Semiconductors & AI Chips Weekly Briefing

March 13, 2026

Nvidia, AMD, and Samsung lead a week of AI chip breakthroughs, new industry standards for optical interconnects, and surging hardware demand.

Nvidia prepared for its GTC 2026 conference with plans to launch new AI processing chips and the Vera Rubin system, while reportedly developing an open-source AI agent platform called NemoClaw to compete with rivals. ^123456789
Nvidia, AMD, Broadcom, Meta, Microsoft, and OpenAI established the Optical Scale-up Consortium to create an open specification for AI infrastructure, aiming to overcome the limitations of copper in large-scale AI clusters. ^1011121314
Samsung and SK Hynix were selected by Nvidia to supply sixth-generation High-Bandwidth Memory (HBM4) as both companies ramp up production for on-device and data center AI applications. ^{151617181920212223}
AMD expanded its portfolio with the Ryzen AI Embedded P100 series and collaborated with Xanadu and Rolls-Royce to advance quantum workflows and computational fluid dynamics on its Instinct GPUs. ^2425262728
Taiwan Semiconductor Manufacturing Company (TSMC) reported a 30% revenue increase for the first two months of 2026, driven by sustained global demand for AI hardware despite a month-over-month decline in February. ^2930313233
Samsung Electronics implemented cost-cutting measures, including requiring executives to fly economy, while simultaneously navigating rising energy and raw material costs and record R&D spending of $25 billion. ^34353637
Other developments include Marvell's introduction of PCIe 8.0 SerDes technology, Intel's announcement of Core Ultra 200S Plus desktop processors, and ASML receiving approval for expansion near a protected nature reserve. ^{3839404142434445464748}

Sources

Company Intelligence on Autopilot

Get your own weekly briefings on the companies that matter to you

Start for free No credit card required