News

NVIDIA has introduced a Synthetic Video Detector designed to help newsrooms, broadcasters, and media teams identify footage that may have been created or altered by artificial intelligence. The tool will be offered as an NVIDIA NIM microservice, making it easier for organisations to deploy inside existing media and verification workflows. It analyses video frame by frame and produces a classification score showing how likely

Samsung, Micron, and SK hynix have reportedly reduced or abandoned plans to develop proprietary CXL memory controllers after major server customers showed limited interest in paying more for tightly integrated products. The three companies had explored custom controllers for Compute Express Link memory, which can expand the amount of memory available to servers beyond traditional processor connected DRAM. The reported change suggests that large

Samsung Electronics America is restructuring its US operations as it prepares to move its headquarters from New Jersey to Texas, affecting hundreds of positions across its consumer electronics business. A regulatory filing listed 739 positions connected to the company’s Englewood Cliffs office in New Jersey. Samsung later clarified that this number includes both jobs affected by workforce reductions and positions being transferred to another

TSMC says building semiconductor factories in the United States costs four to five times more than constructing similar facilities in Taiwan, but the company is still expanding its Arizona operations with another $100 billion investment. The additional spending will support four more facilities and extend TSMC’s presence beyond its first Arizona fab. The company says customer demand for advanced chips remains strong enough to

NVIDIA chief executive Jensen Huang used a mid July visit to Japan to strengthen the company’s business ties while also winning praise for the personal style he brought to meetings with Japanese executives. The trip included discussions about large AI infrastructure projects, robotics, domestic semiconductor opportunities, and NVIDIA’s next generation Vera Rubin platform. It also produced a major agreement involving thousands of Vera CPUs

Intel is expanding memory support across selected Xeon 6 processors, bringing 8000 MT/s DDR5 RDIMM compatibility to current platforms while preparing second generation MRDIMM modules running at 8800 MT/s for early 2027. The faster RDIMM support will be available on selected Xeon 6700P Granite Rapids and Xeon 6 Plus Clearwater Forest processors. Intel plans to enable the feature through a platform release and BIOS

NVIDIA is reportedly acquiring large amounts of unused long distance fiber infrastructure across the United States, a move that could support a private network connecting AI data centres and cloud computing partners. The company is said to be purchasing dark fiber routes with configurations reaching as many as 100 fiber pairs. Dark fiber refers to installed fiber optic cables that are not yet carrying

China’s domestic semiconductor industry could ship around 5 million chips in 2026 as local companies increase production in response to US export restrictions and government pressure to reduce reliance on foreign hardware. The estimate suggests shipments may rise from roughly 4 million units in 2025. Growth is expected to continue over the next two to three years, with the sector potentially expanding at a

China may be able to produce some advanced semiconductors without access to extreme ultraviolet lithography, but the country will struggle to compete commercially if it continues relying on older deep ultraviolet equipment, according to Zeiss chief executive Frank Rohmund. Export restrictions prevent companies such as ASML and Zeiss from supplying China with leading EUV systems and related components. Chinese foundries have therefore turned to

TSMC is reportedly preparing to raise semiconductor manufacturing prices by between 5 percent and 10 percent in 2027, covering both advanced and mature production processes. The expected increases would affect a wide range of customers, including companies that rely on older nodes such as 28nm, 16nm, and 12nm, as well as those ordering chips built with processes below 7nm. This means the impact could

The global NAND flash shortage is expected to continue through 2026, but improving production capacity could bring the market closer to balance during the second half of 2027. Memory suppliers in South Korea, the United States, Japan, and China are reportedly preparing to upgrade existing production lines and add more equipment. These changes should increase NAND output after a prolonged period of tight supply

A Costco shopper has reportedly purchased a powerful iBuyPower gaming desktop for about $1,220 after finding the final display unit marked down far below its original price. The system includes an AMD Ryzen 7 9800X3D processor, an NVIDIA GeForce RTX 5070 graphics card, 32GB of DDR5 memory, and a 2TB solid state drive. It also came with a keyboard and mouse. The desktop was

Samsung is reportedly preparing to begin large scale production of high bandwidth memory using hybrid bonding in 2029 or 2030, a schedule that could align the technology with NVIDIA’s future Feynman AI accelerators. Hybrid bonding is considered an important next step for HBM manufacturing because it removes the small metal bumps traditionally used to connect stacked DRAM layers. This allows the layers to sit

NVIDIA has detailed the architecture behind its next generation Rubin GPU, revealing a 336 billion transistor design built to deliver much higher performance for agentic AI, mixture of experts models, and large scale inference. Rubin is manufactured using TSMC’s N3P process and combines two large dies inside one package. NVIDIA connects the dies through its high bandwidth NV HBI interface so the complete accelerator

NVIDIA has released detailed performance claims for its Vera CPU, positioning the chip as a purpose built processor for agentic AI, data processing, graph workloads, and other tasks that depend heavily on memory bandwidth and low latency. The Vera CPU uses 88 Olympus cores and a monolithic design rather than the multi chiplet structure common in modern x86 server processors. NVIDIA says this helps

NVIDIA has shared the first performance figures for its Vera Rubin NVL72 platform, claiming a tenfold increase in token throughput compared with the current GB200 Grace Blackwell system at the same power level. In an early DeepSeek R1 test, the Vera Rubin rack reportedly sustained 800,000 tokens per second while operating within a 150MW deployment. A GB200 NVL72 configuration delivered 80,000 tokens per second

Philips has introduced the Evnia 32M2N6901A, a 32 inch gaming monitor that combines a 4K QD OLED panel with a 165Hz refresh rate, fast response times, USB C connectivity, and built in speakers. The monitor is aimed at players who want high image quality without moving into the most expensive 240Hz or 360Hz 4K OLED category. It will be available from August for £649.99,

AMD’s upcoming Helios rack scale AI platform could carry a price between $5 million and $5.5 million per rack, according to an industry estimate. That would place it roughly 40 percent above the expected $3.5 million to $4 million price range for NVIDIA’s second generation Rubin systems. The estimate suggests AMD may no longer be positioning itself mainly as the lower cost alternative in

NVIDIA has provided a deeper look at DLSS 5, explaining how the technology will use neural rendering to improve game visuals while preserving the original work created by developers and artists. The company has faced criticism since introducing DLSS 5 earlier in 2026. Some observers were concerned that the technology would replace existing game assets with AI generated content or change a game’s visual

SK hynix expects memory supply conditions to become even more difficult in 2027, as strong demand for high bandwidth memory limits the capacity available for conventional DRAM. The company’s chief executive, Kwak Noh-jung, said next year could be the worst period in the industry’s history from a supply perspective. He also expects customer demand to remain above the company’s production capacity beyond 2030. The