👩💻 Project Rainier—the Most Powerful Supercomputer from Amazon
Amazon Web Services (AWS), Amazon's cloud computing division, has started building an AI computing "ultracluster" called Project Rainier. Anthropic, the AI startup that built chatbot Claude and raised a $8 billion investment from Amazon since September, will be running this project.
"When completed, it is expected to be the world's largest AI compute cluster reported to date available for Anthropic to build and deploy their future models on," says the statement from Amazon.
💻 Project Rainier features
This massive supercomputer will be made of hundreds of thousands of Amazon Trainium 2 AI chips. The exact parameters of the future cluster have not yet been disclosed. However, Amazon claims that Project Rainier's performance will be over five times faster than the current generation of Anthropic's AI models.
Another feature of Project Rainier is its distributed nature. The supercomputer will consist of several clusters interconnected by high-speed data networks. This, for example, is different from Elon Musk's xAI Colossus project, assembled in Memphis from 100 thousand graphics processing units (GPUs) Nvidia H100 with the prospects of expanding it to 1 million GPUs.
Project Rainier will launch in 2025.
🗄 The use
By releasing its own AI chips, Amazon is challenging Nvidia — a de facto monopolist in this market. Today, Nvidia controls 70% to 95% of the entire AI chip market. Chips from Amazon are expected to be 30-40% cheaper, but at the same time, the data transfer speed will be several times faster. Apple has already become a customer.
To keep up with Google, Microsoft, and other tech giants in the AI-technologies race, in the next decade, Amazon is going to invest more than $100 billion in the development of cloud computing and artificial intelligence.
More on the topic:
⏺ A Supercomputer Built the Biggest Simulation of The Universe
