Dell Technologies has partnered with independent AI research company Imbue to build a high-performance computing cluster for training foundation models optimized for reasoning. The $150 million agreement will involve the construction of a Dell PowerEdge server cluster with nearly 10,000 NVIDIA H100 Tensor Core GPUs. Imbue aims to develop AI agents with advanced reasoning capabilities that can perform tasks for people across various domains. The cluster will be used to train AI models and prototype agents, with the long-term goal of creating more capable and trustworthy AI agents that require less user supervision. The system includes smaller clusters for rapid experimentation and networking into a large cluster for efficient training of large-scale foundation models.