production and will begin volume shipments to priority customers this summer. The accelerator is designed for heterogeneous data‑center clusters, pairing with GPUs to accelerate the decode phase of large...
Cadence and Intel Foundry Expand DTCO Partnership for Intel 14A Process
Cadence and Intel Foundry have signed a multi‑year agreement to deepen their Design Technology Co‑Optimization partnership, beginning with Intel’s 14A process node. The deal merges Cadence’s AI‑driven EDA and IP solutions with Intel’s process and packaging expertise to deliver production‑ready...
From Evidence to Authority: Bounded Gate Authority for Governed Semiconductor Realization
The article introduces Bounded Gate Authority (BGA) as a new decision layer within the SEGA‑AI™ framework that converts normalized, admissible evidence into release authority for advanced semiconductor products. It argues that the explosion of data from design, fab, packaging, and...
AMD RDNA 5: Board Partners Apparently Expect New Radeon Gaming GPUs only at the End of 2027 or Beginning of...
Board partners at Computex 2026 indicated AMD’s next gaming GPU generation, RDNA 5, is unlikely to arrive before the second half of 2027, with some expecting early 2028. AMD has only officially launched RDNA 4 with the RX 9000 series in early 2025. The extended gap would...
Synopsys and Samsung Foundry Extend AI-Driven Design Collaboration for Advanced 2nm and Multi-Die Systems
At the SAFE Forum 2026, Synopsys announced an expanded AI‑driven design partnership with Samsung Foundry covering its second‑ and third‑generation 2 nm processes. The collaboration delivers production‑ready digital and analog implementation flows that improve power, performance and area while shortening time‑to‑market....
Linux 7.1 Helping Intel Arc Battlemage Graphics Achieve Better Performance
Recent benchmarks show the upcoming Linux 7.1 kernel delivers noticeably better graphics performance for Intel’s Arc B580 Battlemage GPU compared with the current Linux 7.0 kernel. The tests used Ubuntu 26.04 LTS with Mesa 26.1, isolating the kernel upgrade as the only variable. Results indicate...
Broadcom Told the Truth. The Market Hasn’t Heard the Rest of It Yet.
Broadcom reported Q2 AI semiconductor revenue of $10.8 billion, a 143 % YoY surge, and lifted full‑year AI guidance to $56 billion while reaffirming its $100 billion FY2027 target. Despite the blowout numbers, the stock fell 12 % after‑hours as the earnings Q&A revealed two...
AMD and Imperial College London Partner on AI, HPC and Scientific Discovery
AMD and Imperial College London have announced a strategic partnership to accelerate AI‑enabled scientific discovery and sovereign high‑performance computing in the United Kingdom. The collaboration will leverage AMD’s accelerated‑compute hardware and ROCm open‑source software alongside Imperial’s expertise across climate modeling,...
FuriosaAI Partners with Broadcom to Build Next-Generation Inference Platform for the Agentic Era
FuriosaAI announced a strategic partnership with Broadcom to create a third‑generation AI inference accelerator built on its Tensor Contraction Processor architecture. The new platform will combine a 2 nm compute die, a dedicated I/O die, and HBM4/4E memory using Broadcom’s advanced...
Micron Shipping 245TB 6600 ION U.2 and E3.L Data Center SSDs
Micron Technology has begun shipping a 245 TB 6600 ION SSD, the highest‑capacity commercial solid‑state drive on the market. Built on G9 QLC NAND, the drive delivers up to 84× better energy efficiency and 8.6× faster AI preprocessing while consuming only 30 W,...
Europe’s Third Way on AI Is Easier Said than Done
On June 3, the European Commission released its Technological Sovereignty Package, comprising a revised Chips Act, a Cloud and AI Development Act, an open‑source strategy and a digital‑energy roadmap. The initiative aims to reduce EU reliance on non‑European chipmakers and...