Amazon Web Services (AWS) has announced its plans to launch a new AI supercomputer called Project Rainier, built with the company’s own machine learning chips called Trainium. The ultracluster is expected to be finalized in 2025 and could potentially break records for size.
The project aims to rival Nvidia, the market leader in graphics processing units (GPUs) for AI programs. Amazon hopes to supply cheaper and more efficient chips, which will help it compete with generative AI giants like OpenAI. The company has made significant investments in building out its data center infrastructure and product lines to keep up with demand.
Project Rainier is expected to be used by AI startup Anthropic, valued at $18 billion, to train its AI models. Amazon’s Trainium chips are being employed in this project, alongside a collaboration on improving the technology after a $8 billion total investment from AWS. The company also plans to release a new Ultraserver, which will use its more advanced Trainium chips.
In an effort to capture the boom in demand for AI, AWS has allocated over $100 billion for AI infrastructure over the next decade. According to Amazon’s CEO, Matt Garman, customers would appreciate having multiple choices on the GPU side, citing Nvidia as the current market leader.
The project is a significant development in the rapidly growing AI space, where massive amounts of power and energy are required to function.
Source: https://www.forbes.com/sites/stephenpastis/2024/12/03/amazon-shares-rise-after-aws-announces-ai-supercomputer-nvidia-rival-heres-what-to-know