News

Montage Technology has started sampling a new DDR5 registered clock driver that supports speeds of up to 9200 MT/s, giving memory makers another option for faster RDIMM modules aimed at next generation data centers. The new RCD06 chip is designed for server memory, where bandwidth demand is rising because of AI, cloud computing, databases, and other memory heavy workloads. The launch matters because RDIMMs

TensorWave is raising $350 million to expand its AMD based AI infrastructure, with new deployments focused on AMD Instinct MI355X accelerators. The investment will help the company add more compute capacity for memory heavy AI workloads such as LLM training, high throughput inference, and generative AI production systems. The move gives AMD another visible win in the AI infrastructure market, where Nvidia still dominates

Nvidia’s future Feynman AI chip could become one of the first major products to use TSMC’s next generation CoPoS packaging technology, according to analyst Ming Chi Kuo. The new packaging method is expected to enter mass production in the second half of 2028 and is designed to support much larger AI chip packages than today’s CoWoS technology can handle. The report matters because advanced

AAEON has introduced new edge systems and developer boards powered by Intel Wildcat Lake processors, giving embedded developers more options for compact systems built around Intel’s Core Ultra Series 3 platform. The new lineup includes UP WCL Edge, UP Nexus WCL Edge, and matching developer boards with up to 48GB of LPDDR5 memory. Intel Wildcat Lake is designed as a flexible SoC platform for

Samsung is facing criticism after YouTuber Louis Rossmann said the company offered a $330 refund for a faulty 4TB 990 Pro SSD that now reportedly sells for much more in the current market. The dispute centers on whether Samsung should replace the drive or refund its current market value instead of the original purchase price. Rossmann said his 4TB Samsung 990 Pro SSD began

SK Hynix is reportedly preparing to begin mass production of its 375 layer NAND flash by the end of 2026, marking its next major step in the race against Samsung for higher density storage. The company has completed verification of the new NAND design and is expected to move production into existing fabs after converting them for the newer process. The move matters because

Amazon has made its AWS Graviton5 CPUs generally available, bringing a major upgrade to its custom cloud chip lineup. The new processor offers up to 25 percent better performance than Graviton4, doubles the core count to 192 cores, and adds support for faster DDR5 memory and PCIe Gen 6. Graviton5 is built for demanding cloud workloads, including AI inference, databases, analytics, real time applications,

Japan’s supply of tungsten hexafluoride, a gas used in advanced semiconductor manufacturing, has reportedly collapsed after tighter Chinese export controls disrupted access to high purity tungsten powder. The shortage could affect major chip and memory companies including TSMC, Samsung, and SK Hynix, all of which rely on suppliers that produce this critical material. The issue matters because tungsten hexafluoride is used in chipmaking processes

Nvidia is reportedly trying to turn its upcoming Vera CPUs into a new route for China data center revenue as GPU restrictions continue to limit its business in the country. The company has been speaking with potential Chinese customers about Vera availability, with orders possibly opening ahead of an August launch window. The move is important because Nvidia’s AI GPU business in China has

ASUS’s ROG Equalizer 16 pin cable has reportedly suffered its first burn out case, raising fresh concerns around high power GPU connectors and the safety accessories meant to protect them. The cable was marketed as a premium solution for 12V 2x6 graphics card power delivery, but a new image from a forum post appears to show melted plastic and burn marks on several connector

MSI says memory and GPU shortages are still hurting the PC market, and the pressure may continue into 2026. The company’s chairman, Hsu Hsiang, warned that rising memory prices and reduced graphics chip supply are making it harder for vendors to plan products, price systems, and keep the DIY PC market healthy. The one area that may improve sooner is CPU supply. MSI expects

Intel’s Core Ultra 7 251HX has appeared in leaked PassMark results, and the early numbers suggest the 18 core Arrow Lake HX chip can keep up with its 20 core siblings. The processor reportedly edges out the Core Ultra 7 255HX and Core Ultra 7 265HX in single core performance while also landing slightly ahead in multi threaded scoring. That result is surprising because

LG has expanded its UltraGear OLED gaming monitor lineup with two larger curved models, the 39GX90SB-W and 45GX90SB-B. Both monitors use 800R curved WOLED panels, support a 3440 x 1440 resolution, and offer refresh rates up to 240Hz for smoother gaming. The new models are larger versions of LG’s existing 34 inch curved UltraGear OLED monitor, but they push the lineup into a more

AMD’s Radeon RX 9070 XT has finally appeared in the Steam Hardware Survey after being missing for nearly a year, showing up with a 1.33 percent share in the May 2026 data. The sudden appearance is notable because AMD’s RDNA 4 cards have been visible in retail sales and enthusiast discussions for months, but most of them were absent from Steam’s public GPU rankings.

Amazon could deepen its AI chip partnership with Qualcomm as AWS looks for ways to lower inference costs and protect margins. A new analyst note suggests AWS may become a key customer for Qualcomm’s AI200 accelerators, which are designed for large language model inference and can support up to 768GB of memory per chip. The possible partnership matters because AI infrastructure costs are becoming

Global foundry revenue rose 3.7 percent in the first quarter as strong demand for AI chips helped offset the usual seasonal slowdown in smartphones. The increase shows how much the semiconductor market is now being shaped by AI hardware, high performance computing, and supply chain demand from PCs, notebooks, and consumer electronics. The first quarter is normally weaker for foundries because smartphone demand slows

FurMark is still useful for checking GPU heat, fan noise, and worst case power behavior, but it is no longer enough to prove that a modern graphics card is truly stable. Today’s GPUs from Nvidia, AMD, and Intel change clocks, voltages, power draw, and boost behavior constantly, so proper stress testing now requires a mix of synthetic tests, memory checks, real games, ray tracing

ASUS and MSI have started rolling out BIOS updates that bring AMD’s EXPO Ultra Low Latency support to select 600 series AM5 motherboards. The update helps older AM5 boards close the feature gap with newer 800 series models by adding better DDR5 memory optimization for supported systems. AMD introduced EXPO Ultra Low Latency at Computex as an extension of its existing EXPO memory overclocking

Maxsun’s Intel Arc Pro B70 32GB graphics card shows the biggest step yet for Intel’s Battlemage GPU lineup, delivering strong gains in AI, rendering, workstation tasks, and even gaming. The card is built for professional and prosumer workloads, but its performance also raises a clear question: why did Intel not release a gaming version of Big Battlemage? The Arc Pro B70 is based on

Microsoft’s June Patch Tuesday update fixes a record 206 security vulnerabilities across Windows, Office, Exchange Server, Edge, Hyper-V, and cloud services. One of those flaws is already being exploited, making this an important update cycle for both home users and IT administrators. The actively exploited issue is a Microsoft Defender elevation of privilege vulnerability tracked as CVE-2026-41091. If abused, it can allow an attacker