top line
Amazon Web Services, Amazon’s cloud computing division, announced the launch of a new AI supercomputer built with its machine learning chips. This could become one of the largest used for training AI models, and is aimed at taking on chip manufacturing giant Nvidia.
Amazon Web Services CEO Matt Garman speaks at WSJ Tech Live conference at Montage … (+)
AFP (via Getty Images)
important facts
Amazon Web Services’ (AWS) new supercomputer will be the Ultracluser, built with the company’s Trainium chip. It’s aimed at AI programs, and the company hopes it could rival chipmaker Nvidia’s sought-after graphics processing units (GPUs). The Wall Street Journal first reported on Tuesday.
The supercomputer, dubbed “Project Rainier,” is expected to be completed in 2025 and could break records for its size, Dave Brown, vice president of computing and networking services, told the Journal. .
Following the announcement, Amazon’s stock rose more than 1% to nearly $213.
AWS is also working on Project Ceiba in partnership with Nvidia. This is another supercomputer for AI programs that hosts more than 20,000 Nvidia Blackwell GPUs, chips used in data centers.
The Project Rainier announcement comes alongside the news that Amazon will release a new Ultraserver, a cloud server used to train AI programs that uses its more advanced Trainium chip, while Apple will use its more advanced Trainium chip for its Apple Intelligence AI technology. Benoit Duplin said the company plans to start using the Trainium chip in Apple’s senior director of machine learning and AI said on Tuesday.
Main background
AWS has been in the AI space since the early days of AI technology and is poised to capture the boom in demand for this technology. The company is making major investments in building out its data center infrastructure and product line to keep up with demand, with two $4 billion investments aimed at competing with generative AI giants such as OpenAI. We have strengthened our relationship with our partner company Anthropic. Amazon also hopes to make progress in this area through its Trainium silicon chips, potentially delivering cheaper and more efficient chips.
big number
Over $100 billion. That’s the amount Amazon, which has a market capitalization of more than $2.2 trillion, told the Journal in June that it would focus on AI infrastructure over the next 10 years.
important quotes
“Currently, there’s really only one choice on the GPU side, and that’s Nvidia. We think customers appreciate having multiple choices,” said Matt Garman, CEO of Amazon Web Services. told the Wall Street Journal.
tangent
AI models, the technology used by generative AI companies like ChatGPT and Anthropic, require large amounts of power and energy to function. This model processes large amounts of data using complex algorithms that require the fastest available technology and powerful GPUs. Chipmaker Nvidia has enjoyed astronomical success in recent years as the AI boom has increased demand for its GPUs and chips over those made by competitors. The company is now one of the most valuable companies in the world, with a market capitalization of over $3.3 billion, making it the clear market leader.
Read more
