Amazon Web Services (AWS) and Anthropic are developing a supercomputer that will be five times larger in size and have five times the performance of the current one used by Anthropic for training its artificial intelligence (AI) models. Dubbed “Project Rainier,” this EC2 UltraCluster of Trn2 UltraServers will utilize hundreds of thousands of Trainium2 chips, surpassing their previous cluster’s size.
The new Trn2 UltraServers offer enhanced performance and cost efficiency for customers training and deploying AI models and future large language models and foundation models. According to David Brown, vice president of compute and networking at AWS, the new servers provide the fastest training and inference performance on AWS, enabling organizations to train and deploy massive workloads faster and at a lower cost.
Project Rainier is expected to be the world’s largest AI compute cluster reported to date, with more than five times the number of exaflops compared to Anthropic’s current model. This development comes after Amazon and Anthropic announced an expanded partnership that includes Amazon investing $4 billion in the AI company, bringing their total investment to $8 billion.
As part of this partnership, the convergence of affordable microcomputing and low-code software platforms is reshaping how enterprises approach innovation. Microcomputers like Raspberry Pi are being utilized for enterprise-class systems without the cost and complexity of traditional hardware, facilitating modern industrial edge computing across industrial computers.
The trend has significant implications in payments and commerce, where agility and scalability often determine market competitiveness. Affordable microcomputing hardware combined with user-friendly development platforms enables enterprises to meet complex workflows, diverse payment methods, and regulatory compliance while maintaining a seamless customer experience.
These technologies also empower non-technical staff by providing accessible development platforms with drag-and-drop interfaces, pre-built templates, and step-by-step tutorials, allowing employees in finance, marketing, or operations to create and deploy solutions without writing code. However, security and governance are essential concerns as the democratization of technology brings new challenges.
The partnership between AWS and Anthropic is expected to drive innovation and transformation in the AI industry, enabling organizations to train and deploy massive models faster and at a lower cost.
Source: https://www.pymnts.com/artificial-intelligence-2/2024/aws-and-anthropic-to-build-5x-larger-supercomputer