Magic, a developer of an AI coding platform, partnered with Google Cloud to build cloud-based supercomputers for training large language models (LLMs).
The partnership aims to develop two supercomputers: G4 using A3 Mega VMs with NVIDIA H100 Tensor Core GPUs and G5 utilizing the NVIDIA Grace Blackwell platform. Additionally, the company claims that these computers will achieve 160 exaflops. These systems will support Magic's goal of creating an "automated AI software engineer and researcher" with a 100-million-token context window. Google Cloud will provide the AI infrastructure, including hardware systems, its Vertex AI platform, and expertise to help Magic bring its products to market.
Magic joins other AI startups, such as Arize AI, Character AI, Cohere, and Mistral, in utilizing Google Cloud's services for AI development and scaling.
By using this site, you agree to allow SPEEDA Edge and our partners to use cookies for analytics and personalization. Visit our privacy policy for more information about our data collection practices.