GPX10 Pro AI芯片用于边缘装置,提供100x效率增益和支持多个神经网络。 Ambient Scientific launches GPX10 Pro AI chip for edge devices, offering 100x efficiency gains and support for multiple neural networks.

环境科学已经启动了GPX10 Pro,这是一个为电池动力边缘装置设计的全局系统芯片,其功率、性能和面积在传统的32位微控制器上达到100倍的改善。 Ambient Scientific has launched the GPX10 Pro, an AI-native system-on-chip designed for battery-powered edge devices, delivering up to 100x improvements in power, performance, and area over traditional 32-bit microcontrollers. 芯片支持各种神经网络模型,如CNN、RNNs、LSTMs和GRUs,使得在边缘能够进行高效的人工智能推断。 The chip supports various neural network models like CNNs, RNNs, LSTMs, and GRUs, enabling efficient AI inference at the edge. 其可编程的AI核心可使用星云工具链进行定制,为变化中的AI工作量提供灵活性。 Its programmable AI cores can be customized using the Nebula toolchain, offering flexibility for evolving AI workloads. GPX10 Pro现在可供采样,批量生产预计将在2026年初开始。 The GPX10 Pro is now available for sampling, with volume production expected to begin in early 2026.