Hardware · Top stories
Japan to Mass-Procure 3D-Printed Terra B1 Rocket-Powered Interceptor Drones
Japan's military will mass-procure Terra B1 interceptor drones, which are 3D-printed and designed to counter threats like Shahed drones and IAI Harpy loitering munitions. This procurement allows Japan to rapidly acquire effective counter-drone equipment, following a trend seen in other conflicts where traditional anti-air systems proved unsustainable against drone attacks.
Minisforum Launches AI Agent NAS N5 and AI Mini Workstation MS-S1 with AMD Ryzen AI Max+ Pro 495
Minisforum introduced the AI Agent NAS N5 Max-P495 and the AI Mini Workstation MS-S1 Max-P495, both featuring AMD's Ryzen AI Max+ Pro 495 processor. These devices are designed for local AI model execution and data storage, offering privacy and reduced latency for AI workloads.
AMD Unveils Threadripper Halo Station AI Workstation with 96 Cores and Dual MI350P Accelerators
AMD announced the Threadripper Halo Station, a new AI workstation featuring a 96-core Threadripper Pro 9995WX CPU and dual liquid-cooled Instinct MI350P accelerators, at IFA 2026. This workstation is designed for running large AI models, with AMD claiming it can handle trillion-parameter models, and is expected to be available through OEM partners.
Raspberry Pi Timeline Highlights Key Milestones from 2006 to 2012
The Raspberry Pi project began in 2006 at the University of Cambridge to address declining computer science applications and lack of hands-on experience. The Raspberry Pi Foundation was established in 2009, leading to the launch of the $35 Raspberry Pi Model B in 2012, which quickly became a best-selling British computer.
Lenovo Launches New Yoga Android Tablets with 7 Years of Updates and 4K Displays
Lenovo introduced two new flagship Android tablets, the Yoga Tab Gen 2 and Yoga Tab Plus Gen 2, featuring 4K displays, high-end chipsets, and a commitment to 7 years of Android and security updates. This extended update support positions the tablets as long-term devices in the competitive Android tablet market.
Intel's Core Ultra 400 'Nova Lake' Mass Production Set for Q4 2026, Q1 2027 Launch
Intel's Core Ultra 400-series 'Nova Lake-S' CPUs are reportedly scheduled for mass production in Q4 2026, with the initial 28-core version launching in Q1 2027. This timeline confirms earlier leaks and indicates a staggered release for the new desktop processor platform, which will include a 52-core model later in 2027.
Sugon Teases 64-Thread Mobile Workstation for AI with Domestic CPU and 16GB VRAM
Chinese supercomputer manufacturer Sugon has revealed an upcoming mobile workstation featuring a domestic 64-thread processor and 16GB of VRAM, capable of processing 50 tokens per second on local 35B MoE AI models. This development signifies advancements in China's domestic hardware capabilities for AI applications, offering a portable solution for demanding computational tasks.
Philips Hue Integrates Nanoleaf Light Panels with New $40 Module
Philips Hue now supports Nanoleaf Hexagon, Triangle, and Mini Triangle wall panels through a new $39.99 Philips Hue x Nanoleaf Wall Light Module. This integration allows Nanoleaf panels to function as full Hue lights within the Hue ecosystem, enabling gradient light, scene compatibility, and use with Hue automations and entertainment features.
SK hynix Delays Hybrid Bonding for HBM to HBM5, Citing 775-Micron Thickness Limit
SK hynix announced that hybrid bonding technology for High Bandwidth Memory (HBM) will not be ready for HBM4E and is now projected for HBM5 at the earliest. This delay is due to the industry-standard 775-micron thickness cap for HBM cubes, which necessitates thinner dies and narrower gaps for increased DRAM layers, posing manufacturing challenges for current technologies.
Qualcomm Introduces Adreno Neural Fusion GPU with Dedicated AI Cores for Next Snapdragon Chip
Qualcomm announced its Adreno Neural Fusion GPU, which will be integrated into the next flagship Snapdragon chip and features dedicated AI GPU cores called "Matrix Cores" and an 18MB High Performance Memory cache. This new architecture aims to improve graphics quality and power efficiency by processing AI models directly within the graphics pipeline, resulting in a 40% power improvement when enabled.
Blue Origin Awarded NASA Contract for Mars Telecommunications Network
NASA has awarded Blue Origin a contract worth up to $700 million to develop a high-performance Mars telecommunications orbiter by December 31, 2028. This orbiter will transmit science data, imagery, navigation, and mission communications for spacecraft operating on and around Mars, supporting NASA's future human missions to the planet.
Saab Unveils A3 Collaborative Combat Aircraft Concept for 2030s Deployment
Saab introduced its A3 collaborative combat aircraft concept, an autonomous, tailless, supersonic drone designed to complement its Gripen E fighters by the early 2030s. This development signifies Saab's entry into the high-end uncrewed combat aircraft market, contrasting with lower-cost alternatives and potentially influencing future European air combat strategies.
Chinese court freezes $318M in Nexperia assets amid ownership dispute with Wingtech
A Chinese court in Dongguan has frozen 2.14 billion yuan (approximately $318 million) in assets belonging to Dutch chipmaker Nexperia and its equipment arm, ITEC. This action provides Chinese parent company Wingtech Technology with leverage in its ongoing dispute to regain control of Nexperia, which Dutch authorities intervened in during 2025.
Arm Faces Shareholder Pushback Over CEO's Potential $800 Million Pay Package
Arm is facing potential shareholder opposition to a performance-based pay package for CEO Rene Haas that could reach $800 million if the company achieves a $2 trillion valuation. Proxy advisory firms ISS and Glass Lewis have recommended shareholders vote against the plan, citing its excessive size and the uncommon nature of such compensation structures in the UK market.
Solid-state cooler uses waste heat for refrigeration, potentially for data centers
Scientists at KIT and the University of Tsukuba developed a solid-state cooling system that converts heat into mechanical work for refrigeration, eliminating the need for an electric actuator. This technology could enable processors and data centers to recycle their own waste heat for cooling, offering an alternative to conventional vapor-compression systems that use refrigerants with global warming potential.
US tightens restrictions on foreign-made drones and robots, impacting global market
The U.S. implemented new restrictions and tariffs on foreign-made advanced robotic systems and drones in July and August, citing national security concerns. These measures aim to limit foreign technology in strategic industries, but analysts suggest they may lead to a fragmented global market rather than a clean U.S.-China split, as Chinese manufacturers hold significant scale advantages.
XCENA and Samsung Unveil MX1 CXL Device with Integrated RISC-V Compute and SSD Support
XCENA, in collaboration with Samsung, has introduced the MX1, a CXL memory expansion device that also integrates 3072 RISC-V cores for compute and supports SSDs. This device addresses the growing memory demands of machine learning models by combining high-capacity memory expansion with onboard processing capabilities.
Apple to launch AI-powered smart glasses in 2027, focusing on voice and camera features
Apple is expected to release its first AI-powered smart glasses in 2027, which will integrate Visual Intelligence and Siri for real-world interaction without a display. These glasses will primarily function through voice and camera input, offering features like environmental context, translation, and audio playback.
US Navy's Saildrone Surveyor USV successfully launches missiles and performs EW tasks
The U.S. Navy successfully tested the Saildrone Surveyor SD-3001 Unmanned Surface Vessel (USV) in live-fire exercises, launching two Joint-Air-to-Ground Missiles (JAGM) and demonstrating electronic warfare capabilities. This marks the first time the Navy has integrated an armed drone sailboat into a carrier strike group, indicating a shift towards autonomous combat assets.
ICE Plans to Purchase Boston Dynamics Spot Robots for Officer Safety
Immigration and Customs Enforcement (ICE) plans to spend between one and two million dollars on Boston Dynamics Spot robots and accessories to enhance officer safety and situational awareness in hazardous environments. This acquisition follows ICE's recent spending on other technologies, indicating a continued investment in new tools for law enforcement operations.
Samsung Galaxy S26 Ultra becomes top-selling Android phone globally in Q2 2026
The Samsung Galaxy S26 Ultra was the best-selling Android smartphone globally in Q2 2026, marking the first time an ultra-premium Android device has led Android sales in a June quarter. This indicates a shift in the Android market where a high-end flagship model is outselling budget devices, suggesting a potential change in consumer preference or Samsung's market strategy.
Intel's 14A Process Defect Density Drops Faster Than Expected, CFO Compares to 22nm
Intel's 14A fabrication process is showing faster-than-anticipated defect density reduction, with CFO David Zinsner comparing its progress to the successful 22nm node. This development indicates potential for strong performance in Intel's future manufacturing, with risk production planned for late 2027 and high-volume manufacturing in 2028.
Meta Introduces MTIA 300 Custom Accelerator for AI Recommendation Models
Meta has unveiled MTIA 300, its first in-house accelerator designed for training ranking and recommendation models, which integrates networking and collective communication directly into the chip. This development addresses the communication bottlenecks prevalent in recommendation model training, where network performance can be as critical as computational throughput.
YMTC Aims to Become World's Largest NAND Producer by End of 2027, Plans $4.9B IPO
Yangtze Memory Technologies Co. (YMTC) informed investors it plans to become the world's largest NAND flash producer by the end of 2027, surpassing Samsung and SK hynix. The company filed to raise 33 billion yuan ($4.9 billion) on the Shanghai Stock Exchange's STAR Market to fund production line upgrades and R&D, which is significant for the global memory market and China's semiconductor ambitions.
NASA's HPSC Processor Enters Testing, Promises 100x Performance Boost for Spacecraft
NASA's High Performance Spaceflight Computing (HPSC) processor, developed with Microchip Technology, has entered a testing phase at JPL. This new radiation-hardened system is designed to deliver a hundredfold improvement in performance per watt compared to current space-qualified systems, enabling more advanced space exploration missions.
Glass Substrate Chip Packaging Nears Qualification, Commercial Product Delays Persist
Glass-core substrates, intended to replace organic chip packaging, are in final qualification by Absolics, but commercial product integration continues to be delayed. Despite technical advantages like increased interconnect density and reduced warpage, manufacturing challenges such as chipping and flatness control are contributing to the slower-than-anticipated adoption.
Xbox Executive Discusses Affordability and Digital Entitlements for Next-Gen Console
Xbox chief executive Asha Sharma stated that affordability is a key consideration for the next-generation Xbox console due to rising hardware costs, partly driven by AI spending. The company also plans to introduce a feature allowing players to convert physical game discs into digital entitlements for over 1,000 titles.
Fujitsu's Monaka CPU features 2nm compute die, 5nm stacked cache, and 256-bit SVE2
Fujitsu unveiled details of its 144-core Monaka server CPU, confirming a design with a 2nm compute die, a separate 5nm die for the entire last-level cache, and dual 256-bit SVE2 vector units. This design aims for AI performance and power efficiency in green AI data centers, with evaluation samples available now and volume production in 2027.
NASA develops new bimodal nuclear rocket design to reduce Mars travel time
NASA engineers have conceived a new design for a bimodal nuclear rocket, which could significantly reduce the travel time for crewed missions to Mars. This advancement aims to address the challenges of long-duration space travel, including astronaut health and engineering complexities.
d-Matrix Unveils Raptor AI Accelerator with 100 TB/s Bandwidth via 3D Stacked DRAM
d-Matrix introduced Raptor, an AI accelerator featuring a TSMC 4nm compute die bonded directly onto a custom DRAM die, achieving 100 TB/s bandwidth per card. This design inverts traditional 3D stacking to improve energy efficiency and throughput for generative AI inference, potentially offering higher performance than HBM-based solutions.
Arm details AGI server CPU with two 70-core N3P chiplets and 2 TB/s UCIe fabric link
Arm provided additional technical specifications for its AGI data center CPU at Hot Chips 2026, confirming its design and planned configurations for late 2026 shipment. The AGI CPU features two 70-core N3P chiplets, a 2 TB/s UCIe fabric link, and a 12-channel memory controller, differentiating its approach from competitors.
HP signs multi-year global agreement to license Huawei WiFi patents
HP has entered a multi-year global agreement to license WiFi patents from Huawei, despite Huawei being blacklisted by the U.S. This agreement follows a patent dispute and allows HP access to a pool of 2,000 wireless network connectivity patents.
Intel's Wildcat Lake budget chip uses UCIe for cost reduction and 18A compute
Intel's Wildcat Lake, an entry-level chip, integrates the Universal Chiplet Interconnect Express (UCIe) to reduce manufacturing costs. This approach allows Intel to use its advanced 18A process technology in a budget-focused product, diverging from typical N-1 budget chip designs.
Intel Details Crescent Island AI Accelerator with Larger Caches and Deeper XMX Engines
Intel provided new architectural details for its Crescent Island AI accelerator at the Hot Chips symposium, highlighting increased cache sizes and deeper XMX engines. This accelerator is designed for lower-power, inference-first AI workloads, differentiating it from high-power training-focused GPUs.
Apple updates Mac mini and Mac Studio with M6 and M5 Ultra chips for AI development
Apple released updated Mac mini and Mac Studio desktops featuring new M6 and M5 Ultra chips. These updates focus on enhancing local AI inference and software development capabilities, building on the existing popularity of these machines for such workloads.
China slows exports of germanium and quartz to Taiwan, affecting semiconductor and optics supply chains
China is delaying or restricting exports of germanium and quartz-based materials to Taiwan, impacting the island's semiconductor, optics, and robotics industries. This action creates supply constraints and causes Taiwanese manufacturers to lose orders due to extended customs procedures under China's export control regime.
SiFive Launches BigSky SF-2U870, its First RISC-V Server Platform
SiFive introduced its first server platform, the BigSky SF-2U870, featuring 32 P870-D cores and support for the RVA23 standard. This platform enables out-of-the-box compatibility with Ubuntu 26.04, marking a significant step for RISC-V in the server market.
LG monitors installed McAfee ads on connected PCs via driver updates
LG monitors have been installing an application, "LG Monitor App Installer," onto connected computers through Windows Update, which then pushed McAfee pop-up ads. Microsoft has since compelled LG to remove the McAfee ads, but the method of software installation raises privacy concerns.
Intel Details Xeon 7 'Diamond Rapids' CPUs with up to 256 P-cores and 1.28 GB LLC
Intel released new details on its upcoming Xeon 7 'Diamond Rapids' CPUs, which will feature up to 256 P-cores, 1.28 GB of last-level cache, and utilize the 18A-P process, UCIe-S interconnects, and AVX 10.2. These data center CPUs are scheduled for release in 2027 and represent Intel's next-generation architecture and packaging advancements.
Nvidia to Extend CUDA Support to RISC-V, Outlining Server-Grade Requirements
Nvidia plans to extend its CUDA platform to support RISC-V CPUs, enabling RISC-V processors to integrate with GPU compute applications. This expansion requires RISC-V CPUs to meet specific server-grade criteria, including adherence to server SoC and platform specifications, ACPI support, and PCIe coherency.