In Amazon’s latest push in the rapidly growing AI landscape, the company is set to invest $110 million into its new program, “Build on Trainium.” This new program aims to advance research in generative AI and will support researchers, academic institutions, and students.

In this vein, Amazon will provide free computing power to those who will use Trainium, the specialized deep learning AI chip of Amazon Web Services (AWS). This ties in with the core of the Build on Trainium program, which includes the provision of up to $11 million in credits for participating academic institutions to use Trainium. The credits will be offered to use its cloud data centers, so that universities and researchers to run complex AI models on AWS’s cloud infrastructure. In addition to this, individual research projects outside of AWS’s strategic academic partnerships will be able to apply for grants up to $500,000, (this ensures that a wide range of AI research can benefit from the initiative).

“As part of Build on Trainium, AWS and leading AI research institutions are also establishing dedicated funding for new research and student education. In addition, Amazon will conduct multiple rounds of Amazon Research Awards calls for proposals, with selected proposals receiving AWS Trainium credits, and access to the large Trainium UltraClusters for their research,” the company noted in an official statement. Furthermore, AWS also intends to establish a massive “research cluster,” comprising up to 40,000 Trainium chips. These chips will be made available to researchers and students via self-managed reservations, as well as providing access to the computational power needed to accelerate AI research.

Speaking more about Trainium, the chip is designed specifically for machine learning tasks, especially when it comes to deep learning training and inference. It can handle the large-scale computational demands of generative AI and machine learning models, thus providing a more efficient and cost-effective solution than general-purpose processors. It is also flexible and allows for low-level access to the hardware (which enables researchers to fine-tune their models for specific needs).

“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. It allows us to greatly expand our research on tensor program compilation, ML parallelization, and language model serving and tuning,” Todd C. Mowry, a professor of computer science at CMU, commented on the matter. The Build on Trainium program will also foster a collaborative environment for AI researchers – AWS is already working closely with the Neuron Data Science community, a virtual organization led by Annapurna Labs, to connect researchers to AWS’s broader technical resources and educational programs. It aims to facilitate networking among researchers, students, and institutions, this bridging the gap between academic research and AI advancements.

“Trainium is beyond programmable—not only can you run a program, you get low-level access to tune features of the hardware itself,” said Christopher Fletcher, an associate professor of computer science research at the University of California (as well as a participant in the Build on Trainium initiative). “The knobs of flexibility built into the architecture at every step make it a dream platform from a research perspective.”