Razer joins AI bandwagon with external AI accelerator backed by iconic AMD chip architect (www.techradar.com)

Razer has unveiled a local AI hardware solution designed for developers, focusing on on-premises work with the Tenstorrent external AI accelerator. This innovative device enhances AI computing capabilities through Thunderbolt connectivity, allowing multiple accelerators to be daisy-chained for local multi-accelerator workloads. By enabling direct control over datasets, models, and experiments, developers can avoid recurring subscription fees, thus promoting a more flexible and cost-effective workflow. Significantly, the accelerator leverages Tenstorrent's Wormhole architecture, supporting a diverse open-source software stack for various AI applications, such as large language models and image generation. Additionally, the device can function as a standalone tower or be integrated into rack environments, making it scalable from individual to clustered deployments. With the backing of Jim Keller, an influential figure in chip design, this collaboration between Razer and Tenstorrent aims to prioritize power, flexibility, and mobility for edge AI development, marking a substantial step in making advanced AI technologies accessible to developers everywhere. Pricing and availability details are yet to be disclosed.