News

Hygon is developing a new C86 server processor with up to 128 cores and 512 threads, marking a major step in China’s effort to build more of its own data center hardware. The upcoming C86 5G family is expected to deliver more than 15% higher instructions per clock than the company’s current generation while adding wider AI and vector processing support. The processor is

AMD has released a new Adrenalin hotfix driver to address a Windows 10 installation problem affecting some Radeon graphics card owners. The Adrenalin Edition 26.6.3 Hotfix Preview Driver is intended to fix the Device Manager warning that appeared after the previous 26.6.2 release, which added FSR 4.1 support for Radeon RX 7000 series graphics cards. The issue mainly affected Windows 10 systems using RX

Qualcomm has introduced High Bandwidth Compute, or HBC, a new memory architecture designed for AI accelerators that need more capacity and bandwidth without the growing power cost of traditional high bandwidth memory. The technology places a dedicated compute layer beneath stacked LPDDR memory, creating a tightly integrated design intended to reduce one of the biggest limits in modern AI systems. The company says HBC

Qualcomm has announced the Dragonfly C1000, its first server CPU designed for AI data centers and large scale cloud workloads. The company says the chip will use its custom Oryon CPU architecture, exceed 250 cores, reach clock speeds above 5GHz, and aim for leading single core performance when commercial systems arrive in 2028. The Dragonfly C1000 is part of Qualcomm’s broader push into AI

Qualcomm has officially introduced its Dragonfly platform, a broad data center strategy that combines AI accelerators, server CPUs, networking technology, advanced memory, and custom silicon services. The initiative is aimed at the growing demand for AI infrastructure, where companies want faster inference performance, better energy efficiency, and hardware that can scale across large data center deployments. Dragonfly is designed as a full stack platform

Intel’s Core Ultra 7 270K Plus has dropped to $264 during the Prime Day sale, giving PC builders a lower cost route into a high core count desktop processor. At this price, the chip becomes a strong option for people who want one system for gaming, content creation, multitasking, and general productivity. The processor normally sells closer to $300, so the current discount cuts

ASUS has released a new BIOS update that restores Transparent Secure Memory Encryption, or TSME, on supported consumer Ryzen systems. The feature was reportedly removed through earlier AMD BIOS updates, but the latest ASUS beta firmware brings it back for owners of compatible AM5 motherboards. The update includes AMD’s AGESA ComboAM5 PI 1.3.0.1b Patch A code and is currently available as BIOS version 2401

IBM has unveiled new 0.7nm chip manufacturing research built around a technology it calls Nanostack. The company says the approach could greatly increase transistor density by stacking transistor structures vertically, rather than relying only on making them smaller across a flat chip surface. The announcement places IBM among the first companies to publicly discuss semiconductor technology below the 1nm generation. While the 0.7nm label

AMD’s next EPYC server processor platform, codenamed Venice, could reach higher shipment volumes than NVIDIA’s Vera CPUs in 2027, according to a new industry forecast. The estimate points to 6.75 million Venice units for AMD, compared with 5.75 million Vera CPUs for NVIDIA. Both companies are preparing for a period where CPUs become more important in AI infrastructure. GPUs remain central to training and

Lenovo has warned that high prices for DRAM and NAND memory could remain a long term problem, with costs unlikely to return to the levels seen in early 2025. The company’s outlook suggests that the pressure created by AI data centers may continue affecting PCs, laptops, smartphones, SSDs, and other consumer devices well into the next decade. Memory prices began rising sharply in late

Micron has signed 16 long term agreements with major memory customers, a move that suggests high DRAM prices and limited supply could continue for years. The agreements reportedly secure memory volumes for five years and include terms designed to prevent customers from walking away if market conditions change. The company has described the deals as Strategic Customer Agreements, or SCAs. They cover annual memory

AMD’s Ryzen 7 5800X3D 10th Anniversary Edition appears to have sold out almost immediately after launch, with some listings already appearing on resale sites for far above its official $349 price. The limited reissue was expected to attract owners of older AM4 systems, but early availability has been poor for buyers hoping to purchase the processor at its intended price. Some reseller listings have

Goodram is preparing new DDR4 memory modules that start at just 4GB, a capacity that has become increasingly rare in modern desktop PC builds. The company’s upcoming Rival DDR4 Radiant lineup will also include 8GB, 16GB, and 32GB options, but the return of 4GB sticks reflects how sharply the memory market has changed. For years, 16GB of RAM was widely seen as the practical

Intel’s upcoming Nova Lake desktop processors could push power consumption to a new level, with a fresh leak claiming that the top dual tile models may reach up to 474W under PL2 boost limits. The report is unconfirmed, but it suggests Intel is preparing its next high end desktop platform for much larger core counts and more demanding power delivery requirements. Nova Lake is

Elon Musk has criticized IBM’s use of the 0.7 nanometer label for its latest chip manufacturing technology, arguing that modern process node names no longer clearly describe the physical size of features on a chip. IBM recently introduced its 0.7 nanometer manufacturing research, which uses nanosheet technology, vertical transistor stacking, and wafer bonding to increase chip density. The company has described the technology as

Marvell has introduced two new CXL memory controllers designed to help data centers use available memory more efficiently. The Structera X and Structera A chips use built in hardware compression to reduce the amount of DRAM needed for data, an approach that could become more valuable as AI workloads continue to strain global memory supply. The new controllers are aimed at large server systems

NVIDIA has become the top revenue leader in the data center Ethernet switching market, marking another major expansion beyond graphics cards and AI accelerators. The company reportedly generated $2.1 billion in Ethernet switching revenue during the first quarter of 2026, helping it capture 21.5% of the global market. The wider data center Ethernet switch market reached a record $15.4 billion in the quarter, up

Microsoft has started rolling out new Secure Boot certificates for eligible Windows 10 and Windows 11 PCs, arriving just before older certificates began reaching their expiration dates. The update is important because Secure Boot helps protect PCs before Windows even starts, reducing the risk of malware attacking the system during startup. The rollout replaces aging Secure Boot certificates first issued in 2011 with newer

Anthropic may begin asking some Claude users to verify their age or identity by uploading official documents such as a passport, driver’s license, or government issued ID. The change is expected to begin on July 8 and will apply only in certain situations, not to every Claude account. The company has updated its privacy policy to allow identity checks when needed for account management,

Microsoft has released a Windows 11 update that improves Bluetooth reliability, audio stability, device management, and headphone controls. The update, KB5095093, is optional for now, but it could be worth installing early if you regularly use wireless earbuds, headphones, speakers, or other Bluetooth devices with your PC. The update is especially useful for people who have dealt with unreliable Bluetooth connections, delayed audio pairing,