Tag: #ai-chip44 results
Elon Musk announced Tesla's AI5 chip has successfully completed tapeout, positioning it to become one of the highest-volume AI chips produced. Partnerships with TSMC and Samsung enabled this milestone. Stronger AI6, Dojo3, and other chips are under development.
NIO's 2026 ONVO L90 will launch with a self-developed 5nm autonomous driving chip. This underscores NIO's ambition to control the full autonomous driving stack. The in-house chip enhances performance and integration.
Li Auto CEO announces Mach 100 chip team's paper accepted to 2026 ISCA Industry Track, first for auto industry. Dataflow architecture boosts AI efficiency over GPGPU, with 1280 TOPS per chip (3x Nvidia Thor effective). Debuts in new L9 at 55.98w RMB Q2.
Tenstorrent's Wormhole chip, led by Jim Keller, uses NoC grid of Tensix cores with RISC-V control, explicit SRAM, and Ethernet interconnects, ditching GPU SIMT for deterministic AI compute. Features native 32x32 tiling, double NoC, and scalable multi-chip expansion. SDK tt-Metalium enables low-level programming.
Alibaba's Damo Academy unveiled XuanTie C950, the latest RISC-V CPU core for high-performance cloud and AI computing. The chip targets AI agents amid Alibaba's shift to open-source RISC-V architecture. It was introduced at the company's annual ecosystem conference in Shanghai.
Alibaba Group Holding Ltd. is launching a new chip tailored for agentic AI and inference computing. This addition bolsters its semiconductor portfolio to advance AI ambitions amid surging demand.
Alibaba DAMO Academy likely to unveil major chip tomorrow at Shanghai RISC-V conference, targeting AI Agent compute surge. Event is 2026 XuanTie RISC-V Ecosystem Conference. Builds on prior server-grade RISC-V CPU debut.
AWS invited a TechCrunch reporter for an exclusive tour of its Trainium chip lab shortly after announcing a $50 billion investment in OpenAI. The Trainium chip has won over major AI players like Anthropic, OpenAI, and Apple. This lab is central to Amazon's AI infrastructure push.
Cerebras Systems Inc. picked Morgan Stanley to lead its initial public offering. The AI chipmaker is renewing its bid to go public, per sources familiar.
Moore Threads has achieved full-stack compatibility between its flagship MTT S5000 GPU and Alibaba Cloud’s Qwen3.5-series AI models. This milestone supports China’s tech self-reliance efforts amid US chip restrictions. The Beijing-based firm was founded by former Nvidia executive James Zhang Jianzhong.