China weighs wider AI export controls, Financial Times reports
Beijing is considering limits on AI model weights, training data, chip fabrication abroad and foreign tech acquisitions,...
By Mara Chen-Doyle/section
Beijing is considering limits on AI model weights, training data, chip fabrication abroad and foreign tech acquisitions,...
By Mara Chen-DoyleNvidia published new Vera architecture disclosures, including Olympus core details and early SPEC CPU 2026 figures, as i...
By Mara Chen-DoyleIan Buck said Nvidia has shipped large volumes of standalone Grace servers as the company pushes CPUs harder for AI data...
By Mara Chen-DoyleMoonshot AI’s 2.8 trillion-parameter open-weight model is cheap by API, but running it yourself still means feeding a me...
By Felix ArandaNvidia detailed Rubin GPU changes meant to improve agentic AI inference, including faster tensor math, softmax, memory h...
By Felix ArandaNvidia says its first custom data-center CPU edges AMD’s Epyc 9755 in an unofficial SPEC CPU 2026 integer run.
By Felix ArandaNvidia’s engineering lab demo put Vera Rubin NVL72 racks, liquid cooling and 800VDC power hardware in front of reporters...
By Mara Chen-DoyleFortinet will work with Intel on its Security Processor 6, giving Intel Foundry a disclosed customer outside the AI chip...
By Mara Chen-DoyleFortinet will work with Intel on its SP6 firewall ASIC, giving Intel 4 its first named external customer, though neither...
By Mara Chen-DoyleA federal court left AI training fair use intact while approving a record copyright settlement tied to Anthropic’s store...
By Mara Chen-DoyleBloomberg says the GLM developer has activated part of a giant AI site built without Nvidia hardware, as China pushes do...
By Mara Chen-DoyleThe retailer lists a $400 discount on Apple’s 14-inch MacBook Pro with 48GB unified memory, a 1TB SSD and the M5 Pro chi...
By Mara Chen-Doyle