OpenAI's Jalapeño Chip Sets New Standards for AI Inference Speed and Efficiency
OpenAI's Jalapeño chip delivers 1.5 to 1.9 times more AI work per watt and 1.7 to 3.6 times lower latency than comparison systems, according to first results.
OpenAI's Jalapeño chip delivers 1.5 to 1.9 times more AI work per watt and 1.7 to 3.6 times lower latency than comparison systems, according to first results.
Nearly 90% of global humanoid robots were made in China in 2025, according to a recent report highlighting the country's growing robotics industry.
Cerebras announced the CS-4 AI accelerator on August 24, 2026, with double the performance of its predecessor, achieving up to 4,400 tokens per second per user.
Taiwan indicted nine people, including an Nvidia senior manager, for allegedly smuggling 74 high-end AI servers to China, violating US export controls.
Nvidia AI server prices are expected to increase by about 15 percent due to a memory shortage, affecting systems with Vera Rubin and Grace Blackwell chips.
RayNeo's new AI glasses provide text overlays with 97% transparency and 1,300 nits brightness, skipping the camera and speaker.
Ulanqab, Inner Mongolia, has seen 70% of its 12.5 gigawatt data center projects announced in the last year, making it one of Asia's fastest-growing compute clusters.
Nvidia has partnered with Cloverleaf Infrastructure, a data center developer that raised $300 million in 2024, to support its AI expansion efforts.
Waymo has developed a custom chip for its robotaxis, achieving over 1,000 TOPS performance in 2026.
China is allowing limited shipments of Nvidia's H200 chips to help local AI companies compete with U.S. firms. ByteDance and Tencent each received about 10,000 processors.
Relativity Networks raised $22 million in funding to deploy hollow-core fiber, which transmits data 30% faster than conventional fiber.
Apple’s camera-equipped AirPods, confirmed via leaked macOS 26.7 RC footage, may not be surveillance devices but rather tools to enhance Siri’s visual intelligence.