Amazon has announced a $110 million investment aimed at supporting university research in artificial intelligence (AI) through its new “Build on Trainium” program. Through this project, the company will provide academic institutions access to its Trainium chip infrastructure, specifically designed for deep learning training. This financial backing seeks to facilitate the creation of new AI architectures, machine learning libraries, and performance optimization in large-scale distributed environments on Amazon Web Services (AWS) cloud.
The program will also allow researchers from prestigious universities, such as Carnegie Mellon University (CMU) and the University of California, Berkeley, to access a research UltraCluster consisting of 40,000 Trainium chips. This access to cutting-edge computational power will help universities overcome budget constraints that often limit progress in advanced AI research.
AI Research Within Reach of Academics
Amazon’s goal with “Build on Trainium” is to support researchers in AI optimization and development projects without compromising scientific rigor. To avoid conflicts of interest, the company has ensured that the results of this research will be open-source, facilitating global access to innovations. Additionally, Amazon has announced that it will offer rounds of funding through the Amazon Research Awards program, granting AWS Trainium credits to selected projects that stand out for their innovative potential in AI.
Boosting the Training of AI Experts
Since the launch of its AWS Inferentia chips in 2019, Amazon has aimed to lead the development of cloud AI technology. With this new initiative, it seeks not only to promote fresh ideas but also to train the next generation of AI experts. Christopher Fletcher, an associate professor at the University of California, Berkeley, highlighted the flexibility of Trainium, describing it as “a dream platform for research, as it allows researchers to adjust hardware features to their needs.”
Furthermore, the program introduces the Neuron Kernel Interface (NKI), a programming interface for AWS’s Trainium and Inferentia chips, which provides researchers direct access to the chip’s instruction set, optimizing computational operations and enabling advancements in the field of AI.
Academic Partnerships and Open Collaboration
Amazon has also designed the program to foster collaboration between academics and technology experts through its Neuron Data Science community, which brings together researchers, startups, and AI specialists. Grant recipients of “Build on Trainium” will be able to connect with a network of experts and participate in technical training that facilitates the implementation of their projects.
With this initiative, Amazon aims to establish itself as a key ally in academic AI research while ensuring that scientific advancements are accessible globally.