Amazon has announced a $110 million investment in artificial intelligence (AI) research at universities through its Build on Trainium program. The initiative will provide researchers with compute hours to develop AI architectures and optimizations using AWS Trainium UltraClusters.
AWS Trainium is a machine learning chip developed by Amazon Web Services (AWS) for deep learning training and inference. According to a press release by Amazon, the Build on Trainium initiative will open-source AI advancements, supporting a range of research areas from algorithmic improvements to large-scale distributed systems. It features a Trainium UltraCluster with up to 40,000 chips optimized for AI workloads.
The program also offers funding for AI research and education. Amazon Research Awards will provide selected proposals with AWS Trainium credits and access to UltraClusters. Researchers can collaborate and receive technical support through AWS’s Neuron Data Science community and other resources, as said in the press release by Amazon.
“AWS’s Build on Trainium initiative enables our faculty and students large-scale access to modern accelerators, like AWS Trainium, with an open programming model,” said Todd C. Mowry, a professor of computer science at Carnegie Mellon University (CMU). “It allows us to greatly expand our research on tensor program compilation, ML parallelization, and language model serving and tuning.”
Amazon Web Services is described as a comprehensive cloud platform offering over 200 services, including infrastructure technologies and emerging solutions such as AI and machine learning. It serves a wide range of customers including startups, enterprises, and government agencies, with a broad partner network focused on providing cost-effective, scalable tools across various industries.
Amazon provides various services including online and in-person shopping, delivery of fresh produce, media distribution, devices like Alexa and Fire TV, cloud solutions through AWS, logistics operations, customer service, and entertainment platforms such as Amazon Originals, Prime Video, and Audible.
