Multiverse Computing has released Pulsar 16B, a 16.15 billion parameter AI model that delivers reasoning performance comparable to 30 billion parameter models but with half the parameters. Built on NVIDIA's Nemotron architecture and optimized through proprietary compression technology, Pulsar 16B offers improved speed and memory efficiency, making it suitable for deployment on lower-memory GPUs and single-node setups. The model excels in reasoning, coding, and tool-use benchmarks, and is available open-source on Hugging Face under the Apache 2.0 license, enabling enterprises to run advanced AI locally without cloud infrastructure costs.
TSMC and ASML have both increased their guidance, indicating ongoing supply constraints in AI infrastructure. Nvidia's Rubin N3 node is fully booked, and CoWoS capacity has expanded by nearly 50%, supporting strong demand. Concerns about Nvidia's Kyb...
Nvidia introduced Cosmos 3 Edge, a new AI model designed for robots and vision AI agents to navigate physical environments in real time. The company is expanding its AI ecosystem in Japan by partnering with major local firms like Fujitsu, Hitachi, an...
Netlist has prompted the US International Trade Commission (ITC) to investigate Samsung and its customers Google, Supermicro, Nvidia, and Broadcom for allegedly infringing two Netlist patents related to Samsung's high-bandwidth memory and DDR5 memory...
NVIDIA and Noetra Corp. are partnering to build Japan's first national AI infrastructure, the NVIDIA Vera Rubin AI factory, featuring 13,750 Vera CPUs and 27,500 Rubin GPUs. Supported by Japan's Ministry of Economy, Trade and Industry, this 140-megaw...
NVIDIA announced that leading Japanese companies in robotics and manufacturing are adopting its Cosmos, Isaac, Metropolis, and Jetson platforms to accelerate AI-powered intelligent machines across industries. The new Cosmos 3 Edge model enables real-...
