Elon Musk's xAI has unveiled Colossus, one of the world's most powerful supercomputers designed for cutting-edge AI research, featuring 200,000 Nvidia GPUs, advanced liquid cooling, and a plan to train massive language models that could redefine AI capabilities.
Taiwan's TSMC, the world's biggest chipmaker, warns tariffs on microchips and manufacturing equipment could undermine its plans to expand chip production in the United States, as semiconductor supply chains are global and import restrictions could reduce demand for U.S.-made chips.
NVIDIA, Foxconn, and Taiwan are teaming up to construct a groundbreaking AI factory supercomputer, boasting 10,000 Blackwell GPUs, empowering researchers, startups, and industries with cutting-edge AI computing resources to drive innovation and adoption across the nation.
Leaked details reveal that former Apple designer Jony Ive and OpenAI CEO Sam Altman are developing a pocket-sized, screen-free, contextually aware device envisioned as a 'third core device' alongside laptops and phones, set for release by late 2026 after OpenAI's $6.5 billion acquisition of Ive's AI hardware startup.
NVIDIA unveiled groundbreaking platforms, models, and technologies like Isaac GR00T N1.5, GR00T-Dreams, Blackwell systems, Cosmos Reason, and Isaac GR00T-Mimic for advancing humanoid robot development, with major robotics companies adopting NVIDIA's solutions.
Nvidia's CEO praises the talent of Chinese AI researchers, warning that overly restrictive US policies like export controls on AI chips could hinder America's ability to lead in AI development amidst healthy global competition.
NVIDIA introduced NVLink Fusion, enabling companies like MediaTek, Marvell, Cadence, Fujitsu, and Qualcomm to create custom AI silicon by integrating their CPUs with NVIDIA GPUs, paving the way for high-performance AI infrastructure tailored to specific industry needs.
Warby Parker shares surged around 16% after Google announced a $150 million partnership to develop stylish smart glasses, with the first line set to launch after 2025 using Google's Android XR operating system and AI assistant.
Windows ML, a new runtime optimized for on-device AI model performance, offers a unified framework leveraging client silicon, optimizations from partners like AMD, Intel, NVIDIA, and Qualcomm, simplified deployment, advanced silicon targeting, performance improvements, and cross-build compatibility, empowering developers to create AI-infused applications.
SiFli's SF32LB52J chip with 512K SRAM, 16M PSRAM, and a dedicated display peripheral powers the upcoming Core Time 2 smartwatch, with SiFli offering an open-source SDK and support for porting PebbleOS.