Amazon Web Services (AWS) announced its new artificial intelligence-focused technologies, which are the trend of the year, within the scope of the Las Vegas-based re:Invent conference. The company announced both its own designed artificial intelligence processors and a new artificial intelligence chatbot. Taken as a whole, this news gives us an idea about the company’s long-term goals for artificial intelligence platforms.Amazon developed its own processors for artificial intelligence
AWS introduced its latest generation of AI chips designed for model training and running trained models. Trainium2, which is understood from its name as being designed for model training, aims to provide up to 4 times better performance and 2 times energy efficiency compared to its predecessor. Amazon promises these chips will allow programmers to train models faster and at lower cost due to reductions in energy use. Similar to Microsoft’s OpenAI support, Amazon-backed Anthropic announced plans to build models using Trainium2 chips.
Amazon also introduced the Graviton4 processor. Graviton4, which is for general use, is based on Arm architecture and consumes less energy than Intel or AMD chips. Amazon promises a 30 percent increase in overall performance when using a trained AI model built into a Graviton4 processor. As a result, Graviton4 will enable AWS customers to process larger amounts of data, scale workloads, improve time-to-result, and reduce total cost of ownership.Although Amazon develops its chips for the company, it does not break its ties with third-party providers such as Nvidia, on the contrary, it strengthens them. Amazon says it will offer enterprise customers cloud access to Nvidia’s new H200 AI GPUs. It will also run more than 16,000 Nvidia GH200 Grace Hopper Superchips exclusively for Nvidia’s research and development team. A similar move came from Microsoft, which introduced Azure Maia 100 and Cobalt 100.
50483
